Episode 14 - The Grandfather of Microservices, Fred George

How you doing Fred? 

Not bad, not bad. Got my second coronavirus shot Wednesday. So. 

Cool. party time,. 

Few more days, but yes, it’s like sense of freedom and safety. 

Mm. I bet. 

Vegas is getting an influx of visitors like crazy right now. 

Is it. 

Cold? Cold winter and yeah, we’re casinos going to pitch capacity from 35%, but it looks like people are ignoring that just going Paul hoed. So, 

Wow. Do, do you spend a lot of time in the casinos, Fred, or do you avoid them? 

I, I actually spend a lot of time in them, but I don’t gamble. 

Oh. 

Really? Just, I just walk through them, 

But just observing people. 

Yeah. It, it turns out Vegas reminds me of when I lived in London. 

Oh, really? Of. 

All of all things. It’s the people watching is amazing in both places. You have amazing theater. I mean, we got six standing circuit shows in Vegas. 

Oh, good. 

It goes, all the performers come through and then the restaurants are just unbelievable. I mean, Gordon Ramsey has six restaurants alone. 

Wow. 

If you want entertainment, if you want great food, you want interesting people to see. I think CES is my favorite time. Cause that’s when about 200,000 geeks show up. 

Yeah. 

All of sudden the town looks like it looks like we’re Silicon valley for a week. The, I think it’s the black hat guys come in. Same time the adult video awards come in. You can tell the difference between the two. 

Nice, nice. 

Yeah. The whole town transforms when these events happen. It’s makes it kinda interesting to see, 

We’d like to start off with, what’s your name and where’d you come from and then we’ll go from there if that’s possible. 

Oh, Fred George programmer, where I come from east coast of United States originally is where I was born and raised. And, but I wrote my first code in 1968. 

Wow. 

We are not gonna, we’re gonna skip most of the resume here. 

Well, we know you had a long history with IBM. I think that’s how you got started really well, probably where you got a, a lot of your experience. I imagine. 

Yeah. 17 years at IBM, which was actually really formative experience. First of all, it’s a, I worked in the development side. In building the operating systems, so extremely bright people, great mentors, great training. As a side effect I figured out later is I learned how to work with big companies and be effective. Cause big companies at some level, all large companies are really bright despite what...

How you doing Fred? 

Not bad, not bad. Got my second coronavirus shot Wednesday. So. 

Cool. party time,. 

Few more days, but yes, it’s like sense of freedom and safety. 

Mm. I bet. 

Vegas is getting an influx of visitors like crazy right now. 

Is it. 

Cold? Cold winter and yeah, we’re casinos going to pitch capacity from 35%, but it looks like people are ignoring that just going Paul hoed. So, 

Wow. Do, do you spend a lot of time in the casinos, Fred, or do you avoid them? 

I, I actually spend a lot of time in them, but I don’t gamble. 

Oh. 

Really? Just, I just walk through them, 

But just observing people. 

Yeah. It, it turns out Vegas reminds me of when I lived in London. 

Oh, really? Of. 

All of all things. It’s the people watching is amazing in both places. You have amazing theater. I mean, we got six standing circuit shows in Vegas. 

Oh, good. 

It goes, all the performers come through and then the restaurants are just unbelievable. I mean, Gordon Ramsey has six restaurants alone. 

Wow. 

If you want entertainment, if you want great food, you want interesting people to see. I think CES is my favorite time. Cause that’s when about 200,000 geeks show up. 

Yeah. 

All of sudden the town looks like it looks like we’re Silicon valley for a week. The, I think it’s the black hat guys come in. Same time the adult video awards come in. You can tell the difference between the two. 

Nice, nice. 

Yeah. The whole town transforms when these events happen. It’s makes it kinda interesting to see, 

We’d like to start off with, what’s your name and where’d you come from and then we’ll go from there if that’s possible. 

Oh, Fred George programmer, where I come from east coast of United States originally is where I was born and raised. And, but I wrote my first code in 1968. 

Wow. 

We are not gonna, we’re gonna skip most of the resume here. 

Well, we know you had a long history with IBM. I think that’s how you got started really well, probably where you got a, a lot of your experience. I imagine. 

Yeah. 17 years at IBM, which was actually really formative experience. First of all, it’s a, I worked in the development side. In building the operating systems, so extremely bright people, great mentors, great training. As a side effect I figured out later is I learned how to work with big companies and be effective. Cause big companies at some level, all large companies are really bright despite what we think, but you have to understand what it motivates the company. I kinda learned how to do that in IBM. It was a bit of a shock when I started working for somebody who wasn’t IBM, cause it turned out it not, I not everybody’s IBM. 

I had the same experience when I left Amazon. Actually it was, again, you move out of a company where you’ve got like a comfort blanket of all sorts of things, and then you realize other places don’t have those things. 

Yeah. I, IBM was a very strong decision culture. It was easy to get decisions made. Mm, wound up in a consultant, say where situation, where it was a consensus culture. The positive thing was when everybody agrees, it’s kind of easy to get things done, but it turns out everybody agreed on the wrong thing. It’s almost impossible to get it to change. 

Mm. 

I ran across the decid on the wrong thing too often. All my decision, culture stuff didn’t work. And I sense, 

Did it take you a long time to rebuild that ability to make decisions when you don’t have necessarily the political alignment that you need? 

No, I’m pretty much type a, so I’m always gonna get my way at some level. Being type a you’re pretty ad agile about how you figure out how to do this. If this trick doesn’t work about this trick, and then now I’m at this trick, 

Yeah, I’ve heard you refer to as a hand grenade before. 

That that was, yeah, that was a direct quote from a CTO. I, I worked for in the daily mail. I was described as the hand grenade swing into development, which is a good job description as you can get it because it, there are no rules for hand grenades. 

Yeah. I, I, I remember meeting you briefly, I guess, probably about six years ago. This was at a project in, that was in London and you’d been brought in to kind of shake things up by making all of the developers and everybody around them probably feel sick that they didn’t really know what on earth they were doing with their careers. I was in a position as a consultant where we would pick the pieces up after the, how grade had gone off. 

Yeah. To some degree that was an engagement where I was brought in explicitly to do that. I mean, I was, my job was to be disruptive. I was like, okay, I can do that. And that’s not hard specialty. And, and so basically you change everything. You, you change people sit, you change the type of death they have, they change the program language you using, you change their titles. You do everything to upset the apple cart to then have a reform again, cuz the alternative was just fire the whole group of them and start over again somewhere else, take the whole thing offshore for example. 

How did from leaving IBM and lose losing that decision culture and having that way of working, how did you get from that to being the hand grenade, is there a whistle stop version of how you got from those two between those two points? 

I even at IBM I was a bit of a hand grenade. I, I it’s interesting cuz even though you consider I IBM to be a relatively state organization, their culture was, we believe in our wild ducks. We like our wild ducks. I, I kind of was always kind of in the wild duck category and my, I think my managers appreciated it. They’d always throw me into a situation that was basically broken and had me fix it. What they began to realize is after I fix things well, fixing things is kind of fun. I’ll break it again just for the fun. They would move me to another job where it was a priority broker and leave behind something that was something like you’ve ed do clean up after me, but it’s much easier to clean up me after I already made the change necessary. 

So. 

For, for example, that is, I was, I’m the grandfather of eclipse, the editor. I basically brought objects into my team and IBM and we, we started doing things a little strange way, relatively strange doing entrepre programming, built up a nice structure, flattened the organization out, empowered the managers. They moved me out cause everything was working fine. My predecessors, the one that, finished it up and put, clip out the door. 

Wow. Okay. So, I mean, I know you as the grandfather of microservices, so there is a, there’s a list of other products before that I guess like eclipse. 

Well, even that, but you also, if you’re running on a window system and you see the little dropdowns, you have the consistent spelling, little red X in the corner, that’s some of my work back then at IBM. 

Wow. 

If you ever heard of the net biased interface there, the net buoy associated with it, that’s something the work I did when I was working on land when it was brand new. 

Wow. If you got, what else can we add to the list then Fred? 

Well, I, I own the user interface architecture between Microsoft and IBM for a while. Right when we knew us two days put out the first telecommunication software that let two computers talk to each other. I could log onto one computer and run an app on a different computer. Wow. The first computer networking was some of my early projects. 

What era are we talking about when we’re talking about, getting computers to talk to one another, 

We’re talking about the seventies with the networking stuff. Land is the mid eighties. 

How does that compare to like the stuff that you’re doing today? Because obviously, on a daily basis, speaking of microservices, particularly we’re spinning up Docker containers here, there, and everywhere and getting them to talk to one another and actually frankly, that can be a pain in the us as well. So I mean, how is that? How, how, how have things differed from doing that work on a daily basis today to when you were doing that for the first time in the seventies? 

Well, I, I think every, I think overall I’m always kind of looking for that edge technology. I’m kind of the first guy that try to says, let’s see if I can make this thing, do something interesting. So that was the case of lands. I didn’t invent the concept of local area network, but I put it out the door, Actually the first land versions of stuff. I didn’t, I didn’t necessarily invent go interfaces, but I brought it to IBM. 

Yeah. How, how complex was it working in that I mean, how long did these projects take? Cause obviously we’re talking, I mentioned the Docker containers, you can do that in a couple of hours on your machine or, a few minutes to be honest, depending on how good you are at it. How long was that taking you in the seventies and the eighties? 

Well, I, I think everything always took longer than I wanted it to, but you could, I trans technology transformation is a 25 year process. 

Yeah. 

I’m in that first five years where there’s very few adopt, we’ve got in that early stage, the brave use olds doing things. For example, and you’re not selling to everybody in that environment, you’re selling to a very specific crowd, the crowd, that fears that technology is edge. 

Mm. 

If you’re getting into objects, you’re getting into object and programing languages. You’re probably working with a lot of the wall street banks, which are extremely competitive about having the best software to figure out the best reasons to do everything. 

Mm. 

You’ll find those bleeding edge cans that know that technology is competitive advantage. 

Oh, and is that still the case on wall street? Cuz I don’t necessarily know if that would be the case in the city in London. 

Is definitely the case in wall street. You’ll see it more in the day traders. 

Yeah. 

Day trading is where you do it, not the traditional banking, the, moving money around and stuff like that. The day traders are always, looking for that edge and those are the ones where you see the new languages pop up. 

Yeah. 

Those new things come in. You’ll you’ll see the guys that have the six giant screens. These are the guys that do that work. Yet where you won’t see us say in big banks, you’ll see that some of the private equity firms, the ones that are investing their own money, which don’t have as many rules. Yeah, you always wanna find these guys who figure technologies to pay advantage. Yeah. When I seek out clients, that’s what I’m looking for. People who figure it’s not a matter of, should I, if I’m a hotel change, should I buy more bacon for, or should I use some techn buy some technology? Those guys are looking for the cost to benefit, trade off. They don’t, they aren’t gonna bet on anything. They’re too conservative. You’re looking for the guys to say I’m opening a new hotel chain. I want people to walk through and not even stop with the registration desk. 

I want them go to write to a room and swipe their, swipe, their credit card and walk in the door. You looks for that sort of client. 

And, and how close are we to that? Or is that just to Fred George? I do. 

Oh no, that we’re. We have people that can do that now really general hotel change can do that. You just kinda get a text message on your phone when you’re getting close to the hotel that says, oh yeah, you are, here’s your room number, go walk in, going to the elevator and use your R F I D card to walk in. 

To your point that you mentioned before about being fi you are in working within that first five years, is this technology that we expect to be commonplace in five years or at least starting to break through in the next five years. You know where we at? I, 

It here in some degree in already, the question is when does it get to be like 20% penetrated? That’s in the five to 15 year timeframe. 

Yeah. 

When it gets to be 20 to 25 years, that’s when everybody’s gonna do it or else you get fired. For example, you can’t find a CIO today who doesn’t claim. They do agile. 

Yes. You. 

Don’t say, if you don’t say you do agile, you can’t get hired. Now what agile goes back to 19 95, 19 97. So count 25 years from there. Where are we now? 

Yeah, fair enough. I mean, I think we’ve we’ve come across one or two that are still not necessarily in that bracket, but there’s always gonna be a few dinosaurs, I guess, are gonna. 

Only have a few dinosaurs, a few guys and still for your buggy, whips are gonna be a big industry. 

One thing I wanted to ask you is because I’ve seen a lot of your talks and the, you tend to talk about what you talk a lot about microservices obviously, and we’ll get to that in a bit more detail, but you also talk a lot of about processes and methodologies and having started in waterfall and moved into agile and then there’s some things beyond that I’ll let you talk about, but where does your time tend to split? What are you most passionate about? It the technology or is it the process and how do those things work together for you? 

I, I think I’m most passionate about getting code out the door. Mm. I’m looking for, I go walk into a client. I’m trying to figure out what’s it gonna keep me from getting to code out the door. Sometimes it’s gonna be process. Sometimes it’s gonna be technology. Sometimes it’s just gonna be business attitudes. How does the business treat the, treat the system. We treat the programmers as a giant spreadsheet that they can just play with. You’re kinda looking, I’m looking for, what’s gonna keep me from being successful And because of my broad background for all way, from lots of management, IBM and management training there a business degree as well as obviously some technology background. And, and I, I, run an organization of over 200 programmers at IBM. So I run large organizations. I’m walking in with no biases about what needs to be done, but with an eye to what I can do. 

And there’s a lots of possibilities. 

How do you decide where you’re gonna go first? How do you find what the bottleneck is? 

It’s usually within a couple of few minutes, actually, usually within a few hours of showing up with a client, having that first conversation, and you always wanna kind of start with a person who’s spending the money. You figure out what are they trying to accomplish because what the it manager says and what the sponsor says, sometimes it’s radically different. 

It always one thing or is it always a combination of those three that you have to change in an organization to make it. 

You’re always gonna be touching all three. It’s just a matter to what extent you need to hammer on all three. When, when I walked into some of my engagements in Norway, turns out the talent level was fabulous. The, the team, the people they had were motivated, they were very strong. Technically they knew how to work together, all that collaboration stuff. I didn’t needed to do any work there that was already done. My focus was I on process, which they had not. I, I put the process in place. They were already were using Kotlin as their language. That was really good, but they weren’t using Kotlin as an object language, they were using Kotlin as cobalt. 

Right. OK. 

First of all, you had to teach ‘em what Kotlin is an object language looks like. So what OBjects looked like? It turned out it was a very applications, so we didn’t need microservices. They were like, we should need microservices. No, no, you don’t. You’re very traditional application. You have a set of rules you’re trying to implement. It’s nothing fuzzy about this problem. 

I was gonna ask you about that because the, what the latest talk that I’ve seen you do is on don’t do microservices or don’t why you don’t use microservices and obvious for the man who’s known as being the grandfather of microservices. Could you tell us more on the, on this show, particularly why you don’t use microservices in certain circumstances? 

I guess my favorite model is the conne model from Dave snot, which is a model of how you classify problems, 

Right? 

You have to look at problems and decide whether they’re very traditional. In other words, it’s the rules associated with what the answer has to be like. This is the way trades have to be settled. This is the way bank balances are calculated. This is how salaries, this is how you get tech calculation tax rates. Those are very traditional applications versus ones in a nail call, fuzzy, which are problems that don’t have precise answers. Like which stock should I buy? Should I loan you money? Is this a fraudulent transaction? These are fuzzy problems. So it turns out for fuzzy problem. You have competitive advantage in that space. If you’re able to try ideas out aggressively. The faster I can try something out the better off am I often in a fuzzy environment. If I’m solving a fuzzy problem, I want to be able to have a loosely coupled system, because I wanna be able to bright it, put something out there that doesn’t necessarily impact anything else I’ve done. 

Just to try an idea out. I don’t wanna have to schedule it, think about it, make sure it doesn’t impact anybody else. Do a three month dance or six month dance to get it out the door. I want to try something in minutes. Again, you look at various environments like day traders. This is their world. There are a giant foot is a problem. And they have competitive advantage. If they can put the little things out all the time, not wait for the big it guys to say, now’s the time for when you ship it. 

That, so it’s in those fuzzy problems. That’s the, that’s the place to use microservices. 

Yeah, because microservices right now is, are hard thinking about loosely come system. When you’ve been used to doing step one, step two, step three, to step four. Now it’s about, let’s do all four steps in parallel. We can’t get our heads around that. Cause we haven’t been trained to think that way. It’s not how we get business requirements from our customer. He says, this is how you do it. Here’s the steps we do. Here’s our checklist. It’s very laid out and has to be a fuzzy problem. That checklist is pretty much useless. Mm. He doesn’t, and the business guy doesn’t understand that he doesn’t understand he’s in a fuzzy problem. He’s looking for precise answers, but nevertheless, you wanna be fuzzy, but microservices is hard to get your head around. It’s a developer. Don’t, I recommend don’t do that unless you’re experience in doing it already, I can solve traditional problems with microservices. 

That’s very doable, but you don’t wanna do that unless you have experience in microservices. 

Cause there is quite a lot of set of work to get that working as an organization, having built a few microservices, architectures myself, it is a big learning curve for developers. Who’ve not done it before. 

Oh yeah. And I, I completely underestimated that originally. 

Oh really? Oh yeah. This from the mail online or from before that? 

Oh, well before that, my first microservice project was in 2005, 

Right. 

When I was doing some work for one of the big manufacturing firms in the us. And, and basically I was, had to play with the concept of services and some engagements I had in China hung with some of the, some of the great mentors in that space. Were kind of always asking question, well, how big is a service? I was like, okay, well here’s a chance to play. Let me make a service really small. I, I did that and we had a very successful project. We had a project that was estimated to take 18 months and we delivered it in nine weeks. 

Wow. That is fast comparison to the original estimate. 

Yeah. And, and so, yeah, and so I, and one level was wildly successful, but at, when I rolled off it kind of stopped. They stopped building bigger and bigger services. They stopped focusing on building small services. It was easier just to build bigger ones. 

Yeah. At this time we, what, 2005, what language are we talking in and how big are we talking? The services? 

We, were writing Java and were writing services that probably a couple hundred lines of code. 

Yeah. Okay. 

This was what I call a pinball architecture. We were basically trying to fill in an information gap. We had a, this record of information we needed to collect. You could figure out one piece of information if you had two other pieces, but I didn’t have one of those other pieces. Well, in order to get that piece, I had to get the other information from somewhere else. We built little services that I would take two pieces of information and generate the third. We basically took this little message and we ping ponged it around the services, trying to fill it in. If it’s missing something, you send it to somebody who helped you out. Well, if he couldn’t help you out, he sent it to somebody who could help you out. 

So what was the inspiration for this? Was it actually pinball machine? 

No, but it was my was kind of the best metaphor for it when you got there. Yeah. I think it was a matter of just trying to break things down into very small classes. So remember I I’m object trained. Sure. I, I was trained in objects from about 19 86, 19 87. We got to this, were, I was trained in small talk. We’re all about very tiny classes that are very, single responsibility classes. So I put that into my microservices. Let’s, let’s make every service do one thing. If you describe a service or use the word hand, it must be two services, break it apart. 

I see. That makes sense. In terms of how you were getting these service is to talk to one another, there are there’s this, obviously there’s several ways to implement services and get these services to talk to one another. Right. You could have it via an API, you could have it via events and you could have it via service discovery. Now I know that this has been quite a point of contention in terms of what is the best microservices architecture, and I know which camp you fall into, but I’d like you to explain that to us, to us on the call. Where, where did you start out and did you start out with that straight away and, has your theory about how you do this changed over time? 

Yeah. Theory has changed. I, I think we go back to the pinball, my pinball architecture, every service in which other service they go to get, get it piece of information. Okay. It’s kinda like every service was kind of registered as being the service and knows how to get, get you this piece of information. It turned out that, that created, it, obviously it solved the problem very nicely. There’s a lot of other things about that project that went extremely well. Again, I, I had an absolute stellar team. I had the guy who created the concept of bill process, ? 

Wow. Okay. 

I, I had probably one of the finest program would ever worked with he’d rejoin me on that team. So it was not near average team. This was, this was, a team of superstars. That being said, we did run across some problems. Like all of a sudden we’d drop a message. We’d drop this ball into the pinball machine and it wouldn’t come out. Sometimes it was just running around and around in circles because we had an infinite loop that we didn’t realize. Sometimes the ball would just kind of disappear. We began to worry about how do you detect these other situations? One of the problems about loosely coupled system is when things go wrong, you don’t have a lot of clues about what went wrong. I, I, I began searching for other architectures that helped us address that. I was in, when I was in London, working with a group, this was the early days of Kafka showing up. 

Some of my guys got into the Kafka bus and I’m looking at that event bus. I’m like, wow, this is kind of cool because it’s a permanent history of everything that went wrong. If you have a permanent history where anything went on, when something goes wrong, I can go back and reconstruct it. Mm. In fact, to some degree, you think about an event bus, well in traditional system, when I’m trying to debug a problem, and I got lots of components that are working in various ways, the first thing I do is reconstruct the timeline of what went wrong. So you’re pulling logs from here. Are you pulling logs from there? You try to rebuild that timeline of what works. If you’re building on an event bus, you have the timeline already that works done. 

Yeah, yeah, absolutely. 

Especially if you use an event bus, as I recommend that you have put everything on the event, bus, make it a rapids, don’t try to segregate it. Cuz that means you have to put it together again and put the timeline, skip that step, put everything on the same event bus, because if everything was on the same event, bus debugging is a dream. 

How does your cause you have a concept obviously of using several event buses. How does that, if, how, if you’ve got that approach of putting everything on the main event bus, how does that work with the additional buses that are slower moving, 

To some degree, as long as they pull, the only pull long as you only publish to the rapids, the key event bus, then you can pull off to some event. Bus is all you want to, but there’s nobody who says the key event bus doesn’t have something on it. 

Right. 

And the other thing is driving this. Again, I, I credit some of my right colleagues in London about this was that we borrowed this concept from Google that says, don’t worry about who may need the information, collect it anyway. 

That’s, Google’s whole company policy, I think though, isn’t it? Yeah. 

And, and it turned out it served them extremely well. And, and so to some degree, the KACO bus allowed us to do that. As long as we put everything on the same event bus, I can now write myself a little application that pulls from two radically different things. Here’s some weather information, here’s some sales information I can see if the weather and the sales information influences each other. I can write a little service that does that analysis, but I don’t have to go into this. Each of these two groups and negotiate a new API, all the data’s already there for me to try this. Mm. Now I can try things now I’m into experimentation. You wanna organize myself to exper to allow experimentation. Back to your point about how your trouble of getting microservices going, first of all, you need to be able to deploy rapidly Cause deploy, developing things really quickly and then waiting for three months to deploy them is not rapid response. 

Absolutely. 

I think most good clients that I that know have been successful, start out by saying, we’re gotta do continuous deployment. Let’s get that fixed. Let’s rest control from this it group, the deployment, the guys, the dev guys, the ops guys, and make it, make the DevOps happen. Do what it takes to get that, whether it’s, and a lot of that’s gonna be political, some of it’s gonna be technical. Certainly some of it’s gonna be training of training and developers to let them actually do deployments and feel comfortable doing that stuff. 

Yeah. I’ve got, I mean, there’s several questions I could several route. I could go in off this, I guess, but in terms of, actually getting continuous delivery into places that in my experience that’s been one of the hardest things to sell people on. Would you agree, or have you found a better way of telling people how to do that? 

No, I, in general, I, I probably open a client. It doesn’t, hasn’t already reached that. 

Really. Okay. 

Cause they’re not ready for my stuff. I mean, 

If they’re not ready to do that. Yeah. 

Okay. Certain not ready for my microservices version stuff. I, I can do traditional applications and more traditional shops, But not, not, not the microservices. 

And, and so in terms of the develop then, cuz this is another thing I wanna ask you about in terms of getting the developer more familiar with continuous delivery and some of the other aspects, that not just writing in a particular language, knocking out a service, knocking out an application, but understanding how Kafka works, understanding how to debug things when you’re using Kafka, this starts to leaders in the direction of a full stack developer. This something that, how have you seen that evolve over the last 20 years? I guess. 

I’ve always been a fan of, of a programmer who doesn’t put any adjectives in front of their name. 

Sure. So is that just a developer then? 

Just a developer. Yeah. It’s you just don’t wanna be, you wanna hire somebody who says I’m a Java developer or I’m a lead Java developer. I mean the more adjectives you put in front of the more suspicious I get. You wanna be bring people on who, couple of traits. The two traits I’m always looking for when I hire somebody is I want somebody who is a self learner. So show me your getup account. Tell, tell me what, open source you participate in, tell me what languages you’re playing with. Show me that you’re a self learner. Cause we are in a field that requires always learning. The second thing is, show me, you like delivery. Show me that you have a passion for getting code out the door. It’s not just the technology you wanna get, you wanna ship things. Cause you have those two traits. 

I can teach you all the rest. 

Mm. Are we looking for a polyglot here? Someone who knows all the languages or someone who is just happy to pick up a language and do something more, 

More the latter, but just some degree if I have like 60% of my team is poly skilled. I like to call it. Yeah. Somebody could do more than take more than one role in at least six. I can probably optimize my team to whatever the problem is today. 

Yeah. Okay. Going back to like the, the Kafka side of things, obviously we talked about there being several D ways of implementing stuff. The community in microservices solidifying one way of implementing them or not? This still, is this still a bone of contention? 

I think it’s still a great area of experimentation. 

Right. Okay. 

When, when somebody says they’re a microservices expert, I say, you, there is no such thing. It’s just too new of technology to say, we actually have good knowledge of this. Mm. I think we’re getting to the point where we could probably say for a certain class of applications, these techniques work very well. If I go into a day trading atmosphere where you have lots of events streaming in all the time, a bus oriented architecture is probably the right way to go. And you humbling. You get much debate about that, but in a lot of other spaces where we’re getting to move into other fuzzy areas, and again, it’s all the inference engines and AI stuff where it comes back, how all of that stuff fits in. We don’t know yet. 

Just in terms of that in, it will understanding that microservices architecture on a big scale, do you ever introduce any element of workflow to that as well? Or, how do you, because there’s a number of things popping up a, around the internet about how you get microservices coordinating with one another, obviously even with KAF in place, there’s certain companies, introduc element of workflow. What’s your thoughts on that? 

Well, again, nobody has a lot of experience in this space, but so the trend I have seen when I’ve done microservices at that are event best based is little, the little communities that microservices that form. There’s a set community, a little Southern of microservices that work really closely together and then reach some overall conclusion that they pass on to another cluster of microservices. 

Sure. 

That these services kind of cluster very nicely the same way that modules tend to cluster and traditional systems that there’s a natural clustering of these problems that occurs. It’s not that every service know about every other service know more than every other API needs to know about every other function API in the entire system, in a million lines of code. There there’s modules that form. There’s levels of encapsulation of this, that occur. In my world, I like to take a team and say, okay, you’re responsible for this level of functionality and you’re gonna implement this in microservices. It may be 20, maybe 50 microservices, but you guys own these services. You can change any of these services. You want to, whenever you want to, if you want to change the message formats among these services, it’s all your world to do that. One of the things you need to define is what’s that message that you’re gonna tell the rest of the world mm. 

That they’re gonna trigger on. That one you need to sit down and probably fixing concrete a little more, that becomes a little less flexible. 

You’ve got these clusters of services, then is that where the rapids rivers and ponds analogy comes in the, do they cluster around a smaller river or a smaller pond? 

They turn the cluster around rivers. Yeah. Yeah. Ponds tend to be my acknowledgement that it entities are important. Mm. It, isn’t nice to know what your current address is, but that’s pond stuff. And, and by the way, SQL is a great place for it’s great technology for ponds. 

Right. Okay. Basically. 

What you wanna do with a pond is build reports. You SQL will the reports get along really well. 

Yeah. I was gonna ask you where you go with databases actually. What your, what is your a approach to implementing databases within microservices? 

Well, I, I think every micro, since every microservice is completely a world of its own, every microservice, if it needs persistent data get, makes, it gets to make its own choice. Mm. A lot of times are you doing is basically re pulling reference information together. You’re getting a fee from somebody that maybe it’s the frequent renter status, for example. Well, that’s something you’re pulling outta the data warehouse every night. You’re just matching up some employer, customer ID to some level of status. So that’s a little straightforward table. Key value store is all you need. I don’t need a big relational database. Thank you very much. I can make that choice with my little microservice, but I don’t wanna throw a sequel out either because, if my job is to build reports, my database probably should be SQL. Cause that performs extremely well for that. The nice thing is just because I need a transactional database somewhere, doesn’t force everybody to have to use Oracle in all their services. 

It doesn’t have to be one database fits all. And sometimes you get the argument. It says, well, but I wanna be able to back these things up and say, I don’t need a backup for the key value store for frequent renters. I can pull, I can just pull off the data warehouse again. I don’t need you to save my database. Sometimes they don’t, you get this blanket answer. I gotta, I gotta save all my database permanently. Cause that’s the rule. Well, anytime you use the word rule, I’m I get little gleam in my eyes. Cause now here’s something I can change. Here’s. 

Chaos. 

Cause you don’t know the real answer. Obviously you think it’s a rule. That’s not the real reason. 

Well going into those, into those rule side of things, in terms of that process side, you’ve been known for programer anarchy as a process. How tell us about programer anarchy and what does that mean in comparison to agile? 

So, so first of all, program anarchy was basically a marketing term, 

Right? 

Well, they slapped onto our process, just gain attention and it seemed to work really well. What it really reflects is self-organizing teams. That’s what we’re really trying to say is that the team itself knows how to organize itself. It turns out were working in fuzzy domains. Were looking for a high degree of experimentation. We were having this clustering and microservices up approach working for at this time. In fact, I would say that team, that team basically discovered this clustering phenomena themselves, right? So it turned out. I don’t need a manager in a fuzzy environment when the team knows what they’re trying to accomplish at a gross level, some KPI they’re trying to meet, you don’t really need a manager to tell you KPI is cuz it’s well known and we’re measuring it on a daily basis. Because it’s a fuzzy problem, they’re no experts so that your job’s not to tell me the requirements are, cuz you can’t tell me it’s a fuzzy problem. 

You can’t tell me how it’s supposed to work because you don’t know either. That’s the date, a fuzzy problem. You can’t tell me what will guarantee he, somebody will pay the loan back. You can’t guarantee me that this is always gonna detect charge only transactions. There’s no guarantees there. You decide that you don’t, there are no experts and all of a sudden the business analysts die out. When you tell me there’s nothing that you could tell me to manage my time terms of innovation and how fast I should try I do is out manager’s role goes away. You, and if you’re actually developing and pushing code out as fast as we are, you don’t have a dedicated QA team. You’re you’re testing your coach yourself. You’re putting it into very, user friendly, deployment, friendly environments that are self-monitoring. This dedicated in a QA team, this idea that we have to roll this thing out in stages that goes away as well. 

You start eliminating those roles, you wind up with having all this resource that you could apply to programmers. Instead of having, five programmers and five people to make sure the programmers do the right things by feeding the right requirements, I’m in a world where I can have 10 programmers and 10 programmers can run more experiments than five programmers can run. Even if the half the experiments fail you’re way ahead of the game. 

In this instance, you are clustering around problems, I guess, rather than clustering around domains. Is that, is that right? 

Pretty much. Yeah. Yeah. Cause you’re trying to solve something and that’s why I like people who like to deliver solutions. You want bragging rights to go to the people who come in and say, look how many things I sold yesterday? My, with my clever change and your partner is coming in and saying that to show off with their colleagues, you’ve won. 

How, how difficult is it to get an organization to adopt this approach? Cause I, I don’t imagine that you’re going into organizations that already have this approach. This has gotta be something you’ve you’re selling to people. 

Well, it turns out it’s not hard for the developers. Developers will play whatever game you give them. 

Oh yeah. That wasn’t that wasn’t the thing. The people I were trying to, I was trying to, 

Right. So, so it’s not hard for the programmers to get excited about business metrics. If you tell ‘em that’s the game you want them to fly, what are a few cycles of that? All of a sudden you’re showing the business guys what these programmers are coming up with. They’re like, well, and literally I, I had this happen. They’re like, how do they come up with this idea? I’m saying the programmers, the problem solvers, this is what they do for a living is solve, is figuring out how lever algorithms we just told ‘em what your problem is. And now they’re motivated. 

I suppose in that case, in the case that they’ve brought you in to solve this problem, they’ve already created the space to say, look, Fred is gonna come in and try some stuff for us. And that’s one of your solutions. I presume. How would somebody who wants to implement a Fred George solution go about doing that in your opinion? 

Well, I think your first thing is you gotta make sure you can, you have access to the high level executive is paying the bills. Cause you can’t get to see this guy really understand what the motivation is. You’re in trouble. The other thing is you gotta tell him what you’re gonna be doing. That you’re gonna be making a organizational transformation. You’re not talking about agile. You’re not talking about TD. You’re not talking about prayer programming. You’re talking about organizational transformation. You’re trying to get the organization to a point where they can try an idea out aggressively with minimal consequences. If it turns out to be a bad idea, making them more competitive and you’re gonna be creating some, basically some social chaos in this process. Are they okay with this? Because you can’t have that conversation with that executive and get him to nod his head. 

You may succeed in the short term, you will fire be fired in the long term. 

If they’re gonna give you the space to do this, how long in your experience do you need to be able to show enough of a win that you can then carry on. 

Two months, 

Two months? 

Yep. 

That’s a lot shorter than I thought you were gonna say. Well, 

I was gonna say two weeks, cuz most of the time I can get in two weeks, but I give myself good. Look at two, say two months. 

Right? Okay. Okay. 

I can have a significant impact in the team itself within two weeks. 

I suppose that depends on the size of the team that you’re gonna be taking on in that case. 

Well, obviously I will start with a small team cause that’s the first mistake you make is if you start with 50 people and try to some that something’s different, you need to start with four. 

Mm. 

We got these four people on board. You can bring in another four pair with those guys. Basically you can probably basically double the team size every two months. 

Okay. 

Which doesn’t take you long to go through, get to 50 people. 

Cuz that’s another thing as well that, we’ve talked with a few startup owners on this show and we talk about how you scale a team pretty quickly in terms of scaling a startup. That the same approach that you would take? Cuz most of the stuff that we’ve talked about so far is doing this in an enterprise. 

I think you do it with a startup as well. Startups tend to be more aggressive about willing to change and try technology. And, and frankly startups are basically almost always full stack developers. 

Yeah. 

They’re guys who are thinking about the bigger picture and their options and marketing success and how to marketing words and UI and development and backend security. I mean they’re full developers to start with. You’re starting out in a great spot and they’re also willing to pivot. They’re, that’s where you’re trying to figure out which what’s actually gonna make you the money. What, what idea actually does work. They’re all about experimentation to start with the key is to make sure you stay in that experimentation world. One of the things we did at a startup in month, the first one I worked in London is startup. We, I started out, there were 35 people and we had one fulltime manager who was basically doing management fulltime. We grew to 320 people. We had two full-time managers. 

Wow. How long did it take you to get to 300? Many? 

I would say we grew it that to four years from 35 to 320 and four years. 

And was that the right pace? Could you have gone faster? Could, should you have gone slower? 

I think we probably would’ve as fast as we possibly have gone. 

Yeah. 

I think we could’ve, we it’s possibly, we could have gone faster, but were very reluctant to take any outside cow. We were basically spending our own money. 

It more of a capital concern rather than a cause the thing I’m curious about is, 

We were doubling revenue. We were doubling revenue every eight to 12 months. 

Yeah. Yeah. 

We, were printing money, frankly. I mean we had 50 people, we had, we had 50 people, we made 50 million pounds that year. 

Wow. Well we all need to find out what, what we all need to figure out what project is doing that for us. I mean, that’s definitely a goal that everybody has to try and achieve at some point. 

So, so there are a lot of unique things I learned about that environment that played in a program energy. Like I, walking into that environment, I was a very traditional agile thinker. We had, card walls, we had, some role for business analysts, some role for testing. We walking into that environment, walking, watching how successful they were. I realized that a lot of these roles aren’t necessary primarily because were in a fuzzy problem domain. That’s where I saw programming anarchy and working before I actually named it. It, it wasn’t something we decided were gonna build. It’s something we observed. 

I see. It’s something that you I guess you learning as you’re going along, your, you, you have an open mind clearly in the way you approach these things. You, is that where inspect and adapt comes in for you, but on a minute by minute basis, 

Basically. I mean, I’m, I was fine. I tend to be in bleeding edge technology because I know that’s going to, feed into other interesting things as well, whether it be processing, how you approach problems, how you deliver solutions, new architectures, new, new technologies tend to drive that. 

I was gonna ask you about being on the bleeding edge, cuz you are obviously always on the bleeding edge. How often do you hit blockers with the bleeding edge? 

Well, basically I, I gotta make sure I’m reinventing myself at least every five years. 

Okay. 

Cause otherwise I I’ll miss the next train that’s coming along. 

Right. 

And, and to some degree you always kinda looking for what the next train is, for instance, right now I’m not in containers at all. Containers are kind of dead in my work. 

Okay. Where, where should we all be looking Fred? Where are we going now on. 

Serverless? 

Serverless. Okay. So I functions because. 

I don’t, why should I worry about Kubernetes and Docker and all this other stuff, but I just have to put my code out there, figure out how to scale itself. 

Okay. So functions everywhere. That’s where we’re going. 

Yeah. Why not? Okay. I mean, lamb seem to be the right way to go. 

So, when you are talking about, okay, functions is where we should all be heading and we should give up on containers. If how big are we talking in terms of our services now using Lambda, cuz were talking about Java and a couple of hundred lines of code. Where are we now? 

Well, I, I, I would say for me, I’m in the same size, I’m still talking about service rooms in that same size. It’s just, they’re purely functional services now. You, and you wanna pick a bus, Kafka unfortunately is kinda like first generation web technology. You wanna pick a bus that’s kind of has the auto scaling built into it as well. Once you sort of pick those technologies. Yeah. I, I think you build great systems on that basis. The other thing I’m looking at in terms of X five year trains is virtual reality. 

Okay, cool. Well, let’s do virtual reality then tell us more about that. 

So, so to go back, I was part of the generation that went from green screens to go interfaces. I, I was, I, I have sharp that transaction through. Cause I remember arguing at IBM is like, well, what’s wrong with my green screen? I can only do one thing at a time. Why would want multiple windows? You have to say, well look, you could have your email up and your calendar at the same time. You can do some of things like this and then you get into lands. Well, why do I need anything more than a text speed? Because I just putting to exo might screen, well maybe you’d like to have audio on your computer. Maybe you’d like to have video on your computer. I mean, I’m making these things up in 1985. Yeah. Trying to explain why you wanna land. One of the things I observed was the first goos that came out first, people are using windows basically would take the entire screen and you exactly what would be on the green screen. 

Now it’s in color. There’s no concept of having multiple windows. 

Cause why would you want that? 

It took quite a while for people to understand that when you got enough real estate, you want to have multiple windows up. You don’t really wanna. Again, browsers went through the same thing. Every browser screen was like, have a whole window. I can’t have bulk pains in my browser. That would be, that would be wrong. 

Well, I remember Tams being revolutionary. Yeah, 

Exactly. So, so now you go into VR. Now we’re into a technology, that’s three dimensional. The first thing you see is every ball of the VR, the layouts are, here’s a giant green, that’s 360 degrees, but it’s a set of screens. It’s not taking advantage of the motion. It’s taking advantage of the fact that I can travel through this space. It’s not, I should be moving through this space. And I hear something around behind me. I turn my head and there’s my Twitter feed. And something’s happened on Twitter. I’m not taking advantage of the fact that the UI experience associated with Dr. Needs to be 3d reality experience. Now you’ve seen right now is 2d. 

How, how far away do you think we are before this stuff starts to take off? Because I mean, we do a new section of the front of this show and we’ve just, we covered Google killing off Google cardboard. We’ve covered Microsoft ignite conference where they basically repackaged a load of VR stuff and sold it as a different product. It doesn’t seem to be a huge amount of development, except that we’re all getting excited about what apple might do. What direction do you think it’s going to go in and what steps are we gonna see over the next few years? 

Well, I think you’ll see business being, of course the place that can afford the high end headsets and the software development for it. I think you’re going to basic, see, again, people like day traders putting the VR headsets on to look at all the data trends They can look at and again, navigate through that space, looking for what’s going on, hearing noises. Using sound and stuff like that, effectively 3d sound to say, what’s going on. You know, here’s a noise over here. Oh, that’s, at and T is trading like crazy right now, what the hell is going on? 

To become like a heads up display for us to be able to take in more information, I guess. 

Yes, because basically to some degree, we are still, we have 3d vision. We just still live in mostly a 2d world. I mean, we walk on the ground. We don’t fly. 

Yes. 

Yeah. Yes. Unless you’re having to be a, I guess I’ve seen gymnasts being amazingly 3d, more 3d people. Yes. Scuba dive, scuba dives live in a 3d world, but the rest of us are basically 2d worlds with 3d vision, but we aren’t using the 3d vision yet. 

With the onset of the pandemic, everybody’s moved to almost fully the remote working globally. Does this, I was gonna ask you before, do you see VR being a whole of people sat in an office with headsets, but presumably now that it’ll be sat at home, a headset with headsets on, is that a direction we can imagine virtual offices using VR. 

Maybe, but I don’t see the winning app there yet. 

No, 

It may be nice to be able to move from one zoom room to another, without having to do that. I, I try to teach one of my classes using zoom rooms. Cause I like to break people off into pairs and good work and I go help each pair. 

Right. Yeah. 

In a big classroom, I can look around the classroom and I can see who’s struggling and I go help them first in zoom rooms. I can’t see that. 

Yeah. I guess in, if you headed that direction, you could do that with VR. Couldn’t you? 

You could, I, I could see, if not nothing else I could see, who’s actually talking to each other who seems to be stuck. 

Yeah. I think that’s really interesting cuz that gives you the opportunity to do a whole load of things that you would normally you in a physical place, but you’ve actually removed the barriers for the world then really, which I, I was actually watching a little video earlier on about Arthur C. Clark talking about an, a doctor being able to be in London or in Edburg. I think he said specifically in working on a patient in New Zealand, actually the combination of VR and remote working, that’s pretty much on the cusp of reality, I guess. 

Yeah. I think, I think we’re getting there. 

I’ve seen you talk a bit about I T as well, is that still an area that you are exploring and the automated home? 

Yeah. I, and I it’s kind of also, cause I think it has actually interesting application of microservices cause your home is full of events. 

It is you. 

Almost one like this little home event, bus of your own, something that can say here’s the current temperature, here’s what, you’re where the sunlight’s streaming in. Here’s what, here’s, what appliances are on here’s who’s moving around in your house right now. Here’s at your front door. There’s really just a whole chain of events. You want your IOT devices to react to that. It’s a chance to develop a, a, a relative secure IOT event bus in to the house. 

Well, you’re actually describing my house exactly at this point in time, cuz I’ve been on a renovation for the last three years and obviously you’ve inspired me enough, Fred, that I do have exactly that in my house, it runs on a mosquito event bus. It’s exactly, but that’s probably not a conversation specifically for the podcast. That’s something for geeking out later. How far have you taken the, this in your own home and how do you see that being adopted as technology that everybody has? 

Well, basically it’s kind of came to a giant standstill here because it’s September of 2019. I went to Norway to work for a little while and I just got back to this January where, 

Right. Okay. There’s been of a, a gap there. 

It’s been more than a little gap. Yeah. I I’ve had a massive gap in my home project stuff. Now I’m just beginning to scan the technology space to see what’s out there now. Cause 18 months ago it was one thing and now it’s like, whoa, there’s some new stuff out there. I gotta try. 

New toys. 

A lot of new toys to play with. Yeah. My newest tour is my quest headset. I got to quest two headset and so it’s hard to tear myself away from beginning to think about playing with VR rather than play with my I O T. 

I see, 

I, I’m not gonna lack of choice. I also have some new, some grandkids are getting to the age where I could turn ‘em loose with some custom iOS apps I might wanna develop. So. 

Oh, nice. So you’re building your own development team. 

Yeah. It, it actually worked out well, you, my son’s a programmer for ThoughtWorks since it’s been there for like actually longer than I worked for IB. Yeah. He’s been there before that. Wow. It stuck with at least one of them. Maybe I should start the next generation off. Right. Started him off with a PC junior. Yeah. So, 

What they all starting out with these days. If you’re gonna go and teach your kid to program, what should, what should you start out with in your opinion? Well, probably. 

An I iPad with that. Some of that swift building toys on it. 

Mm. Yeah. They’re pretty cool. They’re pretty cool. 

It seems to be a good place to start. Plus you can ruggedize your iPads and it survives kids. 

That’s true. You can get those great big rubber things, can’t you to make sure that you yeah. 

Make it waterproof, snow proof, whatever it is, you can make them industrial. Either the military or children, either one. 

On the IOT side, then what do you see as making IOT in the home ubiquitous like to get to the stage where everybody has it or everybody wants it? What do you think would be the trigger? 

Well, I, I think first of all, it’s gonna be start almost automatically in new home builds. Mm. All the new homes will basically be wired for fiber bill. People will worry about their excess points for their wifi. They’ll all they’ll be built into the, into your walls and stuff. There’s no reason that you wanna put in copper wire where you can use remote switches for everything. It’s actually cheaper not to use copper wire. 

Oh yeah, absolutely. 

Yeah. Have a friend in California who, built a high end house, many years ago, but he didn’t pull any copper in his house. It’s all done with remote switches. I think new houses will be wired this way. And it’s so many new apartment buildings. We, we wire it this way. Mm. I think you’re probably already beginning to see that in, in say Hong Kong and in Singapore and places like that when we’re putting new construction and that’s the way they’re doing it, 

Do you see the retrofit happening? I mean, I’m sat in a house that’s 120 years old. Right. So, do we expect that is the people are people gonna be doing the retrofit? If they are, that’s gonna be quite expensive, cuz it can require new wiring and all sorts of stuff. Like what would be the trigger that makes a customer have to have that? 

Well, I, I don’t think it, I, I think first of all, it’s gonna be a desire first. It’s the 5% it’s gonna be that early adopters. They’re gonna be the nerds like yourself. 

Yes I’m. 

And do it anyway. 

Yeah. 

I, I pulled a cat five cable in my 110 year old house in north, in the states when I had a house like that, It turns out there’s going down in the crawl space beneath it. I saw two generations of telephone wiring technology. As well. Yeah, I think it’s gonna be the nerds doing it, but you can do this almost all wireless thing. Now you can do wireless hubs and wireless switches and basically you just need power outlets that are just put the wireless adopters in there. You’re gonna find that people like yourself are gonna be able to do that in general. It’s gonna, after a while it’s gonna, here’s some installers that will do it for you. You gotta come in, fix your kitchen up, fixing new bathroom up. The other guy will come in and say, well, you wanted me to do this fancy wiring for you. 

Yeah. 

I, I gave my daughter and her husband, this is my daughter who’s nerd. Did this nerd not a tech nerd, put some few lights in their put house so they could turn on some few lights in the baby’s room and around on television and now they’re hooked. 

Yeah. Yeah. 

It’s kinda like once you have a taste of it, you can’t do without it. 

Yeah. I think it’s just getting to that stage where they become properly useful because a lot of these lights where, it’s in the bulb, as soon as you turn it off of the wall, it’s rendered useless. You can’t do anything with it. So, it’s getting to the stage where that’s not the case and it’s still being useful enough. Do you, I mean, you talk, you’ve, you’ve been on the bleeding edge of all of this technology over your career. Do you have any concerns for this personal aspect of this? We’re in a world now where we’ve got all of these smart speakers in our house, I’m sure, I, I mean, I, as a kid growing up in the eighties, I remember, people being terrified that someone might be listening in, on your telephone, on your landline, and now we’ve got, we’ve got like that exact bug is sat on our desks, ? 

That’s obviously changed people’s attitudes, but what, how have you thought about that over the years? 

Well, certainly talking, if you see my Iott present, one of the things I talk about is putting it on its own land. Let’s build my own little wireless network that has all my devices in my house. I’ll put a, a, a GAteway bridge to the outside world from there. I think it’s gonna be important to set up your own little network. 

Yeah, yeah. 

Secure network that way of yourselves. Again, I think once it gets be commodity, that’s what will be by default. These guys are installing your new wireless system, as well as your new kitchen, as well as your new bathroom. These guys will install a secure, a holy, secure system. The thing is frankly, the current generation doesn’t care about privacy is the way you and I may. 

Yeah. Should I, 

No, to some degree I, I was raised, I, I was in the I system in London. Yeah. So just get into London faster. I put myself into the, rental skin system. I didn’t, that didn’t bother me cause I got a benefit from it. I kinda go back to if, if a massage went you dead, you’ll be dead. 

Well, I think this is an interesting thing when it comes to security, because I think, we probably don’t necessarily need to be as concerned as individuals. I mean, the thing is you potentially on the internet, you become a commodity, right. They’re trying to stick you into a cohort so they can sell you more advertising. I have lot of these speakers turned off cuz I’m sick of Facebook recommending me things that I’ve spoken about, ? So, but in terms of them actually coming after me as a person, I probably don’t need to care about that. Would you agree with that? 

Oh, I, I agree. 

Yeah. 

So yeah. I, I tend to opt out of all the, targeted ad stuff myself, but then you have enough. I mean they know everything about me already. 

Well, there’s plenty of, yeah. There’s plenty of talks that everyone knows what you think. I think, 

Oh yeah. I think there’s no hiding that. To some degree, again, I I’m not, I, I kind of believe in the sunshine thing. I tend not, sunshine principle don’t do anything. You don’t want to be shown in a broad day light. So. 

That makes sense. 

I try to live by that. I think the more that we dump privacy is a primary concern. Now, if I was Q and not I’d care about privacy. 

Yes. Well that’s a whole nother topic. 

Yes. So, so I, I, I’m always a little suspicious of people that are maybe trying to be overly private. Suspicious not to say it’s not to say wrong, but it’s a little suspicious when you’re saying what are you really trying to hide? 

Yeah, absolutely. There are a few other things, a few other points that as we’ve gone through this, I’ve made made notes as we’ve gone through. There’s a, I realize we’re jumping back into conversations we’ve already had, but there’s a couple of things I don’t wanna let you go without finding out about, so your approach to testing, you’ve talked about moving away from not having QA and things like that. I think I’ve heard you I’m pretty sure I would be quoting you if I said, I have heard you say that if you can’t write a few lines of code and make it work, then you shouldn’t be writing code in the first place. In terms of testing, how is your approach to that taken off and what is, well, what is your approach to testing now? What do you use any particular tools? Do you, do you still feel like that about QA? 

I, I definitely feel that about QA. Some parts of the world has QA define. India, QA is defined as being basically process wizards. They really are. You really are associated with thinking about processes more than code itself. I grew up in an environment where if you weren’t smart enough to write code, then you were a tester. And that those days are gone. The testing tools themselves or programming tools, whether it’s Cucu or, or mercury, so mercury tools or other stuff like that, or Seleniums the programming oriented tools and generally in systems we’re building now, you can’t really be test the test, the code unless you understand systems architecture overall. Let see you’re an architect, how to write code. What’s the differe between you and a good programmer. The skill distinction is gone. So first of all, you get acknowledge. There’s no skill distinction now between those two roles. 

So now it just becomes a label. I, I haven’t had a QA team and any been known for the last decade. 

Right. 

Now. That’s because I, I tend to be on bleeding edge projects, but it also tends to say that in those environments, you obviously, you’re not gonna have fast response. You don’t, you can’t do a continuous deployment and have a QA team. Now what we discovered in, again, I’ve credited the team in London for discovering this. I just, I just report it. One of the things we discovered is we put monitoring into our event bus. We built our event bus with testing, built into the bus itself. We’re constantly testing our own software on our own event, bus, into live system. It’s not that we’re not doing system integration testing, we’re doing active monitoring of the real system the same way. We’re probably spending just as much money, just as much talented resource as we would in the QA team doing system test. We’re now doing it by deploying active tools and live environments. 

I was gonna ask you about TDD and whether that is a practice that you follow as a developer, but it sounds like you’re also building that into the, the deployments. 

Yeah. It’s two levels of testing. It’s certainly TDD is kind of the unit testing, I believe in TDD for unit testing for lots of reasons. One of which is it requires me to be lazy about how much I write. Cause I can’t over-engineer things under TD rules. 

Was that a strange practice for you to pick up like learning to TDD? Or is that something you’ve always done in a way? 

Well, not always. Cause I, it wasn’t until, I, I saw the Kent Beck and ward cutting and briefing about this stuff. One of the oops list even before thep book was out that it was like, write the test. I, I was actually in an environment were writing in small talk mostly, and then people write the code and then I’d go home that night and I’d review that code and write some tests for it. I was testing it the evening and evening. I write test for the code, my guys read during the day. They said, write the test before you write the code. It was like, you idiot. Of course. 

It was an obvious that it was an obvious thing for you as soon. It. 

Was obvious once you having written small talk for like, oh, almost 10 years before agile came along, agile is basically just saying, this is the practice you should use small talk. It was once you, once you’ve been a small talk program or everything you said was like obvious. It was like, why I didn’t see it before is just, I’m an idiot. That was the, that’s the beauty of guys, like, ward Cunningham and Ken bank, they see it and they recognize it and they can talk about it. So, big kudos to people who realize that stuff. And, and they guys like Dave S Snowden says they’re fuzzy problems and their traditional problems. Mm they’re like, oh, and by the way, you organize yourself differently for those two types of problems. It’s like, as soon as he says this, it’s obvious. 

Yeah, absolutely. You just, you have to have the first person that makes the connection. 

Dave Snowden, for example, when he said that, I realized why I could not put a so right call center. Cause there’s nothing about a call center. That’s not, that’s anywhere traditional or complicated. It’s a simple problem. You don’t need agile for simple problems. Cause the requirements are not changing. 

Makes sense. Makes sense. We, we’ve not really talked about frontend too much. Are you, is micro front end something you’ve explored or is that the wrong direction for us to be heading in? 

I haven’t played with it, obviously with my UI background. Yes. I’m kind a, I’m kinda a front end designer from the ancient days from the rich client world. It turns out all that translated really well to the web world. It looks like is that play with so far, all that translates right. Is a mobile world. So, so the gooey paradigm is pretty much the same with all those devices and all those technologies. I would like to see my I’d like to have an event bus on my iPhone. 

Nice. I. 

Would, I would like to turn my iPhone experience into a set of UI things that are driven from an event bus. 

What would be the benefit of having that, do you think. 

Experimentation and being able to do micro deployments to that environment? Mm. Imagine, I think about my watch, my apple watch has got this really amazing set of I guess they call ‘em notions or it’s kind of they’re UI elements. Right. I can put, I have like one, two, I have seven different UI elements on my apple watch and I find ‘em all very useful. It’d be nice to be, take my iPad and put, 50 of my own personal little things on there that things I care about. Yeah. 

I can do that. 

If I had an event bus and write little micro apps and just do exactly what I want. 

Yeah. I get it. That makes sense. Maybe that’s another direction for the next five years then? 

Well, I think it’s taking some of the technology. We know works really well, event buss and moving into other platforms. 

I must admit, from having worked in microservices for a good five years now that does feel like this. That’s the thing that’s missing on the front end. You know, we don’t quite have that. Same. It feel, everything feels distinctly monolithic when you’re working on microservices as the back end. 

Yeah. I mean the front end guys, they’ve struggled to get into agile. Obviously one of that is they rooted in print media. A lot of ‘em have been trained in print media, which is extremely static and they were trained to, put your 45 levels of grade scale in there and make it look really good. They haven’t been trained interacting with that. The interaction with that, as you read, read your copy, your eye is drawn to this point, you read that copy. Mm it’s not about clicking and dragging. That other stuff is not part of their education. Hence the UX guys have it now emerge people who are thinking more about the experience, but may not be able to draw the prettiest icons. Now you wanna center your group around the UX guys, not the UI guys. 

Yeah. 

Now you think about the experience from that perspective. Even those guys are, want to almost understand they wanna have all the requirements they’ve been trained. Give me all the requirements. I’ll come up with a great experience and we’ll build the product. That’s wonderful. You have to tell them, I don’t need, you don’t need to tell me exactly all the flows of stuff. Just gimme a hint, give a starting point. Let’s try it out and see how it looks. Let me play with it. What I really wanna do is I wanna take these UI guys. I wanna train them how to build, use my UI building tools. I want them to build their own HTML. I want them to build their own CSS, by the way, when they get hooked into that, I want them to build their own Java script. And now what’s happened. Now they’re programmers, And now they understand how to play with things and try things out. 

Now they begin to understand how I can be granular with this. My first step, when I run into UI people is first of all, wanna sit them in the team they’re working with. Mm. Stop working, sit together with the UI, the L UI guys spread yourself out. This comes from basically one of the principles I talk about, which I picked up in grad school. And that’s three reasons. People talk to each other. One is we say, I have the same hobbies. Kids go to the, in daycare. We both are our soul supporters, but frankly, as a management team, you don’t have any control over that. The second reason people talk to each other, they have the same manager. Managers tend to be people who pull people together and require the conversations. The third reason, this is the guy got tear MIT for this, the chance that you and I community various inversely with a square of distances between our chairs. 

If you and I are sitting close to each other, we’re talking and we double the distance. It’s one chance. Four of us talking, you doubled again, as one chance of 16, he measured the intellectual distance of a staircase. It’s a hundred meters square. That number you might will be in Bangalor if he’s downstairs. 

Yeah. I’ve, I’ve experienced that. 

Exactly. I don’t want the UI people sitting. I want them sitting with the team they’re working with. Cause I want communication between them and the team. I know they’re gonna talk to the other UI guys cause it’s a UI manager. One of the first things I do in terms of senior arrangements is I change the seat arrangements and organizations to create the square distance between the chairs, minimize that for the people who I want to talk. I want the, I want the business requirements guy. I want the man team manager. I want the customer, I want the UI guy and the developers front in the backend, all city as close as possibly other. 

I suppose that’s in the, that’s that same model that Spotify defined as well. Isn’t it of having that slice of a team. 

Yeah. And all comes from the same guy. This is Tom Allen from MIT who basically recognized this. 

When was this? Cause I’m curious as the origins of this model. 

Yeah. Basically again, a guy’s name is Tom Allen. He’s first emeritus. Now at MIT, the Sloan school at MIT he’s basically came up with the square of distances between the chairs. So, and he, he, he measured, he measured that over many studies and he has the concept of, gatekeepers, people who information flow through, which has nothing to do with your title. 

What, what year are we talking here? Fred? 

I, I had the course from him in 1986. 

Right. 

His, his book came out. He had a book came out shortly thereafter about that timeframe, which you could still get hold of. 

Well, we’ll put that on the recommended reading list on the show notes then. I mean like that’s the origin of the Spotify model right there, I guess. 

Yeah. I mean, he’s the one that talked about that. Obviously it was being done cause he actually studied it. I mean, people had indications of this, but he’s the one that kind of formalized it and called it out. And, and I used his techniques almost immediately after walking outta school and every time I tried it worked amazing. 

Wow. One of the last things I want to cover with you just because I’ve heard you say this. I wanna know your, your feelings on it in more detail. I’ve heard you say 80% of microservices implementations are wrong. So why is that? 

I, I, I think people say I’m gonna do microservices cause it’s cool. They immediately, they draw boundaries around mostly entities. So they’re like, here’s my customer service. Here’s my order service. Here’s my inventory service. These aren’t these aren’t functional services. They’re data services. 

Yeah. 

So they’re not making any decisions. They’re actually just data. You’re still having some other service. That’s pulling information from these services and having to make the decisions. So you haven’t really accomplished. 

Anything. That’s just service oriented design really isn’t it that’s domain design to break things up like that. I guess. 

You’re breaking up, but you’re breaking on boundaries that are basically entity based rather than functional base, which you’re I’d rather do is break along functional lines. Imagine that you had an inventory and a customer database, it’s the join tables that have the interesting behavior, cuz that’s where orders come in. That’s where returns associated with that. It’s it’s these joint tables that have the interest behavior. You wanna build your services around the join tables. Having access to these two tables, you build your services around functions that manipulate these two tables and they tend to be functioning in that case. So again, most people build microservices. First of all, they’re gonna be thousands and thousand lines of codes gonna take a team of five people to who must build it. Obviously it’s not micro in any sense of the word. Mm. But they like the term. It sounds efficient. 

Yeah. 

Yeah. So they call it micro. You, so you kill that part of it. You have people doing that, calling microservices, aren’t micro. You have the guys who are breaking on entity outreach, which are in gonna be wrong on and it winds up having no benefit in the long run. All you’re trying to do is joins across databases now. 

Yeah. Yeah. 

Joints work really well inside databases, not outside databases and that’s kind of why. You have people that are taking traditional problems and trying to solve with microservices. They have gone through the learning curve necessary to be efficient and to do that, to handle the, the pinball being lost or pinball going around in circles forever. 

That brings us back to the fuzzy conversation about before. 

That says, yeah, I don’t try to bite this off unless you got a fuzzy problem. If you have a fuzzy problem, it’s worth learning it. You haven’t got a fuzzy problem. Now’s not the time to get into microservices. 

Okay. I mean, thanks for your time, Fred, it’s been a pleasure to talk to you. I’ve been excited about speaking to you for quite some time. Thank you for taking this time early morning in Las Vegas to join us. 

Well, thank you. Thank you for the opportunity.

hello and welcome to that tech show with me Chris Adams and my co host Sam Gregory this week's episode is a big one as we have the grandfather of microservices Fred George joining is if you don't know what microservices are then this is a great chance to learn but in simple terms that how big tech firms building software these days it's a move away from monolithic applications to a suite of much much smaller applications working together in order to form an answer so we're talking thousands of small applications of a hundred lines of code all last instead of one application of millions of lines of code and favors many Piney this way of engineering with the mail online in the early two thousands and so of course we'll be talking about microservices how they came to be what they are when to use them and perhaps more interestingly when not to use them since starting his career in the seventies working for IBM it is always being on the bleeding edge and so we'll ask him about what that's like and how he's maintained that position at the forefront of technology for over forty years along the way will also take in the U. the things he's invented besides microservices and we'll talk about what he thinks is next in the future of technology so without further ado is Fred George asiasat not bad not bad yeah my second coronavirus shot Wednesday so school hi time few more days but yes it's like now since our freedom and safety I bet this is getting an influx of visitors like crazy right now is it cold cold winter yeah we're just seems going to fish like a pastor from thirty five percent but looks like people are ignoring that just going along so wow do you spend a little time in the casino shredded you avoid them I I actually saw one time and then when I don't gamble really I just walked through the just observing people yeah as it turns out biggest reminds me when I lived in London they all talk of all things so it's the people watching is amazing in both places you have amazing theater we we got we got six standing circus soleil shows banks my heart goes all the performers come through in the restaurants are just unbelievable I mean Gordon Ramsey a six restaurants alone so if you want entertainment if you want great food what you should be able to see the other I think CS is my favorite times that's when about two hundred thousand you show up yeah it looks like it looks like we're still looking probably for a week and then but I think it's the black cat guys coming in same time they don't video warts come out you can tell the difference between the two one with the old town transforms when these events source makes candidacy but we'd like...

hello and welcome to that tech show with me Chris Adams and my co host Sam Gregory this week's episode is a big one as we have the grandfather of microservices Fred George joining is if you don't know what microservices are then this is a great chance to learn but in simple terms that how big tech firms building software these days it's a move away from monolithic applications to a suite of much much smaller applications working together in order to form an answer so we're talking thousands of small applications of a hundred lines of code all last instead of one application of millions of lines of code and favors many Piney this way of engineering with the mail online in the early two thousands and so of course we'll be talking about microservices how they came to be what they are when to use them and perhaps more interestingly when not to use them since starting his career in the seventies working for IBM it is always being on the bleeding edge and so we'll ask him about what that's like and how he's maintained that position at the forefront of technology for over forty years along the way will also take in the U. the things he's invented besides microservices and we'll talk about what he thinks is next in the future of technology so without further ado is Fred George asiasat not bad not bad yeah my second coronavirus shot Wednesday so school hi time few more days but yes it's like now since our freedom and safety I bet this is getting an influx of visitors like crazy right now is it cold cold winter yeah we're just seems going to fish like a pastor from thirty five percent but looks like people are ignoring that just going along so wow do you spend a little time in the casino shredded you avoid them I I actually saw one time and then when I don't gamble really I just walked through the just observing people yeah as it turns out biggest reminds me when I lived in London they all talk of all things so it's the people watching is amazing in both places you have amazing theater we we got we got six standing circus soleil shows banks my heart goes all the performers come through in the restaurants are just unbelievable I mean Gordon Ramsey a six restaurants alone so if you want entertainment if you want great food what you should be able to see the other I think CS is my favorite times that's when about two hundred thousand you show up yeah it looks like it looks like we're still looking probably for a week and then but I think it's the black cat guys coming in same time they don't video warts come out you can tell the difference between the two one with the old town transforms when these events source makes candidacy but we'd like to start off with them what's your name where do you come from and then we'll go from there if that's possible friend George programmer where I come from east coast the United States originally as far as born and raised and but I wrote my first code in nineteen sixty eight hello so we're not gonna we're gonna skip most of the resume here we know you had a long history with IBM and I think that's how you got started really well probably what you got a lot of year experience I imagine yeah I was seven seventeen years IBM which was which actually really formative experience for some of our work is development site in building the operating systems so extremely bright people great mentors great training it is a side effect on you know laters I learned how to work with big companies would be affected the company is it someone all mortgage companies really bright despite what we think but you have to understand what motivates the company so I can learn how to do that IBM well it was a bit of a shock when I start working for somebody who wasn't as it turned out it I'm not everybody's idea yeah I had the same experience when I left isn't actually it was again you know you you move out of the company what you call it a comfort blanket levels of things and then you realize that the places don't have those things yeah I mean it was very strong decision culture it was easy to get decisions by our wound up in a consultancy where situation where's consensus culture and the positive thing was there when everybody agrees it's it's kind of easy to get things done turns out there reading reading the wrong thing it's almost impossible to get change I ran across your website on the wrong thing twelve well so all my decision culture stuff didn't work and that consensus did it take you a long time to sort of rebuild glass ability to to make decisions when you don't have necessarily the political line do you need well I pretty much type baits I'm always gonna get my way someone and so bring Taipei you're pretty agile agile about how you're going to do this so we get this trick doesn't work how bout the stretch and I'll bet district so yeah I've I've heard you referred to was a hand grenade full of that what I was yeah that was a direct quote from let's see I see do you know I work for and the Daily Mail I was described as a hand grenade throwing into the vault which is a good drivers refuse you get it all because it there are no rules for hand grenades yeah I remember I remember reading a group meeting you briefly I guess probably about six years ago and this was it's a project in the was in London and you'd been brought in to kind of shake things up a little bit by making making all of the developers and everybody around them probably feel a little bit sick that they didn't really know what else they were doing with that career is and and then I was in a position as a consultant where we would still pick the pieces up after the the hundred going off some to some degree that was engagement where I was born in explicitly to do that I mean I was in my my job was to be destructive so I was like okay I can I can do that that's not hard specialty it's and so basically you change everything you change where people say you changed to talk with us they have to change the programming language you use you change our titles so you do everything was sort of upset the apple cart and then have a reform yeah as the alternate was just fired will group group with them and start over again somewhere else in the whole thing off shore for example so how did how did you know from from sort of leaving IBM and leaves losing that decision culture and having that way of working how did you get from from that to being the hand grenades is there a whistle stop version of how you got from those two to between those two points I even IBM I was a bit of a hand grenade well I I it is usually because even though you can I be able to be a relatively state organization that their culture was we believe in our wild ducks we like our well so I kind of was always kind of in the in the wild die Kategorien in my I think my my managers appreciate it so they always throw me into a situation that was basically broken and help me fix it but once they begin to realize this after fix things well the thing is kind of fun so I'll break it again just for fun so they moved me to another job it was a priority problem and listen leave behind something that was something like you have to clean up after the well it is much easier clean up after me after all right everybody make the changes necessary so for example that is I was I'm sort of the grandfather of the clips together well so I basically brought objects into my team and and IBM and we we we started doing things more strange way relatively strange we got reprogramming built up a nice structures sort of find the organization now in power the managers and then they moved me out because everything is working fine in my process is one of your finished up and close out the door wow okay so I mean I know use the grandfather of microservices so there is a there's a there's a list of all the products before the I like eclipse well even that was not you also if you if you're running on a windows system you see will drop down to have consistent spelling will read actually corner so my work back then yeah well you know if you ever heard the net bios interface this and that but we associate with that that's something use mark I did when I was working on lands it was brandy wow so if you go what what else can we add to the list in front well you know I I own the user interface architecture in Microsoft and IBM for awhile yeah right we were over there was two days of put out the first telecommunications software the lead to computers talk to each other so I could log on one computer and run an app on a different computer well the first computer networking with some of my earlier projects also will severe we talking about when when we're talking about you know getting we're kind of at the seventies when the networking stuff land is kind of like the mid eighties how does that compare to like the stuff that you doing today because obviously you know on a daily basis to speaking of microservices particularly with spending of docker containers here there and everywhere and getting them to talk to one another and actually frankly you know who that can be a pain in the US as well so I mean how how is that how how how how things differ from you doing that sort of work on a daily basis today too when you were doing that for the first time in the seventies well I think you're right I think overall I'm always looking for that edge technology so I'm kind of the first guy to try this is let's see if I can make this thing do something interesting so that was the case of lands I didn't invent the concept of what merry network but I put out the door yeah sure the first land versions of stuff I didn't I didn't necessarily agree interfaces but I. products IBM yeah well how complex was it working in the system I mean how long did the self projects take because we see we took not mention the docker containers you know you can do that in a couple of hours on your machine you know a few minutes to be honest depending on how you good you are at it how long was that taking you in the in the seventies and eighties well I think everything always took longer than I wanted to but I trust I technology transformation is a twenty five year process yeah and so I'm I'm in that first five years where there's very few doctors worked on in that early stage you know the Braves use old story so for example and you're not selling to everybody not an arm you're selling to a very specific route we got crowd that fears that technology is and it should so you're getting into objects getting doctor your primary language is you're probably working with a lot of the Wall Street which were extremely competitive about having the best software to figure out the best reasons to do everything so you'll find those believe yes chants and know that technology is competitive advantage and it is not still the case mostly because I don't necessarily know if that would be the case in the city in London are is definitely the case in Wall Street you'll see a more the day trips yeah so day trading is were you just not through traditional banking no moving money around like that the day traders are always you know looking for that edge all those the ones where you see the new languages yeah does the things come out you'll see you'll see the guys that have the six giant screens all these other guys that do that work well we will see the same big banks will see that some of the private equity firms well the ones are investing their own money which many rules but you know you always want to find these guys with your technologies and yeah I want to see our clients that's why working people who figure is not a matter of should I know if I'm a hotel chain should buy more bacon for breakfast where should I use the technique by some technology those guys are looking for the cost benefit trade off they don't there are going to bet on anything they're too conservative looking for the guys down but we got a new hotel chain I want people to walk through and not even stop with the registration desk with them going right to a room that's why tourists wiped their credit card you walk in the door are you listen that's what our clients and and how how close are we to them or is that just a friend Georgia no we're we have people that can do that really mother motel or hotel change to do that you don't you just kinda get a text message on your phone when you're getting close though tell us says oh yeah your your room number you walk in going the elevator use your RFID card to walk in and see you point the you mentioned before about surfing fight you know your in working within that first five years is this technology that we expect to be commonplace in five years or at least something to break through in the next five years you know you're in some degree in already the question is what does it get to be like twenty percent that's another the five to fifteen year time frame yeah we got to be twenty twenty five years that's when everybody's going to do or how to get fired you can find a CIO today who doesn't claim they do actually yes also if you don't so you do I do you can get higher and what actually goes back to nineteen ninety five nineteen ninety seven Sir Ken twenty five years in there where are we now yeah fair enough I mean I think we we've been we've come across one or two and still not necessarily in that bracket but there's always going to be a few dinosaurs I guess the again and he ordered a few dinosaurs if you guys still for your buggy whips are going to be a big industry so one thing I want to ask you is because I've seen I've seen a lot of your talks and then the you tend to talk a little bit about what what you talk a lot about microservices obviously and we'll get to that and in a bit more detail you will see it took a lot of talk a lot about and processes methodologies and you know having stuff to do more to full moved into my child and then this the some things beyond that I'll let you talk about but I'm what what do you what is your time tend to split what you're most passionate about is it the technology or is in the process and how those things work together for you I I think I'm most passionate about getting caught out the door so I'm looking for a girl walking to a client trying to figure out what's what's going to keep me from getting the car out the door sometimes it's going to be process sometimes it's going to be technology sometimes it's just going to be business at its I was a business treat treat them through the system the program is a giant spreadsheet but they can so you kinda looking I'm looking for what's going to keep me from being successful is it because of my sort of broad background for away from bottom management IBM advanced training their business degree as well as obviously some technology background and I I don't know run a run organization of over two hundred programmers and IBM so run large organizations so I'm walking in the store with no biases about what needs to be done more than I do what I can do and there's lots of possibilities and so on how do you how do you decide where you gonna go first how do you find what what the bottleneck is is usually with them a couple of few minutes actually using a few hours of short with the client and the station we always want to kind of start with the person's spending the money sure out what are they trying to accomplish because what the I. team manager says and what the sponsor says sometimes radically different and so is it always and is always one thing or is it always a combination of those three that you have to change in an organization to make your and yours will be touching all three it's just a matter of to what extent you need to hammer on all three when I walked into some ideations and Norway turns out the talent level was fabrics that keep the people they had were motivated they were sh very strong technically they knew how to work together all the collaboration stuff no I didn't do any work at that was already done so my focus was on process which we had not so I I put the process in place there already we using Kotlin is our language that was really good but they were using Colin is an object using Colin this call okay so first of all you have to teach them what what Colin is adopting language looks like snow will opt obviously it turned out it was a very traditional applications so we do need microservices and they were like we should be back for services no no you don't for her judicial application are you have a set of rules are trying to silence nothing fuzzy about this problem I was gonna ask you about that because that will the latest talks I've seen you do is on I'm don't do microservices old that why you don't use microservices and oversee the man who's known as being the grandfather of microservices could you tell us a little bit more on on the on the show particularly I'm why you don't use microservices in certain circumstances well I guess my favorite model is the Canaveral from when they started which is a model of how you classify problems all right you have to sort of look at problems inside where there were traditional know the words it's the rules associate with what the interest like this is the way transaction be subtle this is the way bank balances are calculated this is how salaries this is how you get back our coach tax rates those are very traditional applications verses one then they'll call fuzzy which are problems that don't have precise answers like which stocks should I buy or should I loan you money this is a fraudulent transaction these are fuzzy problems so it turns out for for all the problems you have competitive advantage in that space if you're able to try ideas out across the the faster I could try something out the better off animal often enough for the environment so if I'm sorry if I was appalled I want to be able to have a loosely coupled systems because I want to be able to write it but something out there that doesn't necessarily impact anything else I've done just to try and idea I don't want to have to sort of schedule would think about it make sure it doesn't impact anybody else go do a three month stands for six months the answer to get it out the door I want to try something in minutes it again you sort of look at various environments like day traders just as their world there are giant for the problem and they have to pay the vantage of they can pull things out all the time not sort of way for the big I. T. guys to say now's not the time for you ship not this is it's in those fuzzy problems that's the that's the place to use microservices yeah because my resources right now is our hardest thank you know a loosely coupled system when you've been used to doing step one step two step three step four now it's about let's do all four steps aren't we can't get our heads around that because we have been trying to think that way it's not how we get business requirements FARC former customer he says this is how you do it here's a step three to use our checklist it's very laid out you have to be a fuzzy problem that checklist is pretty much useless he doesn't and the business guy doesn't understand that he doesn't understand he's on a fuzzy problem is looking for precise answers but then realize you wanna be frozen but micro services hard to get your head around it's a developer so I recommend don't do that unless your experience doing it already I can't solve traditional problems with micro service that's very doable but you don't really do that unless you have experience in micro service because there is quite a lot of sense of work to get that working as an organization having built a few microservices architectures myself he's in a big learning curve for developers you've not done it before yeah huge and I I completely unrest in that region I really and this is from from the mail online or from before that but well before that be my first microservice project was in two thousand five right when I was doing some work for when the big manufacturing firms and US and and basically I was had to play with the concept of services since it began since I haven't China hung with some of the some of that great ventures in that space we were gonna always ask the question well how because the servers and I was like okay well here's a chance to play when we make a service really small inside I did that and we had a very successful project we had a price that was estimated to take eighteen months we delivered in nine weeks wow nice fast comparison to the the the racial estimate yeah and and so you know it's still out and one that was wildly successful when it when I rolled off the project it kind of stopped if we start building bigger bigger sars you stop focusing on building small service is easier just no bigger ones yeah so this time what we're what you thousand five what language are we talking in and how big we took in the services we're we're writing Java it we're writing services are probably a couple months ago yeah okay this is what I call a pinball architect we were basically trying to fill in the information so we had all of this your record information need to collect and so you can figure out what piece of information to get to other piece but I didn't have one of those other features for your to get that piece I had to get the other information somewhere else so the bill will services that would take two piece information generate from Sir and so we basically took this long message we ping pong around services trying to fill in those missing some pieces somebody who help you out well he could help you out he steps somebody who could help you out and so what was the inspiration for this was actually pinball machine well no but it was by is kind of the best metaphor for when you got yeah yeah I think it was a matter of just trying to break things down into very small classes so rob and object right I wish I was I was trained in objects from about nineteen eighty six nineteen eighty seven so we we got to this number I was trying to small talk so roll about very very tiny classes you're single responsibility class so I put that into my mind microservices let's let's make every service to one thing if you describe a service you use the word hands must return services regular part I see that makes sense so in terms of how you are getting the services to talk to one another there is there's this is obviously the several ways to implement services and get the services to talk to one of the right you know you can have a five and A. P. I. you can have a fire events and you kind of I. service discovery now I know that this is being quite a point of contention in terms of what is the best microservices architecture I don't know which camp you fall into but like you to explain that to us Tootie was on the cold and wet when did you start out and did you start out with that straight away and you know has you'll have your theory about how you do this changed over time in my theory has changed I think we go back to the to the I'm all my pinball architecture every every service you which are the service to go to to get yeah this piece of information without calling every service how to register is being the service and knows how to to get you to this piece of information and it turned out that you know that created you know obviously it's all the problem very nicely it was a lot of other things about our project with extremely well and again I I had I had absolutely stellar team and the guy who sort of created the the concept of a bill process well okay I I have probably one of the finest programs ever work with each re joined me on the cheap stuff it was not your average team this was this was the US team superstars with that being said so we did run across some props like I was saying we drop a message should we drop this fall into bed always she anyone come out sometimes it's just running around around in circles because we got an infinite loop the minute relicensing the ball we just kind of disappear and so we begin to worry about about how do you detect these other situations so one of problems about loosely coupled system is when things go wrong don't have a clue about what went wrong so I've been searching for other architectures that help stressed that it was it was in London working with the group this is the early days of Kafka showing and so my guys got into the cockpit bus not I'm looking that event bus and I'm like wow this is kind of cool because it's a permanent history of everything that went wrong if you ever heard history reading what always something goes wrong I can go back to reconstruct it and in fact to some degree you think about an event for us well the traditional system I'm trying to debug a problem and they got lots of components are working various ways the first thing I do is reconstruct the timeline of what went wrong she pulled launching here reporting live very trying to rebuild that timeline of what works you're going on a bus you have the time line already that works done yeah absolutely especially if you use an event buses I recommend that you have put everything on the event I can rabbits to segregate it because that means you have to put it together again the timeline together skip that step whatever the on the same event bus because there are the little statement must be by means of drink how does you could because you have a concept of a sea of using several of them bosses so how how does that work if the account if you call that approach of putting everything on the main event bliss how does that work with the additional doses the slow moving but it is always a point the only the only the only publish the rapid acumen bus and you pull up to some of them also I want to but there's nobody who says the key event bus doesn't have something on it right the other thing is driving this again I I. credit so my my colleagues London about was that we we are this concept global this is don't worry about who may need the information collected anyway it will help a whole company policy I think that isn't it and it turned out it sort them extremely well and so to some degree the Kafka bustle out as soon as you put everything on the same event plus I can now write myself will application it pulls into radically different things here some weather information here are some sales information I can see if the weather in the sales information for each other I will service it doesn't analysis why don't you go to this use these two groups in the ocean you might die all the data is already there for me to try this so now I can try things now I'm into experimentation so you want to organize myself to expect allow experimentation so what about your point about height your trouble getting microservices your first why you need to be able to deploy rapidly does your deployed developing things really quickly and then waiting for three months and for them is not rapid response absolutely so I think most good clients that have been successful start out by saying we're going to do continuous deployment let's get that fixed so let's rest control from this I. T. group the deployment guys a deaf guys that the ops guys and and make it make the DevOps so do what it takes to get that whether it's a law that's gonna be political some of us could be technical it's only someone is going to be a training of trainers and developers to let them actually feel comfortable doing that sort of stuff yes uncle among the several questions I could get several routes I could go in awfully office I guess but in terms of you know actually getting continuous delivery into places that in my experience has been one of the hardest things to sell people and would you agree or have you found a better way of telling people how to do that no I in general I I played one of the client doesn't hasn't already reached really okay they're not ready for my stuff for a similar to that yeah I'm sorry for my microservices first stuff I can do traditional applications the more traditional shops but not not not not the micro service and and so in terms of the the developer then because this is in the another thing I want to ask you about in terms of getting the developer more familiar with continuous delivery and some of the other aspects you know that not just writing in a particular language knocking out service knocking on application but when the stunning how Casco works understanding how to debug things when you using Kafka and this starts to leaders in the direction of a full stack developer is this some think that's and how have you seen that evolve over the last twenty years I guess I've always been a fan of of of of a programmer who doesn't put any actions from their neck sure it is not just a developer them just as well yeah yeah you know it's just a little bit yeah you are somebody who says I'm a job or I'm a I'm a lead Java developer I mean where I choose to put in front of more specials like a so you want to be able to bring people on board who approved that couple traits so the two traits I'm always looking for when I hire somebody else I want somebody who is a self learner so show me your github account jumps how we what you open source you participate in I'll tell you what languages you're playing with so show me that yourself or is it we're in a field that requires always learning was saying that you show me you like delivery show me that you have a passion for getting code out the door it's not just the technology you will get your ship thanks you have those two traits I can teach you all the rest so are we looking for a polyglot head someone who knows all languages of someone he was just happy to pick up a language and do something more more the latter but some degree if I have like sixty percent of my team is Ali skills like to call it yeah might be more than think more than one role at least I can probably optimized but you know whatever the problem is today yeah okay I'm so doing but it's like the and the the calf good side of things obviously we talked about the being several different ways all of of implementing stuff is the community a microservices and solidifying all in one way is implementing them or not is this still is this still a bone of contention I think it's still a great area of experimentation right okay when when somebody says their microservices expert I see you you're just lying thank it's just to newer technology to say we actually have we actually embryologists well I think we're getting to the point where we could probably say for certain class of applications these sort of techniques work very well if I go into a day trading atmosphere where you have lots of events streaming and all the time up a bus worried architectures probably the right way to go we get much debate about that when a lot of other spaces were re getting should move into other fuzzy areas and again a sort of a local the inference engines on A. I. stuff or comes back well how old is still fits in we don't know yet so it just in terms of that in it when the standing that microservices architecture on a on a big scale do you ever into Jeez so any element of work flow to that as well all you know how to use it because this is this a number of things popping up around the internet how you get microservices coordinating with one another obviously even with Catherine place the system companies introducing an element of work flow what's your thoughts on that well again nobody has a lot of experience in this space so the trend I've seen when I'm done micro services like that are sort of been spaced is billable billable communities of microservices are for this other community a little sun micro service work really closely together and then reach some overall conclusion the passenger another cluster micro service sure the service is kind of cluster very nice you'll see way that bottles to the cluster traditional systems there's a natural clustering of these problems are a curse so it's not that every service needs no but every other servers no more than every other every other A. P. I need to know about every other function API in our system in a million lines of code yeah this modules at four news in this sort of levels of the capitalization of the circuit court so in my world I like to take the team say okay you're responsible for this level of functionality and you can if you're gonna put this a microservices maybe twenty maybe fifty microservices do you guys all these services are you can change in the service you want to whenever you want so if you want to change the message formats among the services that's all the it's all your world to do that well one of the G. the finest what's that message you can tell the rest of the world about they're gonna trigger on that we need to sit down and probably fix fixing concrete or more that becomes a little less flexible and so when you call these clusters of services that is that when the the rapids rivers and ponds analogy comes in that you know to the coast around a small a reversible upon US around rivers yes yeah no ponds posted to be my my knowledge but then it ends sees aren't part it is nice to know what your current address this but that's Pufnstuf oh my we seek was a great place for all its great technology for parts right okay really what you would do with the pond is bill reports in sequel reports get along really really well yes I was going to ask you where where you go with databases actually in what you'll what is your approach to implementing databases within microservices well I I think every every Microsoft since every microservices completed a world of its own every microservice seventies persisted data which suggests making some choice in a lot of times are you doing is basically reference only reference information so you're getting a fee from somebody that maybe yes the Africa renter status for example well some people you know the data warehouse every night it's so you just matching up some sort of employer at your customer ID to some level status that's a little strange for table key value stores all you need I don't need big relational database thank you very much so I can make that choice with my little micro service but I don't foresee colonia because younger if my job is to build reports my database partially sequel those that performs extremely well for that the nice thing is just because I need a transactional database somewhere the force everybody to have to use oracle and other services it doesn't be one database that's all so if you get the arguments as well but I want to be able to back these things up save I don't need a backup for the device to work for free for renters I can pull I just pull up to your house again I don't need you to save my database so sometimes they don't you know you get this blanket answer I got it I got it I guess they will be database from because that's the rule well anytime you will use the word role I'm I get I get looked green in my eyes now here's something I could change your system because you don't know the real answer I mean do you think is a rule that's not the real reason well I've got you going into those into those rules side of things and in terms of that process side you've been you've been known for program anarchy as a as a process and how it tells about programmer on occasion what what does that mean in comparison to AJ I'll so so first operator in Ricky was basically a marketing term rights only slapped onto our process just sort of getting attention and didn't work really well well but what it really reflects is self organizing team that's what we're really trying to say is that the team itself does not organize itself so it turns out we were working in fuzzy to me and so we were looking for a high degree of experimentation we were having this clustering a microservices approach or working force at this time hello in fact I would say that that T. nineteen basically discovered this clustering phenomenon themselves right so it turned out I don't need a manager in in a fuzzy environment with the team knows what they're trying to accomplish is gross level some KPI to trying to beat you don't really need a managed to sort of tell you what the case may I ask is this well now we're majoring in on a daily basis well because the fuzzy problem no experts so do you your job not some of the requirements are because you can't tell me this was the problem you can't tell me how it's supposed to work because you don't know either date your policy problem you can't tell me what will guarantee somebody will pay the loan back you can guarantee me that this is always going to check for doing transactions there's no guarantees there so when you decide that you don't know no experts and also the business name was tie dye we we tell me there's there's nothing that you could tell me to manage my time better terms of innovation and how fast I should try ideas out badgers role goes away so use it and if you're actually developing in pushing code as fast we are you don't have a dedicated Q. A. T. you're you're just you're out yourself you're putting into it it's a very user friendly deployment friendly environments that are self monitoring this is dedicated to your team's idea that we have to roll the stand out in stages I goes away as well so we started to use those roles you wind up having all this resource that you can apply to programmers so instead of having you know five programmers and five you will make sure the program is do the right things my feeling right requirements I'm in a world where I could have ten programmers Marcus can run more experiments on five programs can run you have to experiment fail you're way ahead of the game so in this instance you'll clustering around problems I guess rather than clustering around the mains is not is that right pretty much yeah if you're trying to solve and that's why I like people who like to deliver solutions you are bragging rights to go to the people who come and say look how many things I sold yesterday I'm out with my clever change your partner is far coming in is saying that you show off with their colleagues you want hello how difficult is it to to get an organization to adopt this approach because I'd like to imagine that you're going to organizations already have this approach this is going to be something if you're selling to people well it turns out it's not hard for developers developers will play whatever game you get yes yeah I thought that wasn't that wasn't the thing that people are trying to I'm trying to convince so it's not hard for the programs you get excited about business metrics you can tell that's the game you want mine where a few cycles of that also you showing the business ties with these programs are coming up with their like well it literally I've had this happen like honey come of this idea the problems are just what they do for a living this song is very not clever algorithms we just talk about your promise another motivate and so I suppose in that case in in the case that they brought you in to solve this problem and they've already created the space to say look Fred is going to come in and try some stuff forest is now and that's one of your solutions I appreciate how would somebody who wants to implement a friend your solution go about doing that in your in your opinion well I think your your first thing is you gotta make sure you can you have access to sort of the high level executives paying the bills if you can't get to see this guy really understand what the motivation is you're in trouble really is you got to tell him what you're going to be doing that you're gonna be making eight organizational transformation not talking about agile you're not talking about DVD you not talk about their program you're talking about relationship estimation try to get the organization's a point where they can try an idea out of rest of the with minimal consequences if it turns out to be a bad idea making them more competitive you're going to be a preview some basically some social chaos in this process are they okay with this because you can't have a conversation with that executive get him not as head you may succeed in the short term you will fired beget fired a long term and if they're going to give you the space to do this how how long in your experience do you need to be able to show you at enough of a window you can then carry on two months two months yep that's a low show someone thought you were gonna say well I was gonna say two weeks because I most of time I can get you to weeks but I get myself logged in since I took my rights okay okay I can have a significant impact in the team itself within two weeks I suppose that depends on the size of the team the you're going to be taking on and I see all start with a small team that's the first mistake you make is you start with fifty people I try to convince them that something's different yes all four we got these four people on board you can bring another for parents of stocks basically you can probably you're basically double the team size every two months okay which doesn't take you long to go through it to fifty people yes because that's another thing as well that you know we we've talked with a few starts opponents on the show and and you know we talk about how you scale the team pretty quickly and in terms of scaling a startup is not the same approach that you would take because most of the stuff that you know we've talked about so far is doing this in an enterprise I think it will start as well well sort of stream to be more aggressive about really changing try technology and and frankly for start ups are based almost always full stack developers yeah the guys who think about the bigger picture and yet their options I am you know Marge's success and how to marketing words and you why and development and back end J. repeat me the full site developers pharmacy you're starting out a great spot and they're also willing to have it so there are that's where you're trying to figure out which what's actually going to make you that much but what idea actually does work so they're all about experimentation star with the key is to make sure you stay in that experimentation world that's one of things we did a start up in London the first one I worked in a London start up we saw I saw another thirty five people yeah we had one full time manager was basically doing management full time well we grew to three hundred twenty people we had two full time it while and how long how long did it take you to get to three engine twenty well I would say we agree with that that's four years from forty five to three on twenty four years and was at the right pace could you have gone faster could you should you've gone slow I think we probably went as fast as we possibly talk yeah I think we couldn't we as far as we could go fast we were very reluctant to take any outside capital so we were we were basically spending our own money so is that more of a capital concern rather than a because it is the thing I'm I'm curious about doubling revenue we were doubling barbecue every eight to twelve months yes yes we we are printing money Frank I mean we people we are we would get fifty people make fifteen pounds while well we'll need to find out what what what need to figure out what project is doing not for is I mean that's definitely a goal everybody has to try to G. B. as in boy I'm so there are things I learned about that environment that brought played in a program or like you know what that environment well I was a very traditional agile thinker we had young you know card walls we had you know some role business analyst role protesting we walking Harmon walking watching how successful they were I realized a lot of these rules are necessary primarily because we were in a fuzzy problem to me so that's where I saw partner energy working before actually named it wasn't something we did we decide we're going to bill that's something we have to start I see see something useful to them I guess use of learning as you're going along you'll you you have an open mind clearly in the way you approach these things so your ram is not wet inspected the depths of comes in for you but on a on a minute by minute basis basically I mean I'm I was why I tend to be and bleeding edge technology because I know that some of your feet into other other interesting things as well will this be processing are you how you approach problems I deliver solutions you architects do cute you technologies tend to drive that so it's going to ask you about being on the bleeding edge because you obviously always on the bleeding edge so how often do you hit block is with the bleeding edge basically I I gotta make sure read reinventing myself at least every five years okay otherwise I'll miss the next train we all rise it is such a great you always look at what the next train is Frances yeah right now I'm I'm not eight containers of all genders are kind of getting my work okay so where where where where should we will be looking for and what are we gonna start us serverless a case I'm from I don't I why should I worry about what communities and doctor and all this other stuff but I just have to put my code out there let's figure out how to scale itself okay so functions everywhere that's what we got yeah why not okay would be low Landis themes seem to be the right way to go and so you know when you're talking about okay functions is widely used where we still be heading and we should give up on containers if how how big are we talking in terms of our services now using lambda because it we were talking about John over a couple hundred lines of code where are we now well I I I would say that for me I'm in the same spot I'm still going well so talk about sluggish nesting sites you're really functional services now if you want to take a bus you know it's kind of good fortune is kind of like first generation web technology are you a bigger buses coming his as the almost killing built into it as well what you sort of pick those technologies yeah I think you you build great systems on that basis we'll be looking at in terms of sort of X. five your trains is is reality okay cool let's do virtual reality then tell us more about that so so the sort of go back I was I was part of the generation of wind from green screens to go interface so I was I I have structure for that transaction transaction well there are you know IBM is like well what's wrong my green screen you know I can I can only do one thing at a time what what what windows yes I will let you can have your email opt in your calendar at same time you know you could do some things like this then you're good to land what do we need anything more than a text be because I just dropping text on my screen well maybe alike have audio on your computer maybe like have video on your computer I mean these things up in nineteen eighty five yes right exactly why you want to land so we'll reserve Jr was the first movie that came out first you know people using windows basically we take the entire screen and show you exactly what we on the green screen now it's a color because of the heavy mobile windows because why would you want quite a while for people to understand that when you got enough real estate you want to have windows you really want to know can browse with the same thing every browser screen was like I can have a little window I can handle pains in my browser that would be that would be wrong well when the time has been revolutionary exactly so so now you go into VR the number into a technologist read image the first thing you see is a real ball to be our layouts or here's a giant screen less ram sixty degrees but it's a seven screen is not taking advantage of the motion is that take advantage of the fact I can travel through the space it's not I should be moving through the space not hear something around behind me I turned my head to nearest might wait Twitter feed and something's happening on Twitter I'm not taking advantage of the fact that the U. I experienced so simply aren't used to be three D. reality experience now you see right now is twenty so how how far away do you think we are before the stuff starts to take off because I mean you know we do a new section of the phone to the show and we've just recently covered Google killing off Google cobbled we've covered that Microsoft ignite conference where they basically re package delivery all stuff and sold as a different product I'm the doesn't seem to be a huge amount of development except that we're all getting a little bit excited about what apple might do so what what time action do you think it's going to go in and what steps are we going to see over the next few years well I think you'll see business being of course the place they can afford the high and high end headsets software development for so how are you going to base sixty again people like day traders but in VR headsets all look at all the data trends they can look at him against navigate through that space what's going on hearing noises so using sound and stuff like that actively three D. sound sort of say what's going up yeah here's a noise of your little that's yo AT and T. is trading like crazy right now what the hell's going on she's become like a heads up display for us to be able to take in more information I guess yes good because basically disagree with we're still three weeks we have three D. vision we just almost totally mostly to the world we walk on the ground we don't fly yes yeah yes so so unless you have to be a I guess I've seen Jim just being amazingly basically three deep more street people yes when I would never live in a three D. world but the rest of us are crazy to the world of three D. vision well we we are using three D. vision yeah so with the onset of the the pandemic everybody's moved to almost fully remote working globally does this and I was gonna ask you before you know do you see V. are being a whole lot of people sat in an office with headsets but presumably now that'll be set the home had set with headsets on is not is not direction we can imagine virtual offices using VR well maybe but I don't I don't see the winning out there yeah it may be nice to be able to move from one room to another without having to do that I I try to teach one of my classes using a zero I like to break people often appears to work and I go help each pair Brian it big classroom I could look around the classroom activities see who's struggling and I go help them first soon as I can see that yep I guess even if you had that direction you could do that with the Albany your cut I could I could sort sort of seat out yeah it's not nothing else I can see who's using actually talk to each other who seems to be stuck yes I think that's really interesting because I give you the opportunity to do a whole lot of things that you would normally do in a physical place so when you actually remove the barriers for the for the world that rarely which I was actually watching a little video earlier on about the C. Clark talking about an adult to being able to be in London or Edinburgh think he said specifically working on a patient in New Zealand and actually the combination of VR remote working that's pretty much on the cusp of reality I guess yeah I think we're I think we're getting there and I've seen you talk a bit about IIoT is well is that still an area of the your exploring and the the automated home yeah I I I it was kind of also because I think it has to be especially interesting application of microservices because your home is full of events it is the only one like this loophole with bit parts of Europe well something that I can sort of say here's its current temperature here's what you're up with sunlight streaming in yes well is what sciences around here's who's moving around your house right now here's at your front door really it's just a whole chain of events well he which I O. T. devices to react so it's kind of like just sort of all relative security sort of I. O. T. event busted inside of the house of what you're actually describing my house exactly at this point in time because I I've been on a renovation for the last three years always even spine you know Fred the I do have exactly that in my house and it runs on a mosquito event boasts so it's exactly it but that's probably not a conversation specifically for the podcast that something for geeking out later how far have you taken the the this in your in your own home and how do you see that being adopted this technology and the everybody has well basically this is Kerry Katona giants stand still here because in September of twenty nineteen I went out to Norway to work for a little while and and I just got back this January right okay so there's been a little bit of a gap that Furthermore the little gasket so I could have been a massive gap by and my sort of a whole project stuff hello and I'm just beginning to stand the technology space to see what's out there now because you know it was one thing and now it's like wool this this new stuff out there I got try the toys what is the point where yeah it my my newest where's my my quest had step so I got the question had stepped down so well it is it's hard to tear myself away from begin to think about playing with B. R. rather plant my I. T. so I see I'm not gonna lack of choice I also have some you know some grandkids are getting to the age where I could turn loose with some custom iOS apps I might want to double so nice so you're you're building your own development team yeah it actually worked out well he might my son's a four gram football works been there for like actually we're gonna work for IBM is there or that wow so its stock release one of them so maybe I should start with next generation all right sorry for him all with a PC junior yes okay so they are starting out what they'll starting out with these days and if you if you're gonna go and teach your kids program what what should what should you start out with in your opinion probably an iPad with with that some of us were building twice on it yeah that's pretty cool the particle this seems to be a good place to start because you can you're right I sure I passed my survives gets that's true you can get those great big rubber things going to make sure the water proof snow proof whatever it is you can make them industrial so use the military were children you know wants to own only I. eighty sudden what what you see is making our tea in the home you be Christmas like to get to the stage where everybody has it or everybody wants it what do you think would be the trigger well I think first was going to be still be I'd start almost automatically new home builds all the new homes well I'm basically be wired for fiber bill you people worry about their access points for the wifi they'll all they'll be able to do that so your walls is the reason is you want to put in copper wire reviews remote switch for everything it's actually cheaper not to use copper wire yeah absolutely yeah I have a friend friend in California you know built a high end house many years ago but did you pull any coppers houses all the remote remote switches but I think I think you houses would be wired this way starting your current bill is required to sway I think you're probably already begin to see that in in the in say Hong Kong and Singapore place like that will bring you destruction that's where they're doing it you see the retrofit happening I mean I'm signed that houses a hundred twenty years old right so you know it do we do expect to that that is the sort of people are people going to be doing the retrofit and if they all that's gonna be quite expensive because it could require new wiring also stuff like what would be the trigger that makes the customer have to have that well I don't think it I think the first was going to be a desire for five percent it's not me that early adopters gonna be the nurse like yourself yes or no thanks I read how often do it anyway yeah I pulled a cat five cable in my hundred and ten year old house in North Vietnam the the states when I have a house like that well it turns out this is going down in the in in the crawl space beneath that I saw two generations of telephone wire technology as well so yeah I think it's going to be the nurse doing it but you can do is almost all wireless because wireless Hobson wireless sites which is spacey is the power outlets that are just a couple while some doctors so what we're gonna find the people like yourself we're going to be able to do that in general well it is good that after awhile it's going to be here some installers that will do it for you so we have got to come in for your kitchen up in your bathroom up illegal come in say what you want me to do this fancy wiring for you yeah I gave my my daughter and her husband this my daughter who's who's nourishment I did just are not a tech nerd but somebody in there put house should turn on your lights in the baby's room and around the television another hot hello you must have a taste of it do you ship do without it it's just getting to that stage where they become properly useful because a lot of these lights but you know it's in the bold scenes you turn off the wall it's rendered useless you can't do anything with it so now it's getting to the stage where that's not the case and it's still being useful enough but do you I mean you took you know you give he's been on the bleeding edge of all of this technology over your career do you have any concerns feel sort of this personal security aspect of this you know we're in a we're in a world now where we've got all these smart speakers in our house you know I'm sure you know I I mean I got as sick as a kid growing up in the eighties I remember right and you know people being terrified that someone might be listening in on your telephone in the land line you know and now we've got we've we've got like that the exact bogus sites on our desks you know I'm so that's obviously changed people's attitudes and what we how have you thought about that over the over the years well you know certainly are you talking if you see my I. UT presentation one of these I talk about the screen on some land let's build my own little wireless network that has all my my device in my house and I'll put a I get gateway bridge to the outside world from there so I think it's gonna be important set up your mobile network yeah yeah your network that way of yourselves I think once he gets to be commodity that's what that's what would be by default these guys are installing your new wireless system as well as your new kitchen as well as your bathroom these guys will install a secure holy yes you always get your system Delhi is frankly the current generation doesn't care about crime scenes where you and I you and I may yeah do it should well I don't know how I was raised you know I I was in the Irish system and loved it you know just get get get into London Bastrop over something the young runners rental scams system okay I didn't I didn't bother me because I got a benefit from it I can go back to this year it's a missile would you do that you'll be dead well I think this is an interesting thing when it comes to security because I think you know we probably don't necessarily need to be as concerned as individuals I mean the thing is you use initially on the internet you become a commodity right so they're trying to stick you into a co host of the console you more advertising and so I have the the loss of the speakers turned off because I'm sick of Facebook recommending the things that I've spoken about you know and so but in terms of them actually coming after me as a person I probably don't need to care about that would you agree with that I I agree yeah so yeah I I tend to opt out of all the yachts targeted and stuff myself but they have never paid me they know everything about me already well there's plenty of yeah there's plenty of talks but everyone knows what you think I think yeah are you still no hiding that disagree again I I'm I'm not I'm I can't believe in the sunshine thank god to the yacht central principle don't do anything you don't want to be shown in broad daylight so make sense well I try to live by that I think the more that we sort of don't privacy is is a primary concern well you are not a cure our privacy yes well that's a whole nother topic yes so so I I'm I'm always a little suspicious of people that are maybe trying to be a really private well suspicious now say something wrong but it's a little suspicious when you say what we're what we're really trying to hide yeah absolutely and there are a few other things if you with the points that as we've gone through this I've I've made made notes as we've gone through in this if I realize we don't jumping back into conversations we've already had but there's a couple of things I don't want to let you go without without finding out about so your approach to to testing and you know you've you've talked about moving away from not having Q. a and and things like that and I think I've heard you and I'm I'm pretty sure I would be closing if I said I'm I've heard you say this if you call writes a few lines of code to make it work then you should be writing code in the first place I'm so in terms of in terms of testing how is your approach to that taken off and what is fully issue approach to testing now what do you use any particular tools and do you do you still feel about about Q. a I I definitely feel that about Q. eight some parts of the world has Q. is defined differently in India Q. is defined as being basic process was they really are you really are associated with thinking about process where the code itself well I grew up in an environment where if you were inspired to write code then you were a tester and that those that those days are gone because the testing tools themselves are programming tools what is to come or or or mercury store in Berkeley jewels or other stuff like that Libyans if the programming the programming oriented tools I'll enjoy Joni insisted we're going now you can't really be tested test account unless you understand systems architectural role let me see your architect you know how to write code so what's different you want a good programmer this skill distinction is gone so first why getting knowledge is no skill distinction now between between those two roles so now now it just becomes a label so I I haven't had a Q. eighteen any any project in all for the last decade Brian this a little bit because I tend to be on bleeding edge projects but it also tends to say that in those environments you notice you're not going to have fast response you don't you can't do continuous deployment and have a Q. eighty now what will we discover London again I've created the team in London for discovering this I just I just reported what we discovered is we put monitoring into our event bus so we brought her to bed bugs with testing built into the bus itself so we're constantly testing our own software on our own event bus in a live system so it's not that we're not doing system integrate system integration testing we're doing active monitoring of the real system the same way and so we basically just as much money I just as much talent resource as we will in the queue a team doing system tests we're now doing it by deploying active tools and a lot of our and so I was going to ask you about TDD and whether that's a practice that you follow as a developer but it sounds like you're also building that into that the the deployments yes hello testing strongly tedious of the unit testing of I believe a DVD for unit testing for lots of reasons one of which is it it requires me to be lazy about how much I write I can't over engineered things under TV roles what was strange is for you to pay for you to pick up like learning to TDD or is that something you've always done in a way well not always because I wasn't until you know I I saw the ticket back in marketing a briefing about this stuff and it would be useless before the XP box out that it was like all right it's just far I was actually in that environment we were writing in small talk mostly in the people write the code and then I go home that night I I'm sure you like could rise to test for so I was just to get the evening and the merit test will come by guys return today they said right the test we write the code it was like oh you idiot source so it was an obvious thing for me is what you having read small talk for like almost ten years before agile came along agile is basically just saying this is the practice use usual spot talk and it was what's what's your what's your business motto program everything he said was like objects it's like why I didn't see before it's just I mean eighty but that was that that's the beauty of guys like you know we're cutting him can back they see it and they they recognize that they can talk about all so you know big kudos to people who realize that stuff and and they got guys like Dave Snowden says they're fuzzy problems in their traditional like and by the way you organize yourself differently for this year's second problems so I guess he says is it's obvious yeah absolutely you just don't you you have to have the first person to make that connection they started for example he said that I realized why I could not put agile it's not my call center this is nothing about the call center this is not this is anywhere traditional work comp or comp it's a simple problem you don't need actual for some problems this requires are not changing makes sense makes sense we we've not really talked about and from tend to much M. A. U. it is micro frontend something you've explored is that the wrong direction first beheading well I have a play with it yeah obviously with my U. I. background yes I'm a Californian designer from each day from the rich client world and turns out all that translate really well to the web it looks like it's a little bit of players so far all that translates right is a mobile work so it should be going paradigm is pretty much the same on all those devices and all those technologies I would like to see my I like to have it a bit bus on my iPhone nice I would I would like to turn my iPhone experience into a set of you why things that are driven from an event for us what would be the benefit of having that you think experimentation and be able to do micro deployments to that of our so eight thank you I think about my wife we have a well I just got this really amazing set of sort of I guess they call mom notions or it's kind of the ideal you I'll find I could put I have a lake what do I have someone different you I elements on my watch and I final worries be nice to build take my iPad input your fifty of my own personal little things on there that things I care about yeah I can do that bus and and rival biker apps just do exactly what I want yeah I get it all makes sense so maybe that's another direction for the next five years and well I think it's taking some of the technology we know works really well a bit buses in moving into the other platforms yeah because I must admit you know from having worked in microservices for a good five years now that does feel like this that's the thing that's missing on the on the front end you know we we don't quite have that same if feel everything feels distinctly monolithic when you're working on microservices is the back end yeah I mean the funny guys they've struggled to get into agile what is the ruse in print media a lot of work been trying to print media which is extremely static they were trying to you know put your forty five levels of great skill in there and make it look really good but they haven't been trained in interacting with that interaction with and you read it read it re read your copy your eyes drawn to this when you read that copy not about clicking and dragging and then those sums up on their education so since the U. S. guys a bit now remarks people who are thinking more about the experience but it may not be able to draw the prettiest icons so you want to center your group around the U. S. guys not the U. I got yeah and so now you know you think about the experience in that respect but even those guys are wanted almost understand they want to have all the required to betray give me all the requirements come up with a great experience and will build the product that's wonderful so you have to tell them I don't need I don't need you don't need to tell me exactly the all the flows stop just give me I can't be a starting point let's try it out see how it looks let me play with it what I really want to do as a way to take these you are guys I want to train them how to build use my U. I. building tools I want them to build their own H. T. M. L. I. want them build their own CSS by the way when we get hooked on that I wouldn't build their own Java script and now it's happened now they're programmers now we understand how to play with it and try things out now you begin to understand how I could be granular that's so my my first step when I run into you rent do you white people as a personal one assistant in the team they're working with so some workers said together with you while you work I spread yourself out this comes from basically won the principles I talk about which I have picked up in grad school and that's after three reasons people talk to each other what is we sent same hobbies your kids because same day care we both arsenal supporters but frankly as a measure to you don't have a control over that some reason people talk to each other they have the same manager managers tend to be people who sort pull people together are the conversations the third reason this is the guy got teary eyed my T. for this the chance you're not communicating is inversely with the square of distances between are chairs so you are sitting close to each other we're talking we double the distance is what transforms talk the double again is one sixty the measure the intellectual just so staircase sana meters where the number you might as well be in Bangor he's downstairs yeah our experience the backlight so I know what the U. I. people seem to get out of sync with the team they're working with his own communication between them and the team I know you talked the other your guys because you are a manager so far the first things I do in terms of seating arrangements like change the seating arrangements in organizations to create the script history the church minimize that for the people I want to talk I want to I want the business requirements guy with me the manager I want the customer I want you I dive into developers front in the back in all cities closest boss which I suppose that Senate dot that's not same sort of model of Spotify defined as well isn't it having that slice of the team it all comes from the same guys Tom Allen from my T. basically recognized when issues because I'm curious is the origins of this model yeah so basically is again a guy's name is Tom Allen first American now MIT Sloan school it might be he's the one who basically came up with the square of the distance shares so it any any actual dental you measure that where many many studies because of the your gatekeepers you know people who information flow through which has nothing to with your title what what here we took in a friend well I I had the course from him in nineteen eighty six right in his book came out your book came out shortly thereafter about that time frame which you still get hold up I will pull down the recommended reading list on the night on the show notes then so I mean like that's the origin of the Spotify model dry that I guess yeah I mean he's the one that sort of sort of talk about that obviously was being done because you actually started to meet people and indications of this but he's someone I kind of formalized at all in Iraq call it out in the end I use this technique almost immediately after walking out of school and every time I try to work based on so one of the last things I want to and I want to cover with you just because I'm I've heard you say this and so I want to I want to know your your feelings on it and we'll detail I heard you say eighty percent of microservices implementations are wrong so why is that well I I I think people say I'm gonna do microservices from the school and so the mailing big draw boundaries around mostly so that I can use my customer service here's my order service use my inventory service these are these are functional services or data starts yeah they're not making any decisions they're actually just data you still having something other service this point Rachel from the services I would make the decision Sam really accomplish anything that's just sort of service oriented design really is nice to Maine designed to test break things apart I guess you're breaking up you but you're bringing on batteries that are basically entity based Rohingya functional based on which you might try to do is bring along functional lines so imagine that you have an inventory and a customer database is the joint tables that have interesting behavior that's where orders come in that's where returns associated with that this is the joint tables that have the interesting behavior what about your services around the joint so having access to these two tables you go to services around functions that make like these two tables and they tend to be function that kicks so I get mostly people build certain microservices first while they're going to be thousands of thousand lines of code is going to take a team of five people two months ago so obviously it's not micro immediate sense of the word but they like to terminate it sounds sounds efficient yeah we call it micro so you too you kill that part of it so you have people doing that hello my cursor is not right they have the guys to break it to you which which are all it was a happy no benefit in the long run are you trying to do is joins across databases yeah and always worked really well inside database is not outside and that's kind of why and then you have people that taking traditional problems controls all of microservice and so they have gone through the learning curve necessary to be attrition and how to do that handle the you know the pinball being lost or pinball going around in circles for ever and that means back to the fuzzy conversation but before the full that says yeah if I don't try to Microsoft which young pose a problem you ever fully promise we're learning you have got to fill the problem now is not the time to get in my car serviced thank me thanks for your time Fred it's been a pleasure to Tokyo been excited about speaking to you for quite some time so when thank you for taking this time early morning on the Las Vegas to join us thank you thank you for the opportunity