ASF 035: Alex Yates interview (part 2)

Introduction

Audio version

Don’t you have time to read? You can listen to this as a podcast! Wherever you are, whatever you use. Just use the player directly from this site (above), find it on Spreaker, Apple Podcasts, Spotify (new!) or simply download MP3. Enjoy!

Transcript

KN: And how do you prepare yourself for a speech, for talking and leading the technical session?

AY: How do I prepare for doing a tech session? I drink a lot of apple juice, which sounds silly.

KN: Does that help?

AY: Yeah. It’s probably rubbish pseudoscience but somebody told me at a course that apple juice has a fairly unique quality. They said that the pH level is exactly like what it should be in your throat. I’m not sure I believe that but I’ve never looked into it in that much detail and I’ve never wanted to because it works and I don’t want to pull the empress clothes off it. It seems to work for me. Apple juice has this unique effect that it really, really does soothe your throat compared to water, which just kind of dries it out. So yeah, I drink a lot of apple juice. And if I’m presenting from home, especially if I’m doing a long presentation, like a pre-con or like a full-day training session, and that really, really does hammer my voice. Often, after I’ve done a tour or a three-day training class, I do end up losing my voice at the end of it. So on those days, I will have like an entire carton of apple juice per day as I’m delivering the session, if it’s like a full day thing. So apple juice, absolutely.

KN: So that’s from your voice perspective.

AY: If I’m nervous, I think of the butterflies flying in formation, which is an old acting trick that I got taught back when I was an actor in theatre and stuff. Those butterflies are there to help you in your fight and flight. They’re not there to distract you. They are there for an evolutionary purpose to help you perform better, not worse. All you need to do is learn how to handle them. If you get butterflies and it makes you feel skittish and it makes you feel sick, think about them, control them, get them to fly in formation and that gives you a fire in your belly, and it becomes a force you can use. So yeah, that’s one of my favourite tips. But in terms of the more practical stuff, I will often spend a couple of hours rehearsing in my room beforehand, and I will have done dry runs. I try to always record myself at least once before I deliver it live. Everybody hates watching themselves back. I hate watching myself back. I hate the sound of my own voice, just like everybody else does. But I think it’s a really useful thing to do because you are your own worst critic, and so when you play it back, it helps you to see “oh, that’s annoying when I do that, I should try and do that a bit more quickly, I’m waffling here”. I do tend to waffle, I’ve probably waffled a lot in this podcast and I’m sorry for that, but when you’ve got a kind of scripted thing, it allows you to focus on “right, OK, I really need to get to the point quicker here”. Or actually, there’s a key piece of information that I missed here and it made it hard to follow the narrative. So watching yourself back is important. But also from plain presentation style type things it is good. You can see that I’m always looking down or I’ve got this weird habit that I do with my hands and it’s off-putting, or I tend to mumble or whatever it is. You just see kind of plain presentation stage presence-y type things will pop out at you a lot more. The caveat I’d say to that one is, if you’re the sort of person that is very, very critical of yourself and if that’s likely to do more harm to your nerves than it is to give you the confidence, then don’t do it, because nerves are a real thing and they’re important. But if, like me, that helps you to spur you on and helps you do better, then I absolutely recommend watching yourself back. What else do I do? I’ve started always using the PowerPoint subtitles now. The audio caption stuff. People always ask me “how do you do that?”. It’s literally a button in PowerPoint. You go to… I can’t remember the name of the tab in PowerPoint anymore but it’s like the presentation tab. There’s like a check box that you click, and it will start giving you subtitles for everything that you’re saying. There’s an option “Do you want to appear at the top or the bottom?”. You put the top, because if you’re in a big crowded room, then people might not be able to see the bottom of your screen but everybody can see the top. And on your PowerPoint template, what I do is I put a big orange box right where the subtitles are going to be, and I just put the title of the presentation in that box on every single slide. And what that means is that I don’t end up having the subtitles covering up important information on my slides, because whenever I create a new slide it’s like “right, I can’t put anything in this box because that’s where the subtitles are going to be”. So it forces me to design my slides with my subtitles in mind.

KN: Interesting.

MS: You’ve just described all of the technical skills. So for the technical skills you have books, conferences, but what about the soft skills. You mentioned at the very beginning about mentoring, so how do you learn that?

AY: Soft skills. I mentioned a few soft skill tips. I probably mentioned more soft skill tips than tech skill tips so far. I think the best thing you can possibly do if you do struggle with nerves is to find somebody to be an ally or a coach or a mentor. Somebody to speak to about it, somebody to encourage you, somebody who you know has got your back, somebody who can give you feedback, somebody who you can receive feedback from in a healthy way, rather than getting frustrated or angry about it. You can actually say “right, OK, well, thank you for the feedback, I’ll try and do something about that”. Somebody that’s gonna make you feel good about yourself. The best thing you can do is find that ally or mentor or coach who can help you to be the best version of yourself and to have the confidence. Because it is a nerve-wracking thing. That’s one of the reasons I set up speaking mentors, as a way of helping people to find the person who can give them the support they need. And what I actually find is that for most people it’s that confidence boost. So my best piece of advice is find somebody, probably somebody local who goes to a lot of the same conferences and user groups that you do, because nothing beats that.

KN: And what about your work-life balance? Having a small baby boy and working as a consultant, that must be very difficult.

AY: So when Emma got pregnant, I deliberately changed the business model of DLM Consultants to give me a much more reliable income. Because beforehand, Emma had the reliable income that we knew came in the same every month, and my income was… Emma paid for the important things and I paid for the nice things. And obviously, when we had the baby, and Emma takes a year out of work, I need to pay for the important things, which means that the risk profile changed somewhat. And also it means that I really don’t want to be going away from home as much. It now has been two years since I took a flight for a piece of paid work. That’s crazy. If you told me two years ago that I wouldn’t take another flight for a piece of paid work in the next two years, I would not have believed you for a second. My first two years at DLM consultants I barely earned a penny without getting on a plane. So yeah, to have primarily remote contracts, to focus more on longer-term mentoring contracts as opposed to short-term “let me help you build a pipeline” type things. When I started doing those, my work got a lot more predictable and I was able to do it from home. And as it happens, as time would tell, it also set me up much better to be defensible from COVID. So as COVID has happened, I’ve been able to more or less carry on work as normal. The main difference has been the lack of child care, meaning I need to reduce my hours. But I’m able to do that because I’m able to work from home and pick my own hours, which means that I’ve got much more flexibility about how I handle my work/life responsibilities.

KN: So it’s particularly hard to start your own business, especially at the beginning, first one or two years.

AY: Absolutely. We were in a very fortunate position that when I started the company, we didn’t have any children and Emma had a good and reliable job, which meant that based on the savings we had, we could survive without me earning or with me earning a nominal amount, without making any significant changes to our lifestyle. Basically, it meant that we couldn’t put the same amount of money that we were doing into savings and we can kind of save houses and holidays and all those sorts of things. We had to stop that but our day-to-day lifestyle we were able to not change that too much. And we would have been able to fight like that for quite a long time, because we were in a very fortunate position. Had that not been the case, I definitely would have struggled to start. It would have been a much, much scarier thing to take that leap had we not been in that position.

MS: What hints would you give to young people who wanted to start working on the market. So they just graduated and they would like to start on the IT market.

AY: Make things, do something. Even if it’s not a work thing, get involved in a pet project, get involved in some open-source project, build stuff for fun, build a chess programme, build a poker programme, just make stuff and learn about how to make stuff and what goes wrong when it does. Because you’re probably going to be interviewing for junior developer roles or maybe that’s not the type of career you want but you’re going to be applying for some sort of junior IT role. So in that interview, you want to be able to talk about the things that you’ve made, to make stuff. One of the things that I made while I was working for Redgate, while I was learning stuff, my brother is a software developer and we worked together to try and create… It was an algorithm to play tic-tac-toe, kind of noughts and crosses. But the way that we created it is by kind of evolving it by simulating evolution to try and get it to grow itself. So we basically came up with a random generator that would randomly create the logic, various different logic… We basically came up with a way to encode the logic as a random, long binary string, and then we just generated random binary strings, we made them play against each other, and we left it running for a while until eventually it came out with a tic-tac-toe bot that was quite good at playing tic-tac-toe. So that was one of the things that I did as a fun little side project when I was working for Redgate, and it was a lot of fun. So make stuff. And if you like it, wonderful. If you enjoy it, wonderful. You’ve probably found the right sort of career. If you really don’t enjoy that, then maybe a career in IT isn’t for you. If you don’t enjoy picking the projects that you want to build, then maybe you’re not going to enjoy it when somebody else asks you to build something that you’re not particularly passionate about. So yeah, just make stuff, make stuff, make stuff. Make lots of stuff, and if you enjoy it, great, and it’ll give you something to talk about in your interviews, and if you don’t enjoy it, maybe try something else.

KN: That’s a very good hint. And you are co-organizer of Data Relay so I hope all folks from the UK are aware of that free event. But please, could you explain the concept of that for all who isn’t.

AY: So Data Relay is not going to be running during COVID because it’s the worst possible thing you could do during COVID but once we’ve got past COVID, and for the last 10 years before we had COVID, what we would do is we would put on a local free Data Platform training event, and we would run it in multiple cities on consecutive days. So last year for example we ran an event on Monday in Newcastle, then on Tuesday we went to Leeds, on Wednesday we went to Nottingham, then Birmingham, then Bristol at the end of the week. In the past we’ve done slightly longer and slightly shorter versions, but we’ve now settled on five events in five days as being a fairly good sort of size and scale for the event. They’re free events, they typically have three tracks of regular sessions and one track of half day workshops for more in-depth, technical, pre-con style classes. At least one of those at each location is normally run by the folks from Microsoft, so you’re learning from Microsofties about how to do cool and exciting stuff with data in Azure. So that’s a lot of fun. And the whole concept is to bring free and accessible training to you. So that’s why we go to so many different cities. And we try and cover as much of England as we can, or England and Wales. We don’t tend to go up to Scotland because they’ve got some great events of their own and they’re doing good. We’ve not tended to go over to Northern Ireland or the Republic, because there’s kind of a sea in the way so it’s a bit hard. So typically we do England and sometimes we’ll go to Cardiff. But yeah, it’s a lot of fun. And for those people that want to join us for the road trip, we put on a bus for speakers and organizers and helpers to kind of go from city to city, which is exactly why it’s the worst possible thing to do during a pandemic, because I can’t imagine anything worse than having a 200-person event in one city and then another city, and another city, with a bunch of tired and probably immuno-compromised people all sitting together on a bus between each city. It sounds like the worst possible thing to do during a pandemic. So we won’t be doing it again until COVID is fixed.

KN: Which basically we don’t know when it will happen.

AY: Yeah, that’s why we don’t currently have any plans or any dates in mind, and we won’t have until such a time as it looks like it’s under control.

KN: And due to this pandemic, this year Data Relay has been cancelled because SQLBits has been moved to that period of time, right?

AY: To be honest, I think possibly when that happened Relay dodged a bullet. The original SQLBits was going to happen in February or March, which was right in the height of the pandemic and when everything was being shut down. So they clearly couldn’t run at that time. (SQL)Bits were offered three dates: one in the middle of the summer holidays, which wouldn’t have worked for various reasons, one basically over Christmas, which wouldn’t have worked, and a third one, which was smack in the middle of Data Relay. So that’s why they rescheduled to October. And they told us in advance of the announcement and they asked us… They were really, really good, they arranged a call with us and we discussed it. And we agreed that we don’t want to stop SQLBits from happening, we don’t want to get in the way of SQLBits, we don’t want to compete against SQLBits. We actually want to attend SQLBits. And SQLBits were in a really difficult financial situation. They basically had the option to cancel the event and bankrupt themselves or they had three dates that they could go for. So they went for that one. And all of this, by the way, was agreed right at the very beginning of the outbreak, before any of us knew anything like as much as we do now. So at that stage we’re like “yep, fine”. We hadn’t got very far of our planning, we paid one or two deposits for one or two of our venues but we possibly only spent five or ten percent of our budget for the whole thing. So we were early enough that we could call it off without taking a massive financial hit, whereas SQLBits was kind of walled in on it. And SQLBits has a much, much bigger budget than Data Relay does. So I think it would have been enormously mean of Data Relay not to say “look, this year is just things happened”.

KN: Just a special situation. And I think the solution was pretty good, was very good.

AY: As I was going to say, we probably dodged a bullet because had that not happened, I suspect Data Relay probably would have ended up postponing or cancelling anyway, but we probably wouldn’t have done it for until a couple of months later, by which point we would probably have spent more of our budget. So in a sense it probably saved us from ourselves. I think at the end of the day, Data Relay 2020 as it was originally planned would have been cancelled sooner or later anyway. We didn’t realize that at the time but we do now. As for when it does run again, I don’t know. Originally, we were thinking about “do we run a November thing, do we run a spring thing, do we wait until October 2021?”. All of those things are up in the air but based on what’s happened, we’re just not even gonna start thinking about or making any plans until we all know with confidence that the thing is finished, that COVID is finished.

KN: What do you think about SQLBits? Will it happen or not? Because we are closer and closer to the date and the pandemic is still with us.

AY: Don’t ask me that question. I don’t know. What are the SQLBits committee saying at the moment? Because I believe that whatever the SQLBits committee is saying is exactly what will happen. I’m not going to try and second-guess their choices. I can imagine that it cannot be easy to be in their shoes. Will it be safe to run SQLBits in October? Will Americans be able to fly over without self-isolating?

KN: Tons of questions.

AY: Yeah. Given that Microsoft is one of the biggest investors in SQLBits and a lot of the speakers are going to be American, I suspect that the ease and safety of travel between the USA and the UK is probably going to play a significant factor in SQLBits, I would imagine. If it’s not possible or practical, or safe for Americans to take part in SQLBits this year, what does SQLBits do about that? I don’t know. Do they find another 30% of the speakers again to replace all the Americans?

KN: Probably not, or it would be very hard. Let’s leave this question unanswered.

AY: If you go back six months a lot of people were talking about “are you going to go to PASS Summit or SQLBits?” because a lot of people were doing the maths and figuring out that actually even for an American it’s cheaper to go to SQLBits than it is to PASS Summit. So I don’t know how many attendees were going to be expected from the US as well. So yeah, I don’t know, and I feel for the folks that organize it and I wouldn’t want to make any judgment about what decision they would ultimately choose to make.

KN: OK, let’s move on to another question.

MS: From the technology perspective, what is the biggest challenge in IT these days?

AY: What’s the biggest technical challenge in IT? It’s not the question you asked but I’m gonna come back to… No matter what you ask, it’s always a human problem. I actually think that the biggest technical problem in IT is the humans who are writing it. And if you can solve those human problems, then I suspect that you’ll do a lot better. I think you’ll get a lot more bang from your buck from solving a lot of the human and organizational problems. And I think you’ll get more bang for your buck for doing that than you will for most technical problems. However, to answer the question you asked me rather than the one I wanted to answer, what is the biggest technical problem, I think security and data protection within the DevOps world I think is something that people have been gradually waking up to. I mean, everybody thought that GDPR was going to be the thing that was going to do that and it’s not really done it to the level that people thought it was going to, but still, I think that as data grows and people want to be able to test that their changes work on larger quantities of data, the provisioning of development and testing databases that contain appropriate test data that does not contain sensitive production data but which does look and feel like production data and have some similar scale to it. I think that those problems are going to be really big. Related to that, I still see a lot of people using shared development servers for database development, which I think is the root cause of so many of the bad practices that you see in database source control and deployment.

KN: Also, the cause of many mistakes and potential problems with deployment.

AY: Absolutely, it’s the root cause for so many different reasons. It’s tangled into branching and self-service development environments and it works on my machine, all of those things are related.

KN: Let’s go back a little bit to the topic about DevOps and databases then. The topic always discussed and maybe not very well understood by people sometimes is “state versus migration”, SSDT versus Redgate or other tools. Which approach is better? I remember that probably one or two years ago you compared those approaches, so which approach is better?

AY: It depends, it absolutely depends. The reason I started writing those blog posts about it… To explain the core debate that we’re talking about right now: when you’re source-controlling a database, you can either use something like SSDT or Redgate SQL Source Control that is going to define how the database should look. And then you’ll use some sort of schema comparison tool like SQLPackage or Redgate SQL Compare or something like that, to generate an upgrade script and deploy it. And then, when you deploy it, you always know the production matches your source code. The problem with that is the comparison engine that you’re trusting, and whichever one you’re using, whether you’re using the Microsoft one or the Redgate one or something else, they’re never perfect. So you need to design your deployment pipeline with the acknowledgement that something could go wrong at some point, so you need to be testing it, you need to be validating it, you need to be running it in dry run environments, and you need to be alerting if anything ever goes wrong, and you need to have some sort of plan B. In many ways, I think of state- or model-based deployment as the 80/20 rule. It’s actually easier to set up and manage a lot of the time, and it’s a simpler developer experience a lot of the time, and it means that you don’t need to worry about writing the upgrade scripts yourself a lot of the time, and 80% of the time it makes your life an awful lot easier. But 20% of the time you end up having to solve the difficult problems. So it’s one of those. You can get 80% of the benefit with 20% of the effort if you go for a state- or model-based approach. However, if what you want is something that’s going to work all of the time and if you want something that’s really reliable, you’re probably going to go for more of a migrations approach. Now, migrations approach is basically a series of update scripts. Script 1, script 2, script 3, script 4. You run them in order. Typically, it’s combined with some sort of deployment history or schema migration log table on the target database, which records which scripts have been run, so when you run your deployment, you just run anything that’s not yet been run, in order. And that sort of approach frees you from relying on the comparison engine. It means that if you’ve had experience with Redgate or SSDT creating upgrade scripts that either don’t perform very well or they don’t solve various technical problems, like deploying a new NOT NULL column without a default or renaming something and it didn’t realize it’s a rename, it thinks it’s a DROP/CREATE and you lose data. It solves you from these sorts of problems. But at the expense that you have to manage these migration scripts. At small scale that is not a big problem. So at small scale migrations tend to make a lot of sense. If what you have is a dumb data store which doesn’t have very many changes going to it very frequently, a small migration script runner with a few dozen scripts is a relatively neat and reliable solution to that problem and it’s not very difficult to understand. But once you start scaling up your development and you’ve got large teams of developers always making changes and maybe you’ve got one of those horrible monolithic architectures where you’ve got one big SQL database in the centre of everything that backends 150 applications and you’ve got different people from different teams all over the place trying to update it with different source control and release cadences. At that point a migration-based approach becomes really, really difficult to manage, and it just becomes this huge repository full of all sorts of scripts and then kind of ordering and the scheduling, and the coordination of these scripts, there are a lot of problems associated with that. To summarize, the model-based approach will get you most of the way there for less effort, and it can scale better sometimes but it’s never going to be a complete solution. There are all sorts of fixes to that which are never particularly nice, like using pre or post deployment scripts or something like that to solve the nasty issues when stuff goes wrong. On the other side, you can go to a full migrations-based approach, which is probably more reliable, but once you start hitting any sort of scale, it becomes a big administrative overhead problem. So yeah, which is the right one? It depends.

KN: Yeah, exactly, to be honest with the people, both approaches have their pros and cons. There’s no perfect solution.

AY: They do. And you know what? What you just said about the people, a lot of the time look at the people in your team. If your team all know and understand how to use SSDT and have got experience with it and know how to handle the issues when it goes wrong, then why would you pick something else? Because they already know how it works. Stop worrying about your source control process. Let’s set it up and focus on your product. And the same goes for if they’re used to using a migrations-based approach. If your team are all used to using migrations and they’re comfortable with it, just the fact of what the team are comfortable using is probably one of the most important factors.

KN: That’s absolutely true if you’re already using something or if you have some specific skills in your team, that’s absolutely right. But I’m just wondering if someone is just starting that journey and setting up that solution, that’s when that question comes up, which solution to choose.

MS: How does cloud fit here? What do you think about the cloud-first, cloud-only approach? So all of the approach with the database, with the migration or the state approach and then also the cloud. Is it a perfect match or is something different?

AY: Yes, I’m a big fan of the cloud. I think that a lot of people think the cloud is going to solve all of their problems. That’s not true, it’s not going to solve all of your problems, there are going to be new problems going to the cloud. But fundamentally, I think that the cloud is probably a better place to host your stuff. Whether that means you should be lifting and shifting and putting everything there immediately or not is a different question. Whether or not you should necessarily be migrating your mainframe application to the cloud, probably not. I think that the cloud definitely is going to give us huge advantages and already is doing, but it’s going to disproportionately reward applications that are designed cloud-first from the beginning rather than migrating stuff that is already running on-prem. I think that there is a tendency to want to lift and shift stuff into the cloud, and if your application was not designed with the cloud in mind, then you’re not necessarily really going to reap the benefits. The cloud is definitely a good thing and definitely stuff is going to be going there. Do I think that it’s the right thing and we should all be immediately going there right now with everything? No. I think that there will be a lot of people who will choose not to go to the cloud, and I think there’ll be a lot of applications that are running today that never will go to the cloud. And I think the decision has to be made on a case-by-case basis.

MS: What is the hottest feature of SQL Server or the technology that you are learning at the moment, the first item on your learning list?

AY: So at the moment, I’m doing quite a lot of learning some Infrastructure as Code type stuff on AWS, just because I need it for a project that I’m working on at the moment for one of my customers. So I’m playing a lot with Octopus Runbooks and AWS PowerShell and CloudFormation and stuff like that to spin up some infra. It’s not particularly new tech, it’s just tech that I’ve not personally played with that much because I’ve spent a lot more time at the database level, so now I’m not an expert at all by any means in it. I’m just beginning to pick it up and figure out how I build out my infrastructure in AWS, because I need to for a project. And I’m enjoying it because it’s probably long overdue. So that’s what I’m personally working on at the moment. What am I excited about learning, I am very excited to see how Kubernetes and SQL Server are going to play out. How we’re going to start hosting SQL Server in future years, so I’m really, really interested to see what happens there. So what do I want to be learning right now? I’d probably love to be doing a lot more Kubernetes than I’m doing right now. But what am I actually doing? I’m doing a lot of basic automation of EC2 spinning out VMs and stuff, just because it’s long overdue and it’s not as exciting as it might have been five or ten years ago. But it’s what I need to do right now for the project I want at the moment.

KN: Cool! OK, so almost at the end of our conversation, I need to ask you about that because Michał and I, we are Polish guys, so I need to ask you, because I remember that you have some Polish roots, don’t you?

AY: I do have some Polish roots but they’re several generations past. So my grandparents ended up in London at the end of the Second World War as refugees. So my mother was born in 1948 I think in London to a pair of Polish refugees who were a young married couple who had just been through the Second World War together. So I do have a large branch of my Polish family, because they all seem to have about 15 children, so I have a very, very, very large Polish family that seems to cover a broad spectrum of Polish society and we were actually scheduled to have our five-yearly zjazd (reunion) in Poland this year. It got cancelled because of COVID but hopefully we’re going to do it next year instead. And each time we go, it feels like there’s 30 or 40 new little children. I think at the last one, there were about 60 people on the zjazd, and based on expectations, I think there might be over 100 this time, of which the vast majority of them are Polish.

KN: So, I’m guessing that you’ve been to Poland several times at least?

AY: Oh, yeah. I’ve been to Poland a few times. Unfortunately, I don’t speak the language and I feel very guilty about it but as I say, I’m myself second or third generation, I’m not sure how you count it. My dad’s English and you would never know my mom was Polish.

KN: Don’t worry, we can teach you some words. For example, at the end of this conversation, you can say in Polish, like “goodbye”, you can say “do widzenia”.

AY: Do widzenia, dziękuję.

KN: OK, Alex, at the end of our conversation, could you tell us where we can find you, where the people can find you in the Internet, how can they follow you?

AY: The main social media I use is Twitter. I’m _AlexYates_ on Twitter. I don’t really use Facebook anymore. I’m very close to deleting my Facebook account. I actually would have done, had there not been a particular thing for my kid that we do on Facebook, which I’m not going to go into. You can get me on LinkedIn, but normally I’m not very good at responding to messages there because it’s mostly recruiter spam, so the best place to get me is Twitter. Or go to the website dlmconsultants.com and if you want to contact me about potential consulting or mentoring relationship, then dlmconsultants.com and there’s an email address you can use there to contact the company.

KN: Cool, fantastic. Thank you very much again, Alex, for the conversation today and for your time.

AY: Thank you very, very much for having me. Dziękuję!

MS: Thank you.

KN: Dziękuję, do widzenia!

Useful links

Alex’s profiles & websites: Twitter | DLM Consultants | Speaking Mentors | Blog
Conferences: Data Relay | DataGrillen | GroupBy | SQLBits