253: The Life of a Solo Automation Engineer with Chris Kenst

Hit it.

Welcome to the Test Talks podcast, the place to go to geek out on software testing. And now your host whose mission is to help you succeed with test automation, Joe Colantonio.

Hey, it's Joe. And welcome to another episode of Test Talks. In this episode, discover what it takes to become an automation engineer in a startup with Chris Kent. Chris also shares tips on how to improve your automation skills, leveraging education and conferences. You also discover an awesome resource to help you never miss another can't miss testing event again. Chris has been improving software quality for nearly 15 years. He now works at a startup called Blooming Nation as a automation engineer. He's also a board member for the nonprofit association of Software Testing, which is where he's working to build better testing education through ast, VBST classes, webinars and more. Really is dedicated to the craft of testing. You don't wanna miss this episode. Check it out. And speaking about the craft of automation and testing in general, how to get better, I have an awesome resource for you, Secure Guild. Have you joined yet? If not, why not? It's an awesome resource to get you started learning the basics of security testing. I think it's a topic that most testers completely ignore and it's actually a topic that they're perfect for. So if you haven't already, check it out, head over to secureguild.com and register today. You only have one week left if you listen to this when I release this episode, so want to jump on it before you miss out? SecureGuild.com.

Test talks is sponsored by the fantastic folks at Sauce Labs, the cloud based automated testing platform that eliminates the need to maintain your own selenium grid and and test infrastructure. Try it for free today. Visit testtalks.com and click on the Sign up now link under the homepage Sponsor section.

Hey Chris, welcome to Test Talks.

Thanks for having me.

Awesome. So I've been following you and talking to you for a long time on Twitter and love your Testing Conferences site. So I thought we'd talk all about your experience with automation, what you're doing at ast, your role as an automation engineer, all that cool stuff. But before we get into it, can you just tell us a little bit more about yourself?

Yeah, definitely. Like you mentioned, I am an automation engineer at a company called Bloom Nation. We're located in Santa Monica, California. If you go to bloomnation.com, you can put in a zip code and find local florists. Bloomnation.com is a marketplace for connecting local florists to consumers and handling all the e commerce stuff. But our big business is we're actually a software as a service platform where we actually go out and we find local florists and we build them an e commerce site and then handle all of their e commerce. And.

It'S kind of cool because it's basically the company is helping bring local companies onto E commerce and it turns out the floral industry is sort of like behind the times by you know, 10 or 15 years. So it's, it's this very cool e commerce platform and so that's what I do, sort of as, you know, quote unquote day job and it's, it's been a lot of fun. I've been here a little over a year now, but in my past I've been in testing and test automation for 15 years I want to say. So not new to it and I'm also a board member for the association for Software Testing which has been fun and an interesting challenge that we can talk about. I've been associated with the AST for a long time. I've been a long time teacher instructor for their BBST classes which are basically the only university level education in the software testing space. I also do some writing, some blogging and.

Like you mentioned, I created and run the testing the testingconferences.org website.

Awesome. So let's unpack some of that. It's a lot of stuff.

Yeah, the first, a lot of stuff. Yeah.

The first thing I think is interesting, I think you told me for your new gig at Blue Nation. Well, it's not a new gig. You've been there for a year. Are you the only tester there? And if you are, I'm curious to know what that's like being a sole tester in a startup like that.

Yeah. So most of the companies that I join, I am the sole tester. It's this sort of funny thing. I talk to people and they're like, oh, startups don't hire testers. And I'm like, I always get hired startups. I'm also usually the, I'm also usually the sole tester. And then so basically what that means is I'm usually recruited by the CTO and the CTO usually comes and they're like, hey, we have a testing problem. Here's what it is. Can you help us figure it out? So that's sort of been my mode of operation for a long time at the past several companies. But with respect to Bloom Nation, basically what that means is we have a small development team in the US and then some offshore folks in Minsk and a couple other places.

So we are constantly making updates to our platform to help out our florists. And as an automation engineer, I do some exploratory testing and then I also write automation for our UI level tests, for some of our API tests, and on a rare, rare occasion, some unit tests.

Very cool. So how do you balance your time, say, between automation and exploratory tests and activities? Do you have a plan that you usually do, or is it just whatever you feel like at that given moment? Not what you feel like, but whatever is required at that given moment?

Yeah, it's definitely. And I've made it really clear to my CTO that his name's Jason. A lot of fun to work with, but I basically told Jason, I'm like, listen, you can use me as a chess piece, right? Because we have a small team, it's a small engineering team and a small product team and we all sit next to each other. I'm fully embedded in the engineering team. And so part of it is I was like, listen, just feel free to use me as you need to help mitigate problems, put me in the risky areas, that kind of thing. And so we do have an overall automation plan, but in terms of what I do day to day, it actually does vary based on when is our next release, what's in the release, what is my CTO concerned about? The funny thing about doing automation is that in order to write the automation, you have to understand the features that are there anyways. What that ends up being is you end up having to dive deep into a lot of the features, functionality and features in order to really understand what makes sense to automate and what does.

A lot of people say they don't get hired by startups as a tester, but you tend to. Are you doing anything differently or like, how would, let's say a tester wants to get involved in a startup? Any advice for them or how you're able to get yourself into these type of situations?

Yeah, I don't know that I have a big impact and I mean, I have a big impact when I join the company and I can speak very well to what I do and the value that I add and the things that I like to see the business do. So from once I get the introduction, I'm able to, I think, prove pretty well that I would be an added value. But ultimately for startups to, they have to get to a point. And from what I understand, it's usually when The CTO gets bogged down so much with quality issues or testing that they don't have enough time to do the things they want to do. Then they actually actively look for hiring a tester or an automation person. That's been the last three or four companies that I've worked at. That's been the thing where those companies have realized, listen, the thing we're doing is not scaling. We're having quality issues, we're having testing problems. We need to bring someone in. So that's not so much advice, but other than to sort of dispel this rumor that startups do not hire.

Nice. So what's your toolkit like? Probably a lot of people are working with E commerce sites, regardless of what domain it is. Is there any unique tools that you tend to use for this particular project?

I mean, yeah, so that really depends on what we're talking about. If we're talking about. So like at the UI level, if we're talking about tool sets and tools.

We are. We're sort of unique in that we are a Magento PHP shop on the back end and. And then we're JavaScript on the front end. And really our JavaScript has been for our E commerce sites is really plain Jane. We're doing some React, but that's in some newer pieces of the application. And so it's sort of this mismatch of.

Different languages and technology stacks. So actually when I came in, I asked them, I told the cto, I was like, listen, my background's in in Ruby, but you guys don't do Ruby. He was like, that's not a big deal. So I was like, okay, all right, you're willing to work with me. So when I came in, I wrote a framework for UI testing and I wrote that using WebDriver IO that is a JavaScript implementation of the Selenium bindings. I did that because the company uses JavaScript and it uses Chai and it uses Mocha and that stuff. That stuff was already here, so it made natural sense for me to do that. Then outside of the UI on the API level, we have some node APIs and some magento APIs and then we're using a suite called Dred D R E D D for basically interpreting our Swagger specifications and then generate auto generating tests from those. Then we have some other tools and that stuff that we're doing as well. For example, we have a new project where we're starting to use React.

I'm playing around with Cypress and I think we'll probably use Cypress for the UI level testing on that, it really just depends on the individual application.

Cool. Yeah, I've heard a lot of good things about Cypress. What just caught my ear here is Dread. I've never heard of Dread before. Can you give an example how it auto generates tests for you?

Yeah, so it basically we generate a Swagger specification. And so what that means is that we have.

Our APIs written, our API documentation written in a certain, I think it's called Open API. There's a specific schema, a layout that you basically write your tests in and then it generates a Swagger document and the Swagger document is then published. So that, and it's this very nice easy to read API set of API documentation with examples. It's a very nice look and feel. And then, which is great. So you can document your APIs, but then you go, well, how do you then test those APIs? We could obviously write out all our tests, but it turns out there's this, this framework called Dread, which they actually have the logo of Judge Dread, which is pretty funny. But so they claim that they're a command line tool for validating API description documents against that API. And so basically what it does is it reads in the specifications and it says, okay, this is the, the structure of your API. This is what it takes. We give it a little bit more information on the configuration side and some of the hooks and then, and then it reads through that specification and runs all of the.

Get and post and update.

Examples and then it runs them against our system just automatically. So you don't, you don't actually have to code out, you don't actually have to build the test themselves. It just sort of reads and interprets the document and then runs those itself.

You know, I work for a very large company. I always say this and we have a lot of testers, but they all say they don't have time for automation. And I'm curious to know, you are a solo tester and it sounds like you are doing automation. Is it almost easier to be the only person doing it because you can decide where you spend your time?

A little bit, yeah. But it also depends on how you prioritize things. So the thing that I do is that our workflow is we're basically doing a Kanban system. So everything has stories and then we put the stories up on Kanban boards and then those things get prioritized. And so as an automation engineer, but even when I wasn't, when I was just as a, just a regular test engineer and I was Trying to get the time to do something. The way that I did it was to create stories and get those into our project management system such that I could then get those prioritized. So even if I was doing day to day testing, exploratory testing, I would still be like, hey, by the way, we've all decided that this is an important thing so I'm going to go do this for a couple days or I'm going to spend time to get this done. And now this has become a deliverable. And so I don't think, I mean in my experience, yeah, I do get to sort of make things up and I don't really have to clear too much stuff. But even in the cases when I did it made sense to get those into the project management system so that everyone knew, oh yeah, Chris is working on some automation stuff and then when he's done he'll help me or that kind of thing.

So Chris, also, you did mention ast. I'm not sure how many people are familiar with it. Can you just tell us a bit more about it? Because it seems like you are very much involved in also educating your fellow testers. So just curious to know a little.

Bit more about that. Yeah, I've been, I've been volunteering. I took the. So the association for Software Testing is a nonprofit organization. It's an international member supported organization. So basically we have members all over the world, in Europe, in Asia, in North America, South America, just Australia, like everywhere. And the thing that the association for Software Testing is probably best known for are two things. One is the CAST conference. So it's CAST C A S T, the conference for the association for Software Testing. It's a mouthful, but that's why we call it CAST. And CAST 2019 will be in Florida this year, August 12th through 15th and tickets are on sale. So check it out, a little plug there. And the other thing that we're known for are the BBST classes. And so BBST stands for Black Box Software Testing. This is how I actually got into and understanding what the AST was because if you basically go anywhere, if you go to the administrative test forums, if you go to Stack Overflow, if you go basically anywhere and they're like listen, I want to learn about testing and I want like the best classes that are out there. What are they? And inevitably it'll come up that BBST are the best classes and they're the best in the sense that they're the most in depth. And that's because they were created by Chem Kaner, along with some help from a lot of people and.

Chem taught these classes at Florida Tech when he was still there. They're university level based. They got a bunch of, they got some. Chem got some National Science foundation grant money and basically turned them from offline classes into online lectures. And then built, with the help of the ast, built online classes. So they're basically four weeks very in depth classes starting with foundations. There's a bug advocacy class which is about how to write effective bug reports. There's a test design class which is a survey of over 100 test techniques. And so the AST offers these courses year round and they offer them at very inexpensive prices. So a class is $299 for members. So if you become a member that's $75 a year and then you're able to take classes for $300. And if you've ever looked around at the education that is available, education for testers is getting less expensive over time I think. But it's still way cheaper than going to a conference where you're going to spend at least $1,000, if not more. It's still cheaper than most of the other online or in depth courses that you're going to get. And this is a month long course from a university so you're getting hundreds of references, just so much information.

They're time intensive but you come out feeling like you've learned a lot.

The very first book I ever bought in this domain was Testing Computer Software by Kem Keener. Changed my life. That's actually got my first job as a QA person just by reading that book and then going to the interview. So I know. Have you actually taught alongside Kem and what was that like? Did you learn any, any, any funny stories about that or any knowledge that you gleaned from that?

So Chem's back. I have taught alongside kem. I taught alongside CAM for one of his domain Testing classes. So there's sort of a fourth. So there's two basically people that teach the BBST classes, the association for Software Testing. And then Kim does teach on his own, usually in collaboration with, I'm forgetting their name, but with another organization as well. And so you can. His or his company is called Kaner Fiedler and Associates. But he teaches one extra course and that one extra course is the newest one which is called Domain Testing BBST Domain Testing. And I've taught that along with him I think just once. And the thing about Kim that you'll get once you Take his classes or read any of his books, lessons learned, the BBST workbooks, any of those is that he's incredibly thorough. His background is, you know, he has a law degree. I think he has a PhD. So he. His response is, even if you ask him, and I've met him in person, I went through the domain testing in person class at Florida Tech years ago.

Even when you ask him small questions, he's very, very thoughtful and very extensive in his replies, which sort of makes sense. Once you see his material and things that he does, it's really unlike anyone else.

I think you read my mind. My next question was domain testing and you brought it up. What is domain testing? I guess is the next logical question.

Yeah, it's not. It's not a. The common terms that people understand it as are boundary analysis and equivalence class partitioning. So those are the phrases, the terms that people understand in terms of test techniques. But it turns out that those bubble up into this term, an umbrella term called domain testing. And domain testing basically is just a set of numbers. So domain testing is sort of the more academic sense or the more academically used word. So domain testing has been popular for, I don't know, 50, 60 years. It's been used in a lot of academic circles and that kind of stuff as testers. We basically know it as boundary and equivalence class partitioning, but they go hand in hand when you are dealing with certain data types like integers and floating points and that kind of thing. Which is to say that even though boundary analysis and equivalence class partitioning can yield. Can be used at different times and yield different information, a lot of times they work really well together. So you may have, if you're partitioning a domain because you're trying to find a small number of values that will cause problems, sometimes those are actually boundaries. Um, and so, you know, if depending on the data type, like an integer or some integers and I think floating points and that kind of thing, a lot of the times those values that you want to extract that have the most power are usually, like I said, boundaries. So that's why they're sort of bundled together.

Can you give an example? Like, it sounds like it'd be very helpful if someone was testing an E commerce site because you have numbers and currency and things like that. So have you ever used it for your current test plans and did you find any actual issues?

Yeah, so I do it sort of without having to think too much about it. It's actually used a lot in finance. So you're right in the sense that it's actually used a lot of the financial industry. So if you have a lot of input fields where you're accepting, calculating and then display, displaying numbers or values, that's when it's really, really useful. It's used. It's one test technique that you can use. So, yes, I have used it.

And I've found some bugs with it. The bugs that I found have not been very important bugs in that sense, but on at least one occasion, I've personally fixed one of the issues that I found. And then I think we do have a few bug reports of things that have been found with it. But that is a regular part of my toolkit for sure.

Awesome. So another thing I found as I was researching this interview was you have another presentation called Shaping the Identity of a Tester. You just give a bird's eye view of what that's all about.

Yeah, so that was about. That was a number of years ago. It was about basically, if you're a tester. At the time I was recommending that testers should be blogging, that you should be trying to build an online identity.

Of where you want to be in your career, which is to say that you should design sort of the things that you learn around where you want to go. And a part of that would be there's so many testers in the world, there's so many people within the jobs that. But very few people are really known. And so if you were trying to set yourself out a part of that and trying to.

Have better credibility and sort of make your own mark and be able to better market yourself, whether that's with recruiters or companies or for speaking engagements or whatever it is, it really makes sense to sort of grab hold of your identity and sort of try and understand what that is and then make a place for it. Sort of create your own brand in a sense. Now, I would definitely recommend, if you were thinking about doing that, blogs are still good. Not everyone has the desire and time to write, which I think is fine, but you can do other things, including.

Being on Twitter and. And writing on occasion, maybe for a certain publication. But there are things that I think more testers and more people just in software in general should try and own that identity and try and show why they're unique and try and create a portfolio of the work that they've done.

Chris, Another great segue is that actually how I first found about you is I've been searching the Internet for online conferences for testers and testingconferences.org kept coming up. And then I noticed my site, my conferences being listed on it. I'm like, who owns this site? And of course it leads back to you. So that's kind of seems like something like you're talking about as a tester, just getting out there helping folks, and eventually these type of little breadcrumbs will lead to you and open up opportunities almost.

That's right, yeah. And testingconferences.org sort of was just, just a fun, random thing. I'm always thinking about ways to get more useful information out there for testers. And conferences seem like this really, really simple thing where I talk to a lot of people and they were like, oh, I was going to this conference, is going to that conference. I was like, I never heard of these things, like where you find, you know, outside of like Star West, Star east, some of the bigger ones, how do you hear about these? And not even conferences, but like workshops, how do you hear about, you know, so if you think about conferences, there are really, really big conferences, there's really small conferences. So like CAST is a really small conference. And then there's these incredibly small things called workshops where people get together for two or three days and it's this very high intense learning situation and then everyone walks away, sort of better understanding a certain subject. And you go, well, how many of those are there out there in the world? Where can I find them? The answer was nowhere. I went, oh, well, that seems like a good idea for a side project. I got the inspiration from this site called Agile Conferences.

I found that it was using GitHub and I had been wanting to play with GitHub and specifically GitHub pages in Jekyll. I was like, oh, I could probably put something together. And I, I did. And it turns out it's pretty useful. And it's constantly one of the number one search results. Number one or number two on Google. So it turns out a lot of people are also really finding it useful, which is really great because I get a lot of satisfaction out of knowing that it's helping people find valuable resources.

So speaking of conferences, a lot of sessions I've been seeing more and more about are on topics around AI, machine learning and testing. What are your views on that?

Yeah, so I joined, as you know, I joined the test, the Automation Guild this year. Part of the reason that I joined the Automation Guild was because I think we're all sort of seeing that there are a lot of tools that are coming out that are available for testers that are using some form of machine learning. Which is really interesting. Right? They're trying to solve problems. Some of them are unique and useful problems, some I don't think are. But like, Tariq King had that really great talk about using bots essentially to crawl an application and to figure out, sort of actually do some test design and some in addition to test execution. So I think that kind of stuff is really useful. And as someone who's an automation engineer, I'm like, how can I make myself more effective? And. And sort of.

What tooling? And that kind of stuff is really effective. And that's sort of the thing that strikes me as like, wow, this could be really powerful. But so even here at Bloom Nation, sort of segueing on, but staying on the ML topic, like, we are starting to look at ways to use ML to categorize and understand the products that our florists sell. Because there's, it turns out, like we actually have a lot of data about what our florists, the products that they have, the different types of flowers, when they're useful, when they're not. When do they sell the best? When do they not sell the best? And it turns out if you start to build some machine learning models, turns out there's a lot of uses for that. So I see it both on the like, hey, is there a tool that I can use to help test? But also I can see in the near future, like, we're actually going to have applications that are using machine learning and that I'm probably going to have to know how to test. So I think that whole area is just very, very fascinating.

Okay, Chris, before we go, is there one piece of actionable advice you can give someone to help improve their automation or testing efforts? Let us know the best way to find or contact you.

Yeah. So actionable advice. I know you asked this to every person and I thought I was going to be prepared for it. I think it's going to be corny. But the best way to. The best actionable advice I can give to someone, whether you're wanting to learn automation or just wanting to get better at something, I found the best way to make time for that kind of stuff, whether or not you have the time after work, but if you want to get it done during work, I found the best way to do it is literally to schedule the work. It turns out in my experience that my managers and the people around me are really supportive, especially as engineers, to be able to learn new things. And if you tell someone you want to learn something new, as long as you get over the fear of actually expressing that they'll generally help you do it and even allow you to do it on work time most of the time. So if I were to give anybody actual advice, I'd be like, I would say that if you want to learn something new, bring it up, put it out there, and figure out a way to sort of make time for yourself at work.

In terms of getting in contact with me, you can find me most places Kent so Twitter GitHub, I'm ckenst. You can also find me@kentst.com that's where I do most of my writing and I think that's pretty much it.

Thank you Chris for your automation awesomeness. For links to everything of value we covered in this episode, head on over to testtalks.com 2.

And while you're there, make sure to click on that. Sign up for a free trial link under the exclusive Sponsor section to learn all about Sauce Labs awesome products and services. So that's it for this episode of Test Talks. I'm Joe. My mission is to help you succeed in creating automation awesomeness. As always, test everything and keep the good. Cheers.

Thanks for listening to the Test Talks podcast. Head on over to www.testtalks.com for full show notes, amazing blog articles and other automation awesome.