IT/Software career thread: Invert binary trees for dollars.

Vinen

God is dead
2,782
486
Yes, but if they develop code intended for quality assurance, aren't they QA?

No. And I'm stating that the Developers who write the code should also be responsible for owning frameworks and test code for said code.

The only QA that should exist are not traditional QA. Performance Engineering and Security.
 

Deathwing

<Bronze Donator>
16,384
7,385
No. And I'm stating that the Developers who write the code should also be responsible for owning frameworks and test code for said code.

The only QA that should exist are not traditional QA. Performance Engineering and Security.
I would say there's still a need for some form of QA. What developer would own the testing system as a whole? Or do you think there wouldn't be a unified testing system? Who owns the system to collect and display test results? Who owns the actual machines the tests are running on and triages issues when tests fail?

Or does all of that fall under performance?
 

ex-genj

Golden Squire
638
115
yeah no you want me to do all of QA (UAT, regression testing) in addition to all of the dev? You can pay me twice. And by that I mean x1.3 :rolleyes:
 
  • 1Like
Reactions: 1 user

TJT

Mr. Poopybutthole
<Gold Donor>
40,930
102,729
No. And I'm stating that the Developers who write the code should also be responsible for owning frameworks and test code for said code.

The only QA that should exist are not traditional QA. Performance Engineering and Security.

Not sure how efficient that would be. I built up a team of SDETs where I work as I wasn't going to be doing functional QA (those dudes are generally braindead retards).

The biggest benefit I ever had from working as an SDET was not getting siloed. I got to work with everything on all levels of the applications I was working on. I think it will be a very long time before functional retard QA is dropped. Mainly because of people don't like change too much and secondly because vast majority of functional QA I met would be hard pressed to learn Selenium even with it being quite user friendly compared to earlier frameworks (Coded UI, barf).

Deathwing Deathwing It wouldn't fall under performance because if you're automating performance tests you're automating the network calls made to, say, a web application and you would never interact with a UI layer. The test can fail and get a 500 error or fail your transaction time requirement or something but it would never even register an error in the UI.

So you would have to both have something that ran selenium tests designed totally differently and your performance tests with different fail triggers running on potentially the same boxes. Not that that would be too crazy I guess but most wouldn't prefer it.

I am not convinced you would get developers who would be cool with doing all of that as it would be extremely time consuming to code something, then make your automated test suite for it and maintain it. Performance engineering and test automation are jobs in their own right.
 

Deathwing

<Bronze Donator>
16,384
7,385
I didn't think it fell under performance either, I was just curious where most of what I do day-to-day(test automation) would.

At this point, I'm considering a combination of unit tests, spider sanity checks, and headless chrome. And probably some manual release testing. I really don't want to have to maintain selenium tests. Selenium is such a fucking nightmare once you start adding browser dimensions.
 
  • 1Like
Reactions: 1 user

TJT

Mr. Poopybutthole
<Gold Donor>
40,930
102,729
I didn't think it fell under performance either, I was just curious where most of what I do day-to-day(test automation) would.

At this point, I'm considering a combination of unit tests, spider sanity checks, and headless chrome. And probably some manual release testing. I really don't want to have to maintain selenium tests. Selenium is such a fucking nightmare once you start adding browser dimensions.

So in theory you can have really nice Selenium suites. But it requires the devs to work with you and put static xpaths (or whatever you're using) everywhere. Which they will often not do because it is more work for them and some of the more funky tools just generate random shit all the time (Keno UI for example) and make it impossible to do that.

In that way I see what Vinen means but I wouldn't want to have to make selenium tests after I finished developing whatever... every single day. Fuck that noise.
 

Deathwing

<Bronze Donator>
16,384
7,385
So in theory you can have really nice Selenium suites. But it requires the devs to work with you and put static xpaths (or whatever you're using) everywhere. Which they will often not do because it is more work for them and some of the more funky tools just generate random shit all the time (Keno UI for example) and make it impossible to do that.

In that way I see what Vinen means but I wouldn't want to have to make selenium tests after I finished developing whatever... every single day. Fuck that noise.
Will xpaths even help if they are using javascript and <canvas> to draw almost everything?
 

TJT

Mr. Poopybutthole
<Gold Donor>
40,930
102,729
Will xpaths even help if they are using javascript and <canvas> to draw almost everything?

Doesn't have to be an xpath but the devs just need to actively give things static labels for you to look for. Stuff that generates code like Canvas almost never does it in a way that makes sense or easy.

Getting devs to actually do this is another story. As they would first have to have had experience automating tests outside of unit testing and actually give a shit about QA (the latter is very unlikely to happen!).
 

Deathwing

<Bronze Donator>
16,384
7,385
Yeah, I know. It's really frustrating because I know very little about javascript and the guy that's heading this project likes QA but does everything he can to not do any himself. So you have testing people trying to figure how to test a language and code base they are unfamiliar with, when it's the people that should don't care.

At least in our setup, I feel like they should come up with the test strategy and the QA people would integrate into the test system, including how test results get back to the developers. From a developer perspective, is that too onerous/unfair?
 

TJT

Mr. Poopybutthole
<Gold Donor>
40,930
102,729
At least in our setup, I feel like they should come up with the test strategy and the QA people would integrate into the test system, including how test results get back to the developers. From a developer perspective, is that too onerous/unfair?

Not sure what you mean here. Are you telling me your office does not use Jira/TFS/ALM or some similar product? Or do you mean some visible kind of high level reporting to get test results to teams?

My whole department used my SSRS/Data sucking application for like 4 years. It pulled TFS data and reorganized it into some nice reports that you could auto-email out to project teams with native Sharepoint tools. I have done a lot of work when it comes to designing project reporting. If you're in need of ideas on that let me know.
 

Deathwing

<Bronze Donator>
16,384
7,385
Not sure what you mean here. Are you telling me your office does not use Jira/TFS/ALM or some similar product? Or do you mean some visible kind of high level reporting to get test results to teams?

My whole department used my SSRS/Data sucking application for like 4 years. It pulled TFS data and reorganized it into some nice reports that you could auto-email out to project teams with native Sharepoint tools. I have done a lot of work when it comes to designing project reporting. If you're in need of ideas on that let me know.
We have our own homegrown data manager with a myriad of reports. How much developers peruse these varies. But the information on how much tests are failing/passing is there for them to see. And we have other forms of automation for attributing failures to certain bugs and then emailing the owners of bugs about new failures. This is the kind of work I think QA should be direct owners of.

My question was who should come up with the overall test strategy. I think a lot of people, at least my current job, fall into the trap of "well, it says test, so QA gets to do it".
 

ex-genj

Golden Squire
638
115
I've never heard of any developers "coming up with the overall test strategy". I certainly haven't.

Anyone ever go to India for work trip? First time for me in a few days, 18 hour flight, good times. Director says "bring them candy" (?) ok then.
 

Deathwing

<Bronze Donator>
16,384
7,385
I've never heard of any developers "coming up with the overall test strategy". I certainly haven't.

Anyone ever go to India for work trip? First time for me in a few days, 18 hour flight, good times. Director says "bring them candy" (?) ok then.
Who does then? Who decides how your code is tested?
 

ex-genj

Golden Squire
638
115
To be totally honest I don't give a fuck. My job, directly, is to complete user stories, i.e. "this should do this". When I think its done I mark it as complete and move it forward in jira. How it gets accepted is not my responsibility, even as a lead.

On my team stories get accepted ("tested") by the product owners and our lone offshore QA lead. If they fail UAT (breaks stuff), we discuss with the developer about the requirements and if they understood them and what went wrong. From what I can tell we have no overall testing strategy other than our QA guy attempting to make some stuff work in Cypress, maybe, IDK? I do mandate unit testing (and you can't push with any failing unit tests) so that's something but its certainly not comprehensive.

We may be "lucky" in that my team works on an internal product that isn't customer facing so we can get away with nonsense like this most likely. BTW please don't read the above as me being flippant or not caring about testing, I do think its important, but again its just someone else's problem, my job is to deliver features not make non unit test cases. This is ultimately a management problem, not us worker bees.
 
  • 1Like
Reactions: 1 user

Deathwing

<Bronze Donator>
16,384
7,385
To be totally honest I don't give a fuck. My job, directly, is to complete user stories, i.e. "this should do this". When I think its done I mark it as complete and move it forward in jira. How it gets accepted is not my responsibility, even as a lead.

On my team stories get accepted ("tested") by the product owners and our lone offshore QA lead. If they fail UAT (breaks stuff), we discuss with the developer about the requirements and if they understood them and what went wrong. From what I can tell we have no overall testing strategy other than our QA guy attempting to make some stuff work in Cypress, maybe, IDK? I do mandate unit testing (and you can't push with any failing unit tests) so that's something but its certainly not comprehensive.

We may be "lucky" in that my team works on an internal product that isn't customer facing so we can get away with nonsense like this most likely. BTW please don't read the above as me being flippant or not caring about testing, I do think its important, but again its just someone else's problem, my job is to deliver features not make non unit test cases. This is ultimately a management problem, not us worker bees.
Nah, I didn't take it that way. I'm guessing that the amount of testing we do is above average, so your response is not flippant, more that it seems to be the norm for the industry. Whether that's good or not, idk. I've often felt that the amount of testing we do is borderline frivolous and has passed into deep diminishing returns a while ago.
 
  • 1Like
Reactions: 1 user

Needless

Toe Sucker
9,169
3,268
man i cant figure out a super small issue with my react app, it's driving me nuts because i can't replicate it on my xcode simulator (it only appears on devices, im guessing its loading the app too fast?).

basically how it should load is the attached picture, but on devices the textinput's placeholder text doesn't load immediately, so the textinput isn't responsive until you actually touch the device. Is there a way in xcode to set an initial load timer or anything?

Spoilering since its bigger

good
loginscreen.png


bad
loginscreenbad.png
 

Noodleface

A Mod Real Quick
37,961
14,508
So I'm bored and ready to look for a new job. Partially because I had a Google recruiter hit me up. Not that I think I'd pass their interview system, but it got me thinking.

I'm feeling pretty lifeless here 75% of the time. 25% of the time I get to be innovative to a degree, but I don't code everyday. Especially lately they've had me wearing many hats (traveled to a military test facility, did a formal two week long test session here, lots of 'management meetings'). I feel I'd either have more fun somewhere else or should just transition to management.

My old boss invited me to interview for amazon Alexa awhile back and was considering that too. All the jobs are in Cambridge so I'd have to have a miserable commute again. But maybe there are better perks.

At what point does my job hopping look shitty? I just hit two years here.