It's a sign of how the testing feels about this that I've been freed of the person who makes sure my defects aren't sarcastic or condescending, and even more terrifyingly, when I ask my IV for a second opinion on a defect, she blesses me on my way to racking up the defects and doesn't check to see if I'm explaining how to find the ww3 site to look up some goddamn web standards.
The thing that's really a sign of the apocalypse though is, benefit of the doubt is gone, which is why we're suddenly filing higher over smaller stuff or stuff that's transient. In general, about maybe a third of defects are what I call the doubter-defect; it's a toss up whether you go ahead and file it or just wait for the next code bounce, because usually it's something relatively small and not a huge deal, and in general, it'll get fixed the next bounce.
The thing is, testers and developers and the BAs live in a state of general peaceful armstice with each other because in general, we're all on the same side, and getting sixty defects in a day, half of which are obviously just glitches that will get smoothed out before code freeze does not do anyone any good and they also waste time. We all know that the weird thing on X page is because of Y and they're fixing Y now, so it's fine. When your nodes are collapsing and Cisco is using really upsetting words about your hardware failures, you don't sweat stuff that you trust will be fixed when Y is fixed. It's practical, it saves time and energy, it gets things done faster, and no one is homicidal.
It also requires trust, and we don't have any of that anymore; we're running on a deficit of trust, because they're still doing medium to major code changes and epic text changes and from day to day, doing exactly the same test, I find something new and strange that no one's seen before and we have no idea where its come from. To put this another way; we should have deployed this last Saturday. As of today, about a quarter of our tests are either failed or we can't start them because of code problems. And they keep redesigning it to be--and I can't believe this is possible--even less user friendly than it was. Gone is the autoredirect to login when you finish setting up an account; now you get a wall of text. This program is painfully counter-intuitive for anyone who has ever in their live used a webpage, and I don't say this lightly, I feel personally validated about building sites that were white text on a black background now because they are trying--with terrifying results--to work against how pretty much every other webpage in existence is designed.
But even that is you know, a matter of opinion even though seriously, what the everloving fuck, if it's functional, and it's not. Today I filed one after the other five defects, all of which would make it impossible for a user to use the site but three at least were still the doubt-defects; they looked and felt transient, which is the natural result of fixing a lot of big problems so smaller stuff gets through. Two weeks ago, I would have paused my test, emailed my supervisor and my IV, and told them about the problem and got a judgment call that more likely than not we would have held off until the next code bounce to see if it was just because of a major code change that was in the process of getting fixed. Since last week, if anything goes wrong, we file it on the spot.
...this would be a terrifying number if our testing environments didn't have to be bounced several times a day to fix the node issues which lead to web service errors, so we can't do anything.
We're reopening fixed defects; that is not common. We're filing defects on fixes for defects, which is goddamn freaky. I wrote sixty tests, all but three of which passed; they are breaking things we already tested, and changing them so much that some of the tests don't even test anything that exists anymore.
Then I got an email that they were revising the design document. The design document. The business rules. The entire backbone of what the code is supposed to do is still being revised and honestly, what the hell.
However, all is not doom and gloom. Mostly. This is because hilariously, our web service errors are a string of numbers followed by Clone1 or Clone2 then a letter combination, and they are goddamn annoying. My last email to my supervisor had the following header:
UAT2/SSP - Account Creation - CBO - Attack of the Clones
(UAT2 - our testing environment this build; SSP - the program we're testing; Account Creation - the part of the program I'm testing; CBO - a part that goes with Account Creation; Attack of the Clones - a terrible movie but useful for subject lines.)
It's a sign of the fact that we're kind of out of it when that email was forwarded upward and outward to um, everyone before my supervisor read the subject line fully, and dude, I know you're thinking not a big deal; no. Management sold it's sense of humor when they sold their soul for fucking over WHP. It's creepy.
For all that, it does make the day go faster, and it's exactly the wrong time for that to happen. We need more time, not less.
My work related rant for the week.
Interesting, I realized it's been about four months since my last episode of hard depression; I'm frustrated and tired and angry about work and everything, but that's actually a positive sign.
It's been two years and three months since I realized it was happening, and ten months since it began to lighten, and four months since the last time I couldn't even breathe deeply because the effort seemed wasted.
I've been writing since March; this is the longest I have written daily in two years and three months, and there's this incredible--I don't even know the word, gratitude or relief or something. It's stupid to think it went away for good, but I kind of thought it was going to. I literally hated looking at my old fic because of certainty.
I need a vacation soon, maybe. I don't know. Maybe just rent a hotel room downtown and sleep a weekend away with room service and a lot of new books. Something.
Posted at Dreamwidth: http://seperis.dreamwidth.org/934501.ht