Hi!
Correct me if I'm wrong, but here's what I can gather: Total spending was $1.7 million less than budgeted. Tech spending was $1.7 million less than budgeted. And $1.7 million was sitting in the bank accounts at the end of the fiscal year.
We did not spend on hardware, because, well, I explained already in some other email. We did not reallocate the money to hire lots of developers because we didn't know how the hardware spending will look like.
The solution to not being able to afford "proper .com-level salaries" is to offer people nothing?
We're not competitive on the job market. So, if we don't get qualified engineers willing to work for less, what should we do - hire less qualified ones?
I'll fix the dumps for minimum wage plus daycare for my two kids.
;-) Are you for these conditions for other projects too?
I don't know about that. It's a pretty cool problem, it's just a difficult one to solve. Or maybe it's a cool problem because it's difficult to solve.
I find it very boring problem myself. Probably thats because I'm spoiled by really cool problems at work, or maybe somewhat cooler problems at Wikimedia ;-)
It would certainly help. The problem with the dumps is that they're so huge.
They're small. :) And yes, I know what is the problem with dumps :-)
Not being able to test solutions on a system just as huge is a serious constraint.
Problems are known well enough to create quite some work. Resolve 10 "what ifs" and you're nearly done.
Plus you have to remember that the WMF's particular installation is not the common one.
Agreed.
There's probably enough information out there to pretty much replicate it, but that's another serious constraint. I'm certainly not willing to deal with those unnecessary constraints.
OK!
It was the existence of the history dumps that enabled Judd and I to find the oversighted SlimVirgin edits.
That encourages production of dumps, right? :)
Once again, I've heard that for three years now, so forgive me for not believing it until I see it.
OK!
Domas