the wrong sunrise

Another damp and grim morning, this one with the added joy of trekking out to the far reaches of industrial NE Portland to subject the old MG to the rigors of DEQ emissions testing, followed by quite a bit of PHB-herding. It was really quite ugly out there this morning, and I had to venture out into the elements around sunup, which is quite unusual for me. It was an ugly morning and I took no photos, but I did have a few sunrise photos from back in late January gathering dust in the archives, so here they are, instead. I realize these pics are of the wrong sunrise, but it was better, and this is my blog, dammit.



In any case, despite today's pathetic excuse for a sunrise, I really ought to be in a good mood today, even though I'm not. First, a miracle happened and the car passed emissions without having to visit the shop, which is extra nice since I'd procrastinated about this until almost the end of the month, and my tags were set to expire at midnight on Halloween. So that was an unexpected nice bit, even though I got rained on a lot in the process. I guess on the bright side this means I didn't waste any part of a nice sunny day waiting in line at the DEQ station.



Second, I (probably) saved my employer a mountain of cash today. Remember the overseas outsourcing thing I mentioned a couple of days ago? It turns out that the widget this firm in India (which will remain nameless) wanted to build for us -- well, the one part our management thought we actually wanted, out of the huge baroque architecture they were proposing -- is actually just a thin wrapper around the inotify mechanism that comes with Linux kernel 2.6.13 and later. They'd led us (and by "us" I mean "our PHBs") to think this "component X" would be vastly more complicated than that, and wanted gazillions for it, where in reality I could probably write the damn thing over the course of a weekend. Hell, probably a couple of hours would do the trick, to get the core functionality down. Granted, it still wouldn't work with kernels pre-2.6.13, but their stuff wouldn't either -- although they were quite happy to do the "mumble, mumble, subsequent version" song-and-dance instead of leveling with us. Hell, I understand they even led our PHBs to believe this would be super-easy to port to Solaris, Real Soon Now, which is absolutely, positively untrue.

Maybe I'm overreacting, but it really feels like they were trying to pull one over on us. It's vastly more infuriating when someone who is supposedly an engineer tries to con a fellow engineer; you expect that from marketing and bizdev folks, because that's what they exist for. It's a big world, and there's a place for that. But when someone who on paper ought to be a fellow geek tries to dazzle me with slick Powerpoint slides and a bunch of handwaving, that's a mortal insult, and I prefer not to do business with that sort of person. I wish I could just be happy about saving the company money, but I'm still feeling insulted, and that takes priority, I'm afraid. The worst part about it is that they thought we (and by "we" I mean "I") wouldn't catch on to what they were up to. They thought I wouldn't get curious what was so special about kernel 2.6.13. They thought I hadn't done my homework, and wasn't familiar with stuff like inotify/dnotify. They disrespected my Google-fu, and that's something I simply can't abide.

Which is not to say the deal's off, necessarily. I've been around the industry long enough to know things often happen for nontechnical reasons. Right or wrong, it's a fact of life we all have to live with. And I gather that our PHBs and their PHBs go wayyyyyy back, possibly all the way back to wherever it is PHBs spend their larval years (a fraternity, maybe?) So we may still end up paying them to build component X for us, but at least maybe we'll get a better deal on it. Or maybe we'll still get gouged, but in return their PHBs will let our PHBs win at golf, or they'll send us a swanky fruit basket that doesn't quite make it over to Engineering, or something useful like that.