> In my experience working at Facebook, not a single thing you said was true. Well good for you. Either one of us has had an anomalous experience, or... > Peo

> In my experience working at Facebook, not a single thing you said was true. Well good for you. Either one of us has had an anomalous experience, or... > People plan and think before they code. "Planning and thinking" != design process. Of course people plan and think. I plan and think before I go to the grocery or hardware store. Of course it's only for a moment, and the typical process I've seen at FB is a lot closer to that than to the design processes I've seen elsewhere in nearly thirty years producing software. > People are rewarded for impact, not for landing useless code. Again, you're playing extremes against the middle to paint a picture more favorable than accurate. Very little code is totally useless. However, I've seen plenty of "impact" awarded for code that made a very tiny increment in functionality, far outweighed by the missing/misleading metrics or spurious alarms or outright bugs in something even the author knows will be replaced next half. Every day I encounter stuff that's broken as a result. This morning it was a distcc script that nuked a build, because someone's desperate to get their impact in before the end of the half. I can't count the number of times I've followed the trail back from an incident to a diff with "test in prod LOL" or some such for its test plan. So "carefully designed and tested" just isn't true in general. The common case is very far from that.