Back in college, all of the computer science lab machines used shared NFS home directories. It wasn't unusual to see "NFS server not responding" on your console when the network was experiencing a hiccough.
There was a running joke about how crafty students would swallow SEGFAULT in their assignments and just endlessly print out "NFS server not responding" in a delay loop. "Crash? No, it's just the network. Look!"
[edit] I just realized that he was debugging a third-party application. The narrative is a bit hard to follow.
I can't quite figure out why someone so smart was doing something so stupid.
The way to solve crashes is not to crash, unless you're erlang and functional and build the software to recover fully -- but that's not something remotely feasible to do for an application written in a standard imperative language.
Otherwise trying to continue on blindly results in a hang like the one he experienced at best, and a data corruption at worst.
My takeaway from this is that next time I encounter a really challenging problem, I will close my eyes and imagine having a conversation with House, M.D.
Man I'd be pissed if I wasted two days on getting my finger behind some idiot's stupid 'hey let's mask errors and hope they go away' tricks.
Also, while the story itself was sort of interesting, the format (with the chords and the House and whatnot) made it hard to follow - why not just describe the issue rather than making up schizophrenic discussions?
Why? Because that's Ridiculous Fish's writing style. You may find it hard to follow but I (and presumably others) find it amusing and read his articles because of it :)
Well, there could have been some irascible bug in a previous version of the OS that the app kept running into, and this was the programmer's method for getting around that. Or, it could have been some really sneaky bug in the program itself, something that was beyond the programmer's ability to figure out, so they did this instead and followed the industry standard advice to "just ship it". Or, maybe it's part of some framework somewhere, because somebody thought it was a good idea.
It's a bit of a stretch to jump to "some idiot"; most of us are "some idiot" often enough.
Sure, but all of that still doesn't make it OK. I've shipped things with non-optimal solutions - and if somebody ever had to waste two days like this on those things, they'd have every right to be pissed. And yes some of these things does make people idiots - if it was beyond the programmer's ability to figure out, or if a 'framework' thought it was a good idea, that makes them idiots.
Because the author of the article works for Apple, and it sounds like they discovered a compatibility regression in a particular piece of third-party software and assigned him (or her) the task of figuring out exactly why it was broken.
Microsoft is famous for hacking up their OS in order to keep quirky or poorly-written third-party software running on newer releases. I don't know if Apple would go quite that far, but it sounds like they care enough to spend expensive developer time investigating these problems.
Back in college, all of the computer science lab machines used shared NFS home directories. It wasn't unusual to see "NFS server not responding" on your console when the network was experiencing a hiccough.
There was a running joke about how crafty students would swallow SEGFAULT in their assignments and just endlessly print out "NFS server not responding" in a delay loop. "Crash? No, it's just the network. Look!"