I didn't really take last week's Twitocalypse that seriously, but as you probably know by now, it turned out a little worse than expected -- we'd been told that Twitterrific (and, we assumed, most other Twitter apps) would be fine, and of course, as Craig Hockenberry explains on his blog, things ended up not-so-fine. Desktop app developers, of course, could publish updates as quickly as they could code them; iPhone developers were in a different situation.

When the Iconfactory's app stopped working, most people (including me) got an API error all weekend. Craig found the bug, then he and his team were able to leverage their contacts at Apple Developer Relations to help expedite the release; in short order, an update was pushed out to the App Store. I downloaded it yesterday, and can tell you that things are fixed... at least until the numerical limit on Twitter's tweet identifier raises its head again (or the Newton flips out, but that's another story).

Hockenberry also has ideas about how to keep issues like this from happening again. Not the actual issue of a variable overflow (that will undoubtedly happen again at some point, on Twitter or any other API that scales way faster than anyone expects it to), but the issue of iPhone apps needing a quick fix. He says that Apple should give every developer a number of "incidents" -- situations rarely used, in which a high priority fix can get sent out to apps in major emergencies. He says, and it's true, that for most developers, it's not a question of if you'll need to send out a critical fix, it's a matter of when. And support by Apple, obviously limited to one or two instances per developer, would help developers, distributors, and consumers.

Of course, it's up to Apple, and it's not like they've smoothed out the approval process so well already that they can start adding wrinkles to it. But clearly, given that the Twitterrific update went through quickly, there's room for exceptions to be made.

[via DF]

Oh, computers. Every time we think we've figured you out, you come up with another number-related disaster for us to fear. This time around, it's the Twitpocalypse -- apparently Twitter decided to use a unique identifier in their API for each separate tweet that goes out, and come sometime this weekend, the number will reach 2,147,483,647 and keep going. Which just happens to be too much for the variable that holds it, and that means that many of your favorite external Twitter clients, from Tweetie to Twitterific, to all of the other magic Twitter apps out there, could go down in a ball of numerically-ignited flame.

The good news is that just like the Year 2K panic, this one is more or less unfounded. We've already spoken to our friend Ged Maheux of Twitterrific, and he says that the problem has already been fixed in the latest version of the software. We'd imagine that most of the really big Twitter apps have already addressed the issue, and taken care to make sure they'll still work (either that, or they'll get on it ASAP). Plus, a Twitter outage wouldn't really be anything that might cause riots in the streets (or different from the usual downtime, actually -- rimshot!).

But it is funny and strange that as long as we have people who can't foresee the future (so, all of us) programming computers, we'll keep running into problems like this. If your favorite Twitter client goes out on Saturday, just find a basement to stay in, stock up on shotgun shells and water, and put your head between your legs -- help is on the way soon.