Categories
iphone

Perennial problems of launching an iPhone app…

We launched an app this week that’s been exhaustively tested:

  • 2G, 3G, 3G S, 4
  • iPod 2/3, iPod 4
  • OS 3.1, 3.1.3, 4.0, 4.1, 4.2
  • 10+ different human testers
  • Apple’s own testing (App-review process)

…and yet we’re getting 1-star reviews on the app store, citing “this app crashes on startup”.

Meanwhile, according to Apple:

“No crashes for this app”

It’s always possible that someone’s writing shill reports, but I find that very unlikely. I find it far more likely that there’s some bug somewhere that’s affecting a narrow range of existing users.

Yet there’s literally no way for us to uncover this. Apple still provides no means for a developer to contact users. You are forbidden from sending replies or messages to the people who review your app.

Those people are assuming (we know, because they say this in their review!) that we know about the crash and can do something about it – we can’t. We cannot do *anything* until one of these people contacts us and gives us some crash info. Or allows their phone to upload crash reports to Apple.

Sigh.

5 replies on “Perennial problems of launching an iPhone app…”

Is the app memory-intensive? Have you tested on jailbroken devices? Jailbroken devices often have more services running in the background, so less memory is available for apps. If you are borderline, you can be fine on a stock device but run out of memory on jailbroken devices.

Only thing I can recommend about getting in touch with reviewers is to prompt them to communicate in the app description.

As for crash logs, if you use something like Flurry, you can log crashes over the web as they happen.

Eventually, we managed to reproduce *a* crash (whether that’s *the* crash is unknown – likely, but not guaranteed).

We’ll update, and see what happens…

Nope, sadly. Although we’re definitely not using OF on that project.

Right now, it’s waiting with the client to decide what to do next. We did find a few oddities in the live build that can’t be reproduced in the adhoc build – but without the client pushing for it, we didn’t dig very deeply.

Comments are closed.