Vivisecting Apples iPhone App Review Process – Insights from Inside | Hacker News Reader

Vivisecting Apples iPhone App Review Process – Insights from Inside

scottylabs.com

9 pointsmicrokernel7 comments

I asked a reviewer on the phone once. He said this:

There are 3 levels of review (Technical, policies and X). Random reviewers test different aspects. There is also 'Escalation', which is when the reviewer is not sure and the internal policies are unclear.

Furthermore, he says that the policies in the review team change every week sometimes. Examples of policies:

- No women with things in their mouths are allowed

- If a butt is shown, a reasonable amount of torso has to be shown

- You cannot make fun of religion, you can only talk about religion in a neutral manner

So the reviewers have a long and complicated list of rules they are to follow, and this rule list is dynamically updated when an app comes in with new type of offensive stuff.

Issues that stand out:

1. The location data is interesting but I have to question it accuracy.

2. Ten minutes of testing is way too little to spot subtle problems even in a simple app. Is this really the standard quality of review?

3. <cynically>Will Apple now re-review the app and reject it? Or will they reject later versions?</cynically>

2. We're assuming from that 10 minutes that there isn't a lot of semi-automated testing that does not involve a live internet connection, and I'm willing to bet theres a large amount of static testing against the binary. </respectful>

3. Why even hide behind a re-review? The fully covering NDA on the SDK has been lifted, but Apple still has contracts in place to use and publish apps. If publishing an unknown reviewer's (presumed) home address fits under a clause in the contract, Apple will most likely straight up revoke the app's approval. </cynical>

2. Our app is "In review" right now. For the first round our app reported back to us that it was active for about 55 seconds.

Concerning the location data I have configured it to be exact to around 100m (kCLLocationAccuracyHundredMeters) which in my experience is easily achieved. Maybe the use a misplaced skyhook WLAN to achieve that result?

Am I the only one a bit freaked out that this guy thinks it's cool to publish someone's address that he detected through their iphone? He's even said he thinks it's a home worker.

Or have I totally misunderstood what's going on here?

I just worry some poor tester is going to have hordes of disgruntled developers turn up on his door step demanding approval.

No, you're not the only one. I came over to say the same thing. Highly irresponsible and self-destructive. (He's going to have to interact with these folks on some level during subsequent approval applications.)

Reply on news.ycombinator.com