6 ms·

Can someone explain why both the Google Play Store and the Apple App Store give opaque explanations for rejections? Why don't they just tell developers what's w

Can someone explain why both the Google Play Store and the Apple App Store give opaque explanations for rejections? Why don't they just tell developers what's wrong and what needs to be fixed instead of pointing to some broad rule and forcing an interpretative song and dance?

nvrspyx 4y ago
Probably because giving a clear explanation would demonstrate how often and how badly app reviewers misinterpret their own rules.
- TheRealPomax 4y ago
  Bots don't really care about that, though.
mrmuagi 4y ago
I think they are using automation in detecting "violations" perhaps, given the other commentors did seemingly no-op updates and resolved things -- and any pushback would show the wizard behind the curtains, so they keep it vague as possible? Unless someone who actually works behind the scenes can chime in.
CobrastanJorji 4y ago
While I think it would benefit Apple or Google as a whole to explain their reasoning clearly, it'd make things harder on the app approving teams, who probably don't want to do this for a number of reasons: it'd take more time, it'd expose inconsistency/mistakes/incompetence (why was this okay last time but not this time?), and worst of all, it'd lead to the organic creation of a binding network of labyrinthine precedents.
ivanmontillam 4y ago
Whilst I don't agree with the opaqueness policy, I understand why it's in place. It's because it leaves them at risk of using previous precise answers to others as precedents for future cases. Being opaque allows them to not be too committed, and outside of situations like "but you accepted XYZ app which does exactly the same, I don't understand" You can still claim XYZ app does what you did and didn't get punished by that, but they can never admit it in paper/words.
- stevenkkim 4y ago
  Yeah, I think this is probably the right answer. I once worked on an iOS app that took about 5 cycles of "guess what's wrong?" -> submission -> rejection before we got it to pass. Each time, we asked "Can you just please tell us what to fix?" and each time the answer was "refer to rule x." So frustrating.
- MBCook 4y ago
  There is also the intent vs letter issue. Scammers will try to skirt rules by making sure to adhere to the letter of the law, even if it’s plainly obvious they are violating the intent of the rule in question. Rule 407b: no real money gambling apps Scammers: it’s not real money. It’s Linden Dollars. Or our company script. Or our new NFTs. You can’t reject us. Should the rule have to list out every single thing that could possibly used as currency? Because there are people who will argue that point.
  - three_seagrass 4y ago
    Yep, anti-fraud and anti-scam is an exercise in game theory. The more explicit you are with the rules, the more specific bad actors will try to weave around them. It sucks because it means the scalable, machine-assisted review process comes up with more false positives, but it's a two sided market and the marketplace owners care more about the users than the devs. I feel like the U.S. tax code is a good example of when you try to go more explicit with the rules.
- mkmk3 4y ago
  Maybe theres a way to functionally build up towards the same destination by collecting feedback and accepted changes. I don't know how that integrates with law, but is there something that influences the legal implications beyond what this kind of system would cover (provided a sufficient number of samples)?
- np- 4y ago
  Isn’t that unethical though? Preserving the right to make arbitrary decisions and favor some people over others? Most of the entire human history of conflict seems to be because of this reason of arbitrary unfairness.
jstummbillig 4y ago
The same reason they are not thrilled about support in general: Manual labor is expensive, and at the scale that Apple or Google operate, the amount of abuse and unwarranted complaints you receive at high traffic, low entry barrier, outward facing services is absolutely insane and can not realistically be shouldered without heavy reliance on tools and templates.
- fluidcruft 4y ago
  But how can it be hard to make a bot produce the input it flagged? I can see that if you are talking about humans asking them to compose detailed and specific reports has costs. But we're talking about machines, not humans. Basically all anyone is asking for is a stack trace so they can identify and debug the issue. I don't understand why that's unreasonable for a bot. It's like a complier that aborts with nothing more than "syntax error". Most people know compilers can be much more helpful than that.
  - jstummbillig 4y ago
    I assumed that the report is actually not bot generated, just software assisted?
Maxburn 4y ago
First off, there's no human involved in the first couple passes with this. Also bad actors could then use that detail to tweak things and get around policy. It's not a great answer either but possible.
artdigital 4y ago
In my experience AppStore Review is usually pretty direct. They tell me I violate this and that because of these things, and even include screenshots where the violation is
bliteben 4y ago
I'm pretty sure its because the reviewers don't speak english and thus can't type a fluent response in your language.
- gjm11 4y ago
  The (plausible) speculation in the article is that the reviewers "don't speak English" because, being machine-learning models rather than human beings, they don't speak any language at all.
counttheforks 4y ago
It's a power move.
CPLX 4y ago
Because some of the animals are more equal than others. A good analogy is dating. What they’ll allow you to get away with is highly dependent on how attractive they find you, from a business standpoint.
- jrootabega 4y ago
  Yes, it's unfair by design, while still giving the illusion of fairness. If they obfuscate the process and the rules, some parties will be falsely flagged. But the provider can also unfairly allow certain parties to violate the rules, while pretending that there is some secret rule that means they didn't violate anything. It's not the ONLY reason, but it's a big one.
avree 4y ago
I wouldn't lump Apple in with Google. In my experience they are extremely direct, and very quick to put a human on the other end of any message. Google, on the other hand, is a black box until the end - if you don't have backchannel contacts or a "friend" at Google, good luck getting any support.
naasking 4y ago
The more direct and transparent they are, the easier it would be for spammers to figure out how to game the rules.

nvrspyx 4y ago

Probably because giving a clear explanation would demonstrate how often and how badly app reviewers misinterpret their own rules.

TheRealPomax 4y ago
Bots don't really care about that, though.

mrmuagi 4y ago

I think they are using automation in detecting "violations" perhaps, given the other commentors did seemingly no-op updates and resolved things -- and any pushback would show the wizard behind the curtains, so they keep it vague as possible? Unless someone who actually works behind the scenes can chime in.

CobrastanJorji 4y ago

While I think it would benefit Apple or Google as a whole to explain their reasoning clearly, it'd make things harder on the app approving teams, who probably don't want to do this for a number of reasons: it'd take more time, it'd expose inconsistency/mistakes/incompetence (why was this okay last time but not this time?), and worst of all, it'd lead to the organic creation of a binding network of labyrinthine precedents.

ivanmontillam 4y ago

Whilst I don't agree with the opaqueness policy, I understand why it's in place. It's because it leaves them at risk of using previous precise answers to others as precedents for future cases. Being opaque allows them to not be too committed, and outside of situations like "but you accepted XYZ app which does exactly the same, I don't understand" You can still claim XYZ app does what you did and didn't get punished by that, but they can never admit it in paper/words.

stevenkkim 4y ago
Yeah, I think this is probably the right answer. I once worked on an iOS app that took about 5 cycles of "guess what's wrong?" -> submission -> rejection before we got it to pass. Each time, we asked "Can you just please tell us what to fix?" and each time the answer was "refer to rule x." So frustrating.
MBCook 4y ago
There is also the intent vs letter issue. Scammers will try to skirt rules by making sure to adhere to the letter of the law, even if it’s plainly obvious they are violating the intent of the rule in question. Rule 407b: no real money gambling apps Scammers: it’s not real money. It’s Linden Dollars. Or our company script. Or our new NFTs. You can’t reject us. Should the rule have to list out every single thing that could possibly used as currency? Because there are people who will argue that point.
- three_seagrass 4y ago
  Yep, anti-fraud and anti-scam is an exercise in game theory. The more explicit you are with the rules, the more specific bad actors will try to weave around them. It sucks because it means the scalable, machine-assisted review process comes up with more false positives, but it's a two sided market and the marketplace owners care more about the users than the devs. I feel like the U.S. tax code is a good example of when you try to go more explicit with the rules.
mkmk3 4y ago
Maybe theres a way to functionally build up towards the same destination by collecting feedback and accepted changes. I don't know how that integrates with law, but is there something that influences the legal implications beyond what this kind of system would cover (provided a sufficient number of samples)?
np- 4y ago
Isn’t that unethical though? Preserving the right to make arbitrary decisions and favor some people over others? Most of the entire human history of conflict seems to be because of this reason of arbitrary unfairness.

jstummbillig 4y ago

The same reason they are not thrilled about support in general: Manual labor is expensive, and at the scale that Apple or Google operate, the amount of abuse and unwarranted complaints you receive at high traffic, low entry barrier, outward facing services is absolutely insane and can not realistically be shouldered without heavy reliance on tools and templates.

fluidcruft 4y ago
But how can it be hard to make a bot produce the input it flagged? I can see that if you are talking about humans asking them to compose detailed and specific reports has costs. But we're talking about machines, not humans. Basically all anyone is asking for is a stack trace so they can identify and debug the issue. I don't understand why that's unreasonable for a bot. It's like a complier that aborts with nothing more than "syntax error". Most people know compilers can be much more helpful than that.
- jstummbillig 4y ago
  I assumed that the report is actually not bot generated, just software assisted?

Maxburn 4y ago

First off, there's no human involved in the first couple passes with this. Also bad actors could then use that detail to tweak things and get around policy. It's not a great answer either but possible.

artdigital 4y ago

In my experience AppStore Review is usually pretty direct. They tell me I violate this and that because of these things, and even include screenshots where the violation is

bliteben 4y ago

I'm pretty sure its because the reviewers don't speak english and thus can't type a fluent response in your language.

gjm11 4y ago
The (plausible) speculation in the article is that the reviewers "don't speak English" because, being machine-learning models rather than human beings, they don't speak any language at all.

counttheforks 4y ago

It's a power move.

CPLX 4y ago

Because some of the animals are more equal than others. A good analogy is dating. What they’ll allow you to get away with is highly dependent on how attractive they find you, from a business standpoint.

jrootabega 4y ago
Yes, it's unfair by design, while still giving the illusion of fairness. If they obfuscate the process and the rules, some parties will be falsely flagged. But the provider can also unfairly allow certain parties to violate the rules, while pretending that there is some secret rule that means they didn't violate anything. It's not the ONLY reason, but it's a big one.

avree 4y ago

I wouldn't lump Apple in with Google. In my experience they are extremely direct, and very quick to put a human on the other end of any message. Google, on the other hand, is a black box until the end - if you don't have backchannel contacts or a "friend" at Google, good luck getting any support.

naasking 4y ago

The more direct and transparent they are, the easier it would be for spammers to figure out how to game the rules.