Well there's the problem
"as a single human I don't scale well"
Well duh. Google has billions in the bank and can't be arsed to hire more than a single person to manage app rejections ?
Popular Android app Podcast Addict has been suspended from the Google Play Store, apparently for mentioning COVID-19. "Are you for real Google. Can you please check your own Google Podcast App?" the frustrated developer said on Twitter. The app claims to be the top podcast app on Android and is certainly very popular, with …
We had a problem with Google a few years ago and it was very sobering, how you can best deal with Google.
One of their servers suddenly started spamming our Internet connection (10mbps connection being stuffed with around 100mbps of traffic from the Google server) - possibly a misconfigured server that was supposed to send information to another part of Google and our IP address was mistakenly entered?.
Tried contacting them at abuse@ and webmaster@ addresses. Automatic answer that this account is not read and the message had been automatically deleted!
Okay, call them... Over 20 minutes bouncing back and forth over their automated telephone system, before being spat out and left with a dead connection.
Twitter didn't bring any help either. In the end, we had to contact our commercial ISP and get the Google IP address blocked at their border, which was a subscription service costs nearly 3 figures a month! In the end, we just accelerated our move to a new ISP and got a new IP address, it was quicker and easier to arrange than to get Google acknowledge a problem. For all I know, the Google server is still spamming the old IP address.
I had a similar sounding issue with Microsoft at one point.
I logged an issue, got a standard boilerplate acknowledgement saying the usual "we'll look at this issue when hell boils over" and left it like that. I kept getting flooded with shit for another few weeks.
I then set up a complicated fudge of scripts to automate recording the incoming problem, crafting a support message and attaching the requested debug into and logging it with Microsoft Support and filing their email acknowledgement. I then felt a twinge of conscience before setting it live.
It didn't take long to get an email from Microsoft support diplomatically inquiring if I would kindly desist from logging issues at a rate of about 30 per second as they'd got the message that the issue needed to be investigated. With it being late on a Friday afternoon I had a much more serious crisis of conscience, resolved by heading off to the pub with everybody else.
The problem was apparently fixed over the course of the weekend as there were no more reports being generated on the Monday morning. I do hope their helldesk system allowed programmatic closing of tickets; ~30 reports per second is around 1800 requests per minute and 108,000 per hour, which is something like 2.59 million individual reports over a 24 hour period.
Brilliant. It's only when *they* become inconvenienced do they respond.
As for the original poster, well, the guy who lodged the appeal was a developer advocate. It wasn't his job to lodge the appeal. His job is to make sure relationships between developers and Google remain well and good.
This is the way of the world, unfortunately.
I do to this day wonder how they dealt with it from their side though; I have never met a servicedesk system that would allow any easy way of dealing with 1800 requests a minute. Even if you could group select and bulk delete say 200 in a go, you'd still have to do that nine times a minute to keep up with the incoming volume of problems being logged.
Mind you, if they flagged them all as "complete" after dealing with the problem then it would have had an interesting and statistically significant effect on their average response times.
If you have access to the database, then sure.
But in a large multinational organisation, i'd say that they wouldn't have wanted the helpdesk deleting any tickets that messed up their SLA's, so wouldn't have had access to this easily. They'd have to have done it that way I suppose though, can you imagine doing it through user level tools? The mind boggles.
"Well duh. Google has billions in the bank and can't be arsed to hire more than a single person to manage app rejections ?"
Indeed. People love to talk about how monitoring apps, videos on Youtube, and so on isn't easy because there are just so many of them. But the fact is that it is incredibly easy - you just need to employ enough people to actually do the job. Is that more expensive than waffling about automation while refusing to actually address the problem in any meaningful way? Almost certainly. Might that mean some business models might not actually be profitable? Quite possibly. You do not have the inherent right to a profitable business.
The thing is though, about that "profitable business" thing. Personally, I just ignore Google's profitability and look at Google's parent company Alphabet where all of the money actually ends up. Alphabet makes ten billion profit a quarter. That's forty billion profit a year. Numerically, 0.1% of that profit would be forty million.
In the UK you could hire people for 15k to do a simple job of "is this objectionable" quite easily since a lot of people get paid less than this to do far more unpleasant jobs.
At 1.5x the cost which is the rough ballpark used to account for tax, pensions, office space, equipment etc that'd cost 22.5k p/a per employee so for 40 million you'd get ~1,777 people doing the job for around 0.1% of the companies yearly profit so it's not really even a case of the business model not being profitable.
"At 1.5x the cost which is the rough ballpark used to account for tax, pensions, office space, equipment etc that'd cost 22.5k p/a per employee so for 40 million you'd get ~1,777 people doing the job for around 0.1% of the companies yearly profit so it's not really even a case of the business model not being profitable."
Sure, but if you're going to look at the entire company's global profit, you also have to think about the cost of doing the moderation globally. Maybe you could handle the work in the UK with 2000-odd people. Then you need another 2000 or so for each other country in Europe, and for each similarly-sized one around the world. Probably 10 times that for a bigger country like the US. That means it's easily 1% of profits just moderating a single country, and probably at least 20-30% in total. And that's assuming minimum wage is enough to cut it, when they're (certainly Facebook at least, I think Google as well) already having issues with lawsuits and claims for psychological treatment even with the minimal amount of people involved. Could they do it and still be profitable? Maybe, I don't think we have anywhere near enough information here to do a real calculation. But it would certainly be a significant hit to profits even if it's not enough to kill things off entirely.
It's also worth noting that the Google/Alphabet thing isn't really worth worrying about. Google accounts for 99.4% of Alphabet's revenue. Alphabet isn't a real company, it's just a shell to collect Google's stuff together under a different name and funnel some of the money to speculative projects (like self-driving cars). Any distinction between the two names is completely meaningless; it's all either just Google, or Google wearing a false nose.
Too many apps to moderate proactively?
Too many posts to look at by humans?
Too many video uploads to check them all?
It's simple - employ more people. Maybe the ones your rent-seeking efforts are putting out of work.
And if you cannot manage your scale, then don't scale!
Which leads me to another conclusion, this time for devs.
Stop treating the tech giant devs and their issuances as the cream of the crop and the founts of techie excellence. They prove daily that they don't know how to scale, or how to do AI/ML, any better than anyone else. There is a myth associated with them, and it's undeserved. They are good but no better than the rest.
So much of the techie world is full of little emperors with no clothes.
Small companies are great because they have people that are motivated and care.
As they get bigger they get filled by people that are just there for the paycheck.
Small companies can't afford to have everything checked by humans. Facebook sees the increasing demand for moderation as a barrier to new competitors, therefore protecting its market share (albeit with a cost).
That is part of the problem, Google and Facebook are considered big companies based on revenue and users, yet they are tiny companies, in terms of employees per customer, they have scaled their tech without ensuring that there are enough people in place to deal with the expansion.
The same is true about the law. When they are small, they circumvent the law and they scale up, circumventing the law until the authorities come a-knocking and suddenly their systems have been scaled up so much that it is "impractical" to comply with the law - i.e. it would be prohibitively expensive to obey the law, it is cheaper to keep paying the fines and lawyers than it is to actually tackle the problem.
They should follow the law from the beginning and they should scale their solutions for staying inside the law as they grow, that way the revenue model grows taking into account compliance - but that isn't sexy, because revenues will be lower than they could be - instead of getting "sticker shock" when they suddenly have to actually comply.
There is usually a warning period and if it is a change in policy there will be multiple emails before it comes into effect that explain how to be compliant.
Having developed stuff to work with adwords, this isn't new behaviour. They've been doing this sort of sudden change with little to no warning for at least a decade.
Yes they'd send emails, but I never once remember receiving one that explicitly warned me that an upcoming change was going to set a bin fire again that would then be my problem to fix again. Even more fun when they drop documenting API calls. That was a fun few weeks trying figure out camelcase/bem/strongly typed APIs with only a vague error response to go on.
Which is about as clear as receiving a letter for a traffic violation with a fine.
No mention what rule was broken or how the traffic was violated.
So nothing to give you an idea of just needs remedying nor how to challenge an injust accusation.
Everyone operating within the Google/YouTube environment has signed up for these problems. If a small part of Google's business is no longer needed then it can just die and they won't care.
I noticed with Amazon that they were duplicating their orders for each address you held on the system. They sent out 3 printers but only charged me for one. This happened more than once and to my friend as well. Once we realised that Amazon did not want the items back and we're not going to charge, we ended up with quite a few things.
They have fixed the bug now but can you imagine a small business allow such a thing to go unchecked for months?
If software is pre judged to be deliberately abusive then you'd want it blocked at once. I presume there are naughty people out there whose actual job is to put malware into trusted channels and who spend all day just doing that. Giving them the benefit of the doubt is over generous.
I absolutely love Podcast Addict, even the endlessly convoluted menus.
More and more I look for alternatives to Google. Whereas the biggest problem used to be Google shutting down a Google app or feature that you need, now it's perfectly usable third-party apps that stand to disappear.
I for one look forward to Huawei developing an Android and Google free phone OS.
Perhaps it's time for developers to ensure that they provide the installation APKs on their own website as well as via the Google Store, so that there's a way for users to obtain the latest versions even when Google muck things up^Hfail to scale again.
This has the useful benefit of allowing their software to be sourced by those who either don't have access to the Google Store (eg Huawei users) or who wish to avoid it (eg privacy advocates).
I agree as long as there is an easy way to check the apk is signed by a particular key, or the key that the version it is replacing was signed by.
I have a few phones which do not have the Google Play store. Some of them are running old versions of apps that are useful to me. But if I find a new version of the app I want to be sure it has been created by the same people as my existing version and not substituted with a trojan.
If the APK is downloaded via a secure https connection to the company official site, the key is only going to be verifable across a https connection anyway, so that sets the upper limit of verification as https.
Every company should publish the official APKs on their own site downloaded across https. Otherwise they're locking themselves into Google, and Google can play companies off against each other for ad money.
Worse, if you ever try to disable Google Play Store, it will uninstall everything you've bought in the store to force you not to uninstall it.
>Perhaps it's time for developers to ensure ...there's a way for users to obtain the latest versions
Its more than just the latest versions.
Do a factory reset and you may find that many apps can't be installed because the latest version doesn't run on your particular version of Android.
However, with a little digging around you can often find a site - of unknown repute - that has a few old versions, which you can try and find one you can install and then update.
Podcast Addict is the best app I have. The developer has put a link to the APK file online, link in his Twitter feed, that doesn't let him make any money from sales or advertising that is tied to Google/Android though. Googles attitude to developers and appalling communication is outrageous. Might be considering a Huawei for my next phone.
The Google Play Store is a commercial venture. If they think something won't make them money they can remove it. Censorship is the wrong word to use for a marketplace, this is business pure and simple.
When they fuss your search results, that's censorship, when they ban an app from their store, that's their prerogative.
It's Google's market. They have complete discretion over what's in their store. If you based your business on them allowing you access you took a massive risk.
Lots of downvotes but sabroni is right.
If you're part of google's ecosystem, you're there because it benefits you commercially. If google says jump and you don't ask how high, tough.
It may be difficult to exist outside the dominant culture but plenty of people have been pointing out the problems with monocultures forever. If you want to make use of google, facebook or microsoft, play by their rules or play somewhere else.
Better yet, do your best to bring down monocultures. In the end, they only benefit themselves.
1. Publicity seems to have bought it back and more notorious than ever!
2. Never heard of it. Just tried it. Somewhat better than Google's app (listened to the latest More Or Less, and The Inquiry, over breakfast).
The takeaway? Don't depend on Google's services; be promiscuous with your app stores (Huawei will thank you).
I assume things like this get flagged because the code or some metadata related to the app references COVID-19/coronavirus and the reference is spotted during some static analysis. Things like podcast names and descriptions are only going to be available at runtime via some feed or API call.
If my assumption is correct, then surely there is no need for the code itself to mention either of these terms? Even things like categories have the option to be externally driven to get around the filters.
Biting the hand that feeds IT © 1998–2020