Hacker News Reader

Comment by zmmmmm | Hacker News Reader

That's crazy. Google can't tell the difference between legit downloads and "abusive" material? Google Code projects are already tied to a Google user account, what difference does it make if that user puts their abusive material on Google Drive instead of the downloads section of Google Code?

There are tonnes of code hosting sites going back as far as SourceForge and none of them ever came up with something like this.

There's something extremely weird about Google making this kind of developer hostile move. Why would I use Google Code now when there are any number of alternatives that have no problem with hosting downloads? It's like they are trying to drive us away. I hope it doesn't hint at Google Code being on the "sunsetting" list in the future.

View on HN

The difference is: you pay for Google Drive (including bandwidth).

To give you some perspective: I have a popular open source project hosted on Google Code. I host my binary downloads via S3 and since it's popular, it costs me $200+ a month in bandwidth.

If I used Google Code for hosting those binaries, Google would have to pay for that and why should they?

And that's a grey area example i.e. I would be using the service as it was intended to be used.

I'm pretty sure plenty of smart alecks simply abuse the service by putting and random stuff there to avoid paying for bandwidth.

This is a tragedy of commons: Google offered a free service, people started abusing it via excessive usage and Google chose to no longer offer that free service (as opposed to e.g. spend their resource on never-ending battle of policing abusers).

That is very much their right.

People download 2TB of binaries from you a month? What do you work on?

I'm guessing it's Sumatra PDF; author: Krzysztof Kowalczyk, hosted on google code, and fairly popular.

Thanks. At 4MB that's 500,000 downloads a month. Awesome work.

A little off-topic, but thanks Krzystof for Sumatra PDF. Easily the best PDF-viewer on Windows.

A little too much yellow. But yeah, I too find it to be the best.

Why not get a Hetzner box or something?

http://www.hetzner.de/en/hosting/produkte_rootserver/ex4

It's $60 for 10 TB/mo.

If it was purely a cost of bandwidth issue they could easily limit the size of downloads. Most projects on Google Code only need to offer downloads of a few MB in size to fulfil the basic need. It's really hard to imagine that being a significant cost to Google, and the convenience of having downloads integrated into the service is huge.

GitHub also made this move a while back (so it's not completely unprecedented), though I agree it's massively inconvenient.

Thanks .... I must have missed that github did this too (been a while since I tried to post a download there, but I have a number of active Google Code projects).

I wonder if it gives them some kind of safe harbour from the copyright industry. Perhaps a Google Code download is a "download from Google" vs a Google Drive download being a "download from that user".

> I wonder if it gives them some kind of safe harbour from the copyright industry.

FWIW, when I search, I see quite a lot of abuse reports for malware; this may have little to nothing to do with the copyright industry (against whom they already have a safe harbor for user-submitted content due to the DMCA.)

I imagine it could still be a kind of "safe harbour" issue.

If a user downloads malware from Google Code, it seems like there's at least some tiny chance they might try to sue Google for it. If they download it from a personal Google Drive account, it will be a lot harder to convince a court that Google had a responsibility for the content.

It's sad that your ignorant comment is at the top of this thread.

Actually after GitHub's move, I see many GitHub projects just provided their download link at Google Code. Basically they created a Google Code project just for the download section, even not bother to push the code there.

> There are tonnes of code hosting sites going back as far as SourceForge and none of them ever came up with something like this.

Actually, [github phased out their uploads/downloads features starting last December](https://github.com/blog/1302-goodbye-uploads) too. They didn't mean "abuse" as part of the reasons it was going away, but I wonder if it played a role.

Frankly, I hope they do drive developers away from google code. Every time I run into a project using google code to host they start off in the red in my assessment. Google code is a mess, it's always been a mess, I avoid it as much as I can.

Always been a mess? You must have never used sourceforge. There were times before github, and Google Code was great for what it was.

I use Google Code because it has a functioning issue tracker (try setting a priority at GitHub and sorting by it), and because you can have multiple repositories per project. The latter is very helpful as it lets you have one issue tracker, wiki, downloads etc for the project but still have things organised better behind the scenes.

Other services like Github could do the multiple repository thing too. In fact they already all do - having one main repository and a second for the wiki. Github said there is no way they would add this functionality when I asked.

Losing Downloads from Google Code is a big pain. That they never made a paid for private version is also annoying.

Github also no longer has this feature ...

> Google can't tell the difference between legit downloads and "abusive" material?

Obviously, they can, or they wouldn't be able to tell that there is an increasing frequency of abuse.

They could estimate the frequency of abuse just by taking a random sample of files and manually evaluating them. That's not inconsistent with it being hard to find all of the abusive files.

> They could estimate the frequency of abuse just by taking a random sample of files and manually evaluating them.

That still requires them to be able to tell the difference between abusive and non-abusive files, which ability was what was questioned in the post upthread.

Obviously, this wouldn't be an issue if they could in an efficient (and particularly, automated) manner identify all abusive files before they were exposed to the public.

Google is tidying up. Why have two storage facilities on an island when Google Drive is as good as it is universal.

It doesn't look like they were improving it either. I guess with Google it means only one thing...

Reply on news.ycombinator.com