Hacker News Reader

> That's a safe place for them, right?

Fuck no! Double fuck no!

Do people actually think that? That's beyond ignorant/naive. Data in the hands of a for profit corporation, safe? Did you read the terms of service / limitation of warranty? The short version is "fuck you and your data". It's not safe from loss at the whims of market/share holder needs/etc. It's also not safe from being compromised either by internal employees or external hackers. It's also not protected by most of the Bill of Rights (those mostly apply to government vs people not business vs people). It's easily supeoned, if the business doesn't just turn over the data on request.

Safe is encrypted on multiple devices geographically (and preferably jurisdictionally) separated.

The lowest acceptable level that can still be considered safe (a very loose definition of safe) is; a copy in plain/open formats on a device I physically control/hold(my desktop harddrive), and a copy somewhere else. Mom's house, Dropbox, yahoo mail, etc.

To us engineers/founders/people with knowledge on how the software business works, it might be obvious, but not for the average Joe. Google (or Yahoo) will be a name brand and they'll hold high expectations of them, especially with Google's "Do No Evil" mantra.

That's "Don't Be Evil," actually. I think that's a pretty important distinction in tone.

> Do people actually think that? That's beyond ignorant/naive.

Yes, people actually think that! Have you met 90% of your (ok, Google or Yahoo's) customers? Barely anyone tells them to worry about saving their data locally and these companies actively ask you to load your data into them listing safety as a feature. "It's in the cloud!" blah blah

You can call people ignorant/naive when Google/Yahoo/Hotmail et al. display prominent "Your data is going away in 6 months AND HERE'S HOW TO GET IT." Emphasis on the "here's how to get it" otherwise they're suddenly at a loss of how to get their data back.

Most "normal" people don't assume there's a moratorium on data. Data is forever. You put it on a floppy disk, it stays there forever. All you can do is lose the disk. Same thing with flash drives. Same thing with online services.

The different here is, as you said, safety is pushed even harder with the cloud. There's nothing to lose anymore. The people who are savvy enough to realize that storage can and does die are now being told that people far more competent at it than them are managing backups and such. Why would you think your stuff would go away? Who is telling you that it might? I don't think it's ignorant/naive at all.

I do think that what Google is doing sucks, because I don't think it costs them much/anything when they're handing out 7GBs to anyone with a Gmail account. There's orders of magnitudes less Groups than Gmail users, and Groups was a great way of organizing small project teams privately, in a way that Sites isn't (eg. mailing list, files, pages all in one place).

I'm looking forward to the day when we have a widely used p2p archiving network like OceanStore or Free Haven. Then we can stop worrying about this stuff.

I can see how a Dropbox-like P2P filesharing system could be beneficial for groups, but how would a P2P network benefit individuals trying to safely and securely archive personal documents?

Encrypt the documents. Put them on a distributed hashtable, using something like Reed-Solomon codes to distribute them over the network, so you're only taking about twice as much total storage as the document itself but you're very resilient to nodes disappearing. Maybe add some kind of resource exchange system so you pay for the storage by participating in the network.

Google the projects I mentioned and you'll find a lot more detail on these sorts of ideas. Both projects intended to build permanent, uncensorable storage at global scale. It's not a totally solved problem yet but there's been a lot of interesting engineering.

My Master's Thesis covered exactly that. I'm currently working on commercialising the idea, so you can expect a "HN - Review my startup" within a year :)

If you don't mind linking to your thesis I'd love to read it!

Sure, but it's in Swedish. If you still want it I can e-mail it you.

Don't worry, I'm sure the NSA has it on a server somewhere along with the rest of our communications.

If the data is important don't entrust it to someone else. It doesn't matter if it is Google Groups, Facebook, or Yahoo mail.

While that is a good prescription for individuals, what the OP is getting at is that trust is a part of a brand, and if you could measure it, deleting data is penny wise and pound foolish for these companies.

Also, don't trust Seagate, Western Digital, Maxtor, or any other hard drive manufacturer..

(the serious point - if it's important, don't have a single point of failure. It doesn't matter if it's hardware or cloudware)

Totally, 100% understand that people make that kind of assumption. It makes sense if you don't look at it closely.

However: free account. Expecting a free anything to store your anything indefinitely is expecting the sky to rain diamonds.

Expecting a warning about such actions, especially for a low activity account, is not unreasonable (though still: free). And > 1 year warning is about as good as you can hope for, and would be utterly wonderful. But companies don't plan to stop things that far ahead and just absorb expected loss for over a year, so it's unlikely to occur, in which case 6 months is very generous, and about a minimum for people's expectations.

> However: free account.

Indeed, you get what for pay for.

"...is expecting the sky to rain diamonds."

Lovely analogy! :)

It does happen, though: http://geology.com/nasa/diamonds-in-space.shtml

...not that anyone should expect it. ;)

Awesome find, I'll have to keep that one handy for other similar analogies :D

> Personally, I think deleting data like this--especially when the storage requirements are so tiny--is unforgivable.

I think the problem is that people haven't adjusted to the new ways of keeping data. They assume their free Yahoo account is theirs, the way their closet is theirs. You know, if you leave a box of pictures in your closet, it'll stay in your closet barring a catastrophe.

Online services are available to serve a certain set of goals and objectives. If those objectives change, your stuff might get thrown out. It's more like asking a hotel to let you keep a box of stuff behind the desk for a while than it is like your own closet. They might keep it for a really long time, but if the hotel changes management or remodels, your stuff could get tossed. Especially if you're not regularly staying at the hotel.

It's not even "you get what you pay for," because paid services shut down too. In this era of cheap storage, there's no excuse to not make multiple backups of key data. Ideally at least one physical copy of important things. Actually print the emails out, put them in a binder, and put it in your closet. Keep one copy at Yahoo, one at another online backup service, one on your PC, and one on an external harddrive. That sounds like a lot, but it's not. Once you get a sensible backup policy and figure out what's important, it doesn't take very long to do backups.

It's a sad story you shared, but not unforgivable. Someone's keeping a box of your stuff behind the desk at a hotel. You can't count on that box to stay there. Now, the problem is people haven't realized that yet. But slowly, it'll become widely known and understood culturally. If it's important enough, keep multiple online and offline backups. Don't count on anyone to hold on to your stuff for you, because if management changes or the place gets remodeled, it might not be there any longer.

"especially when the storage requirements are so tiny"

Hardware is not the only contributor into the cost of keeping data available on the web site. There is electricity cost, servers administration cost, cost of keeping up with the updates, legal cost of dealing with requests to put data on your servers. Finally, there is opportunity cost -- the same team that manages these Group features could be doing something more useful. That's why Google is not maintaining unsuccessful applications.

There is another reason to forget old data: reducing noise. Usually newer data is more useful than the old one, so it forgetting old data reduces noise and makes it easier to find more useful information.

I hope your friend managed to get her data back. That makes my blood boil.

Do you believe it's reasonable to expect an account to not be deleted after 14 months of inactivity? That's not a short amount of time.

Particularly for a free service. It's not reasonable to expect Yahoo to archive everyone's email forever. If you want something with more guarantees, you should probably expect to pay something for it.

OK, you are not getting his point. You don't have to explain it to us, all of us here are aware of it.

But, if there are users that don't understand it, than it means that they didn't deliver the message in a proper way. And obviously there are users that didn't get it (I know several cases for erased email accounts).

Obviously, she didn't. It's important to set the expectations clearly up front with the client.

I lost my old yahoo to, went to look at it one day for the nostalgia factor and it was gone, not a big loss but it's always fun to look back at old chats and emails and see how much you have changed over time.