Hacker News new | comments | show | ask | jobs | submit login

I imagine (and I'm not speaking for my employer here) that it has to do with cost: if the easiest way to backup your email is by using IMAP or POP, it's essentially an incremental backup mechanism. If the easiest way to backup your email is by clicking a link which contains all your email in a .zip file, it would be a non-incremental backup system and its cost would be much higher, because people would be far more likely to repeatedly download the same emails.

Maybe there could be a few options? Download all mail, download by month, download by year? Seems like they could solve that problem pretty easily and would be very awesome.

Offering different zipfiles for months/years wouldn't solve the problem. "The problem" is that when you have hundreds of millions of users, the easiest mechanism needs to be an efficient mechanism. Adding links for months/years would still leave the easiest mechanism as the least efficient: I know that I, personally, would continue to download the "all mail" zipfile just to ensure that I didn't miss anything. I cannot imagine my mother or grandmother doing anything different.

I don't think it would. That would then be something I would have to set up reminders for and then what if I forget it for a few days? I would have to remember when I last backed up and set that as the starting date. Of course google could remember that for me but we're now getting closer and closer to what IMAP already does better.

When I first set up offlineimap, it downloaded my entire account history on the first day - I'm not sure how that would be less expensive than a single zip file to download on that same day too.

I imagine that most people would download their entire e-mail history, stick with what they know, and re-download their entire history week after week.

Sure, but it would be easy to wrap that inside a Google 'download my data' tool' only downloaded the diff of all your data (not just email) and provided that in compressed form.

I'd prefer the direct zip, sure, but I'm just saying that it's not inconceivable.

You can argue that's it not technically inconcievable but if anyone were to implement it this way, it would likely become a nightmare of usability and broken expectations.

If I'm 'downloading' or otherwise 'exporting' data from any service, I expect it to be all the data. Not a diff wrt the last time I clicked the button.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact