I realize part of the benefit of it is to share files - but what about things I want to make sure are private?
I see photos and public when I first sign in - are these both public folders? How do I store things, lets say my resume I want to be able to access from anywhere, but want private?
Other than that, it appears to 'just work' amazingly. Having built a somewhat similar backend to a system like that, it really is quite a feat.
Edit: Since the startup I was at is now defunct - email me and I may have some a good potential market for your product (very slightly modified)
15. How public are Public links? How Private are my other files?
* Any file in your Public folder is accessible to anyone who can guess your public file link and the path to the file. No one can browse the directory, though.
* Any file in your Dropbox outside of the Public folder is by default only accessible via computers linked to your account or to you via the web interface.
* If you give out a shareable link to a portion of your Photo Gallery (from here), people who know the link will be able to browse any photos in the album you linked from and any sub-album.
Thank you! If you guys can pull off the "Just Works" (and I think you have; I haven't had a chance to play with my beta yet) this has great mass market potential in my opinion.
For example, I'm working on some code with my cousin; we have Subversion etc.; but his PC is the host and we're never on our PC's synchronously so using Dropbox could solve that problem. Also, my folks manage a large apartment complex and they have files they share across multiple PCs at work and at home, I think DropBox would be simple enough for them to use without causing me stress as the de facto family IT guy.
1) Get a Linux version out, it shouldn't take too long as you developed the thing using Python (not sure what OS specific bindings you have)
2) Add support for LAN connections, when two boxes sit on the same LAN there is no reason for them to wait for S3 to be updated, get the update done locally and then propagate to S3, this will save you bandwidth and shorten latency.
In regards to #1, a simpler solution for me is to just use sshfs. It hooks into the kernel's fuse module and allows you to mount any SCP-able location as a local filesystem. I don't know if it supports any kind of caching or not, but it seems like it would be quite easy to build that on top of it.
Fuse also has support for things like curlftpfs, so if you can't get ssh access somewhere, as long as you have a host with ftp access, you can do the same thing.
At the risk of getting grabbed by the MPAA...I want to use a system like this to globally access my media archive, which includes all my legitimately acquired games and music, but also a decent amount of pirated material. Obviously storing copyrighted material without a license on a server in other people's control is a very bad idea. This would be a very useful thing to do though; maybe in a different world where the copyright system has been reformed... I often find myself at a friend's house wanting to see some movie I have on my disk, but I never bother to bring my portable HD unless it's planned.
I have about..50 gigabytes of images of games that I own, i think? Having access to those anywhere would be really nice. I would run my private server to do this kind of thing anyway, but this networking solution would make things a lot easier. However, I wouldn't want anyone to instigate a raid based on what to the casual observer seems like a fat load of illegally acquired copyrighted material. This would be a problem even if I only uploaded what I own.
And then there's the problem of storage cost, 1.5$/10GB/month turns into a considerable number very quickly. So I suppose you intend this kind of thing for the small stuff; images, documents, small-to-medium music collections etc.?
That's my problem with all these online storage things. It's too expensive to save my 150 gigs of FLAC/mp3 on their service. Not even S3 is cost effective.
And I do not want to waste all the time to rerip all those CDs if my computers gets stolen (again), so I finally had to settle for rsync and Dreamhost @ $12/mo for 500 GB, but Dreamhost is far from fast, so it's not perfect.
I'd like to keep it up to date, so just a harddrive is inadequate, since all my trusted friends and relatives are 3 hours away. I'd need a whole computer, on 24/7 with some ports forwarded and stuff.
I don't want to impose the... 8 or 10 bucks extra electricity per month a 24/7 comp uses.
(Also, there's the nice side benefit of having a webserver to host my random forays into web design, personal blog, etc.)
Congrats! What's ironic is that I was lucky to record a video of Drew coding an early version of Dropbox last March, but as I didn't have a painless solution for backing up, I don't have a copy of it anymore. :)
Gave it a swirl just now. I have tested Wuala, AllMyData and a couple of other similar services before and DropBox does stand out in a general look&feel as well as the implementation details.
IMO it has two major problems at the moment -
* On Windows, if DropBox is not running, it causes Windows Explorer to lag like hell; entering a folder on a local disk takes seconds and the Explorer basically plays dead at this time
* There is no detailed description of the security model - what's encrypted/authenticated where and how exactly. While some (most ?) people don't care about this sort of thing and tend to take devs' word for it, the model needs to be disclosed to allow independent review and evaluation. Otherwise it's just a proprietary crypto, which should be assumed broken until proven otherwise.
Very interesting. I actually built something like this myself (I was going to call it satchl) and I eventually stopped after looking long and hard at the online storage market and deciding that it wasn't worth the effort.
What I wanted to do with satchl was provided the ability to drag and drop files onto a drive/mount point which would also appear as a web page (a 'desktop' if you like). There would be a couple of versions of the desktop available: a public one and a private one with a single password for making changes. I was targetting a family that needs to share files across the world, or a small company. I had fun writing all the Ajax stuff for the virtual desktop...
I figured that the key to this was ease of use, and by that I mean really, really easy to use. I didn't think that I'd be able to get it to be so easy to use that my mother could use it.
And therein lies what the DropBox guys figured they could do. I didn't they could either until I saw them demo it. It's really quite slick, and is a completely different product than the usual "online storage" product. It's not yet another "web-based file manager"...it's a file system that works seamlessly over the Internet. With revisions.
So, just curious, why haven't you guys bought dropbox.com from whoever owns it now? Seems like a smart move because if I tell people to "try dropbox", they'll go to dropbox.com first.
Techcrunch says 15k of funding so if they've had an angel round it hasn't been announced. It could also be a small angel round. At this point, it's probably not worth spending a large percentage of their bank account on the domain. The dropbox.com domain owner also might want a number that is close to their entire bank account balance.
So I used dropbox as a beta user for close to a year now... the thing I really like about it is the "set it, forget it" feeling it affords me--it's there when I need it, and stays out of my way when I don't. What I would really like to (eventually) see is some kind of a dropbox server I could download to e.g. my SAN (external hard drive attached to my router) that would work the same it does now, i.e. over the internet.
Does the Windows version work with applications that don't understand shell extensions (e.g. DOS/Win16 apps, Cygwin &c)? For all the time that I've been searching through syncing programs, I haven't yet found one that presents as a drive/real folder (ala FUSE).
How well does this work with large binary files? (I'm not interested in using this to share DVD rips, I work with legacy blob databases at my day job at the salt mines which are usually several hundred MB in size)
i haven't gotten a chance to test this yet, but how does dropbox do conflict resolution? ie: i'm offline at one place, make a change, offline at another place, make a change, go back online at both places.
for text files, i suppose this could potentially implement svn-like text merging. how would that feature feel, from a design perspective? unintuitive? error prone? or excelent?
edit: damn, this is a really cool app. goodbye emails to self.
If two users share access to a file (either through sharing the same account or within a shared folder) and a conflict between their two versions appears, the Dropbox server retains only the first version to fully make it to the server.
The machines still in conflict with the server will get copies of the file with a suffix like "(Drew's local copy 1-25-2008)." Users can then manually fix the conflict and remove the extra copies of the conflicted file.
Plus, when you read the FAQ, it's even cooler than it sounds at first: automatic photo albums, public/shared/private settings, automatic versioning, secure sending, etc. This is one sweet product.
http://getdropbox.com/beta/yc50