I spoke with my friend about S3. They was hosting some images uploaded by users (something like imageshack), and Amazon just disabled account and deleted all data. Sorry, but I need something reliable, I wont study their API, code that for 2 weeks and use it for 3 months.
Currently, I am making analysis and specification of new opensubtitles, we are going to reprogram it all - also change design and so on. So all those things is for new OS.
Nowadyas I add some small things:
- in user profile user can list requested subtitles
- in subtitle detail page is fetched last 10 subtitles for same language...
Now I will try to focus of better handling duplicates subtitles and some old bugfixes.
I don't want to imply your friend hasn't got this right, but S3 is being used by pretty large sites, like SmugMug, that host pictures of all types. It also hosts every image in Twitter, which ranges from copyright infringements to porn.
Also, they wouldn't close and delete an files without warning, something they clarify in their FAQ. They would, at worst, give a warning to retrieve all files.
There is no API to learn how to use. There are tons of clients to upload files to S3, including command-line and PHP. And the URL is static. You can start using S3 in less than a day and if you keep a local copy of every file you can stop using it at any time.
If you're worried about leeching then that is something else, but not really related to S3. It's a pretty simple change that would have an immediate benefit, even if it worked for just a few months in the worst-case.
Anyway, you know better. I'd advise against the Second-System Syndrome. Planning ro re-design OS from scratch might be too much to do in one go.