So last night, I started making good on the implied threat contained in yesterday's idle musings about Amazon S3. I've started building S3Ajax, an AJAX-based wrapper for Amazon's S3. Between S3, the OPML Community Server API, and what will hopefully come out of the Atom Publishing Protocol - I think web-based filesystem APIs have the potential to be a Next Big Thing.

Imagine if Amazon S3 offered filtered RSS feeds of new files uploaded to a bucket. And, imagine if that RSS feed listed those new files as BitTorrent enclosures, with a few item elements fed from metadata headers (ie. title, description, etc). Instant podcasting.

Also, imagine if Amazon connected an XSLT processor to S3 like they do with product searches. You could turn a bucket listing into an RSS feed with an XSLT uploaded to S3 itself.

Archived Comments

  • Les, Please make clear the license and known issues with S3Ajax. I may use it as a basis for the GM API...

    Cheers, Jeremy

  • Hmm... license - one of these days I should settle on a license for my stuff, more specific than Share and Enjoy.

    For S3Ajax, as it stands right now I'm happy to call it Public Domain. I'd like it if my name were mentioned somewhere in relation to it, but I won't raise a fuss if it isn't. If you can find a use for it, you're welcome to have at it!

    At some point in the future, I may attach something like an MIT or BSD license to it. The SHA1 implementation I'm using is under MIT, which is the lynchpin of the authentication bits.

    As for known issues, it's had about 4 hours' exploration and testing so far and seems to work fairly well. It's had no stress testing, but I've been able to authenticate, upload/download content, and set metadata attributes through headers. However, it could certainly use some test cases.