WireWorld » Hacks » Wirehead on Hacking » New web content engine, part 11

New web content engine, part 11

So I've been wanting to better topical-ize my content. And I also wanted to better support FriendFeed. So I spent some time on the syndication support...

One of my most popular presences on the web is Wirehead arts and I like how I can keep it fairly topical. But, overall, I have too much stuff going on to really give each and every interest of mine a unique site. So I'm experimenting with ways to group the useful content that one might be curious about into a unified clump with the sections. This has been changing over time. I started out with one blog, but now I ended up giving each section a blog.

The problem is that, while any section of the site could have an individual Atom feed generated for it, I hadn't made that available. Now I have on this site.

I also decided to attack the problem of FriendFeed. I've found that the best feature of FriendFeed is pretty much better handling of image browsing. In fact, I feel like it's better to view an image-oriented stream in FriendFeed than any other external reader... and many times, it's easier to keep track of people in FriendFeed than it is to keep track of them on Flickr.

I had gotten some vague pointers about supporting Yahoo's Media RSS extensions, so I kind of approached it from a few different directions. First, I wrote a Atom-to-RSS converter. See, Atom (and hAtom) form the basis for the formatting engine. I use XSLT and more recently using just bits of Ruby to translate the Atom and other XML markup languages to XHTML, so that I can change the look of a site without changing the underlying architecture. But I'm keeping that in reserve for the moment.

It turns out that you can just insert the Media RSS tags into an Atom feed and everything just works.

I also made a sample implementation of the Atom Media Extensions just for comparison. So I've got three ways to denote data on the thumbnail. The proposed Atom Media Extensions, the Media RSS tags, and the proprietary internal mechanism. I'll probably remove my properietary internal mechanism for thumbnail data as soon as I can get around to it.

Main concern, of course, is that Atom standards are always much better thought out and usually also usually just a few seconds too late to the party. Because I find it cleaner to use a link tag, but I'm sure it's too much trouble to retrofit everybody... especially given that Media RSS is neutral enough to work with both RSS and Atom.

But then again, it's tag soup all around...