Machine Tags of Loving Grace
One of the highlights of Refresh Edinburgh for me was listening to Dan Champion give a presentation on his new site, Revish. He talked through the motivation, planning and production of the site. This was an absolute joy to listen to and it was filled with very valuable practical advice.
Revish is a book review site with a heavy dollop of social interaction. Even in its not-quite-finished state, it’s pushing all the right buttons with me:
- The markup is clean, semantic and valid.
- The layout is uncluttered and flexible.
- The URL structure is logical.
- The data is available through microformats, RSS and an API.
There’s some really smart stuff going on with the sign-up process. If your chosen username matches a Flickr username, it automatically grabs the buddy icon. At the sign-up stage you also have the option of globally disabling any Ajax on the site—an accessibility option that I advocate in my book. Truth be told, there isn’t yet any Ajax on the site but the availability of this option shows a lot of forethought.
Also at the sign-up stage, there’s a quick’n’dirty auto-discovery of contacts wherever there’s overlap with Revish usernames and your Flickr contacts. This is very cool—one small step toward portable social networks.
One of the features dovetails nicely with Richard’s recent discussion about machine tags ISBNs. If you tag a picture of a book on Flickr with
book:isbn=[ISBN number], that picture will then show up on the corresponding Revish page. You can see it in action on the page for Bulletproof Ajax.
Oh, and don’t worry about whether a book has any reviews on Revish yet: the site uses Amazon’s API to pull in the basic book info. As long as a book has an ISBN, it has a page on Revish. So the Revish page for a book can effectively become a mashup of Amazon details and Flickr pictures (just take a look at the page for John’s new microformats book).
I like this format for machine tagging information related to books. As pointed out in a comment on Richard’s post, this opens up the way for plenty of other tagging like
book:title="[book title]" and
I’ve started to implement this machine tag format here. If you look at my last post—which has a whole list of books—you’ll see that I’ve tagged the post with a bunch of machine tags in the
book:isbn format. By making a quick call to Amazon, I can pull in some information on each book. For now I’m just displaying a small cover image with a link through to the Amazon page.
That last entry is a bit of an extreme example; I’m assuming that most of the time I’ll be just adding one book machine tag to a post at most, probably to accompany a review.
Machine tags (or triple tags) is still a relatively young idea. Most of the structures so far have been emergent, like Upcoming and Last.fm’s event tags and my own blog post machine tags. There’s now a site dedicated to standardising on some namespaces—MachineTags.org has a blog, a wiki and a mailing list. Right now, the wiki has pages for existing conventions like geo tagging and drafts for events and book tagging. This will be an interesting space to watch.