Raw

Summary : this tells you how you can easily create your own dynamic "reading lists". All you need is a regular del.icio.us account.

Yep, I like the idea of Grazing the Web.

So, "Reading Lists" then, short lists of feeds someone finds interesting, that may change over time and are made public. They're promoted as being OPML things, but the idea's just another nice little app of data sync a la syndication. OPML sucks, but (as with RSS 2.0) if people are using it, and providing usable data* in this form it would be churlish to ignore it.

Ok, a quick way of collecting a list of resources is to use del.cio.us through its bookmarklet. So I'll find a few feeds (the feeds, not the blogs) and tag them "readinglist". I may want to have more than one list, so I'll add another tag, for example "tech".

Now right away there's an RSS 1.0 feed available at:

http://del.icio.us/rss/danja/readinglist+tech

In it are item resources, the feeds I've chosen.

In the RDF world this would be enough, and all the other metadata would be immediately available too. But to engage with the Really Simple world, extra processing is needed to change the format and throw that stuff away.

RSS is XML, so XSLT is available. Here's delicious-to-opml.xsl, 30 or so lines of info-stripping.

So, got feedlist in RSS, got transformation, all we need now is a processor. The W3C have one handy.

Paste in the URI of the XSLT, the URI of the del.icio.us RSS feed and click submit.

Voila: my tech reading list

(PS. So the OPML readers don't get confused, you probably want to map this to a TinyURL, e.g. http://tinyurl.com/b7d8w)

PS. Rowan tweaked OPod so the original URI now work - thanks!

Everything's dynamic, so every time I link to a feed and tag it "tech"+"readinglist", the data will be available in that OPML.

It's been noted that Reading Lists are best kept short, so the XSLT only pulls out the top 5 items.

This isn't particularly tested, and if anyone wants to use this I suggest copying the XSLT to your own server (it's only a static file, just upload it and use its new URI in the XSLT machine) - mine's a little unpredictable ;-)

But the result is apparently Valid OPML ** (but only after I removed the totally XML-valid XML declaration).

PS. The lllooonnggg URI causes the kind of parsing problems I was musing over earlier, the link above doesn't work with OPod. But I just remembered an encoding-style workaroundby the name of TinyURL.

Ok, so now who want write the Javascript to take the keywords, wrap up the URI as above and get the result of http://tinyurl.com/create.php?url=.

* this won't happen in the general case until there is a credible specification for feedlists in OPML

** see what I mean

Uh-oh, this post just showed up at tech.memeorandum, I'd better point to some background on the RDF references : Semantic Web Starting Points.

Turns out OPML from del.icio.us has been done before (using Magpie parser plus a little PHP) : vrypan.net.

[Danny]