- From: Toby A Inkster <tai@g5n.co.uk>
- Date: Mon, 15 Sep 2008 17:05:21 +0100
- To: Martin McEvoy <martin@weborganics.co.uk>, RDFa <public-rdf-in-xhtml-tf@w3.org>
On 15 Sep 2008, at 16:17, Martin McEvoy wrote: > Talking of "hacks" why is this NOT a "hack" : http:// > wiki.digitalbazaar.com/en/HAudio_RDFa? it looks like a direct rip > from the Microformats wiki, That article is published by the same Manu Sporny who is editor of the hAudio spec. (At least I assume it's the same Manu Sporny. Manu Sporny has, I think, a name like mine, in that it can be considered an "inverse functional property".) > Old News I'm afraid Microformats do have a generic parsing model > just because you cant read it on a page somewhere , again this is > Work in progress by many members of the Microformats Comunity, > Brian Suda, Ben West, Myself are some of the people who spring to > mind who are actively working this Problem. more notably Toby > Inkster with this page http://microformats.org/wiki/parsing- > brainstorming .. If you read the blurb at the top, you'll see that the page covers parsing a good deal of microformat properties, but not 100%. For example, hCard N optimisation, the "tel" and "email" properties for hCard, "item" in hAudio, lots of crazy stuff in hResume regarding intermingling hCards and hCalendar for experience and education properties. (And that last one, I don't think there's any fool-proof way of handling - the hResume spec itself is way too fuzzy.) These are all outside my parsing algorithm. Even if you ignore those issues, the parsing mechanism is still far from generic in the same way RDFa is. If you look at part 4 of the document you cited, step 5.3.1, you will see that the parser needs to "find the value of element e". To do this, it needs to choose one of five different parsing methods. To know which method, it needs to know what "type" of property it is parsing: a plain text string, a datetime, a URL, etc. This requires special knowledge of each property - e.g. the parser needs to know that "fn" should be parsed as a string, but "photo" as a URL. Also, in step 3 it needs to know which properties may contain an embedded microformat. (e.g. an hCard may be embedded in hCalendar's "attendee" property.) This contrasts markedly with RDFa which can be parsed without special knowledge of any particular terms (except for a handful of unprefixed rel values which are grandfathered in, mostly from HTML4/XHTML1 - and the special handling they get is only very slightly special). -- Toby A Inkster <mailto:mail@tobyinkster.co.uk> <http://tobyinkster.co.uk>
Received on Monday, 15 September 2008 16:06:32 UTC