On 20/1/09 11:42, Toby Inkster wrote: >> Another problem with having to define prefixes upfront is that it makes >> streaming production of output really hard -- if you don't know ahead of >> time which namespaces you'll need you will be forced to include *all* of >> them (whatever "all" means here) in the header. > > That is true. It's not a problem I've had with RDFa per se, but in > programmatically producing Atom, where the feed's<updated> date is > required to come before the<entry> elements are listed. But the feed > <updated> date is calculated by taking the most recent of the entry > <updated> dates, so after writing the entries, I need to go *back* and fix > the feed<updated> date. > > I wouldn't suggest creating a situation where similar annoyances will > occur with RDFa. This is a strong argument against requiring CURIE > prefixes to be declared in the<head> - the ability to declare them close > to where they are used is very useful. And essential for copy/paste integrity, as well as bloggability, use in wikis, etc. cheers, DanReceived on Tuesday, 20 January 2009 11:11:47 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 20 January 2009 11:11:48 GMT