On Wed, 2002-02-27 at 14:40, Tim Berners-Lee wrote: > If you say that a namespace file should *only* contain pointers, > then you may impede a perfectly goo application which has > a bunch of useful stuff in a single language > (e.g. RDF with a style sheet) and make such applications work at > half speed because you force an extra indirection. I don't think "half-speed" is plausible here. Downloading a document containing pointers and then downloading appropriate resources, while it has a one-time retrieval and processing cost, is also a one-time affair if an application is written with a vaguely sensible caching strategy. The benefits of acknowledging "there's more than one way to do it" this way seem to pretty drastically overwhelm the value of slapping one kind of "useful stuff" at a location. Multiple ways seem to be a given in the XML world, from CSS/XSLT to DTDs/W3C XML Schema/RELAX NG/Schematron to the different kinds of resource needs at creation time and (various kinds of) run time. -- Simon St.Laurent Ring around the content, a pocket full of brackets Errors, errors, all fall down! http://simonstl.comReceived on Wednesday, 27 February 2002 15:15:45 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:04 GMT