On Sat, 20 May 2000, Clark C. Evans wrote: > It does seem to me that the primary hurdle now > is defining a 'reasonable' "default" base URI > for relative references.. or simply forbidding > relative references until <xml:base> comes along. Relative URIs break whenever you want to use them to solve more than one dimension at the same time. Reliance on a single xml:base attribute to control all URIs is just another step down this road, and the more-broadly people try to apply it, the more obvious it is that is wrong, IMO. xml:base seems as inappropriate for namespaces as the document URI. Similar things should use similar defaults. An element type's identity is not like the location of a document instance's picture files, as much as xml:base would try to pursuade us it is. Reasonable software -- even Microsoft software -- establishes different bases for different types of things. Often there is a whole dialog with a list of where different types of things are located in terms of a base location. I do not see how the web changes this requirement. Ray Whitmer ray@xmission.comReceived on Monday, 22 May 2000 14:16:25 UTC
This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 23:13:58 UTC