- From: Elliotte Harold <elharo@metalab.unc.edu>
- Date: Tue, 19 Jul 2005 13:29:06 -0400
- To: Misha Wolf <Misha.Wolf@reuters.com>
- CC: www-tag@w3.org
Misha Wolf wrote: >> Whatever the other pros and cons of QNames, it seems to me that you >> should use them properly and consistently if they are used at all. >> I'm not clear on why Dan's suggestion of: >> >> I'. <e demoAttr="_3" xmlns="http://example.org/demoNs"/> I don't think that was Dan's suggestion at all. Remember relative namespace URIs are deprecated and undefined. he was proposing something more like this: <e demoAttr="3" xml:base="http://example.org/demoNs"/> This makes a little more sense because the namespace mechanism is not being overloaded to handle something it wasn't designed for. This still has problems because xml:base has meaning for other elements and attributes as well that you might not want. Better still would be: <e demoAttr="3" base="http://example.org/demoNs"/> where the base attribute is defined only to have meaning for this particular application. Best of all would be to ignore this whole attempt to map these numbers to URIs somehow and just use the numbers themselves: <e demoAttr="3" /> If you want to define these in URIs somewhere , fine; but there's no need to clutter your XML documents with them. Not every datum needs to be or have a URI. Sometimes a number is just a number. Sometimes a number is an identifier from a space other than the Web. This URI everywhere insanity must stop. -- Elliotte Rusty Harold elharo@metalab.unc.edu XML in a Nutshell 3rd Edition Just Published! http://www.cafeconleche.org/books/xian3/ http://www.amazon.com/exec/obidos/ISBN=0596007647/cafeaulaitA/ref=nosim
Received on Tuesday, 19 July 2005 17:29:13 UTC