- From: Jim Davis <jdavis@parc.xerox.com>
- Date: Sun, 19 Oct 1997 19:52:06 PDT
- To: w3c-dist-auth@w3.org
2.1.1 Shouldn't this cite RDF as a metadata standard? 2.6.6. Example Either I don't understand the XML Namespace processing instruction, or the example is wrong. It seems to me that the Link, src, and dst elements need a D: prefix. (Or does XML have a lexical scope for providing a "default" namespace, so that Link within D:Prop is taken to be D:Link? If so, this should be explicitly documented, because it's an obscure feature. Likewise for the example in 2.8.3.6 Also, either I don't understand the description of Link, or there's a bug here. The values for Link are 1*Src and 1*Dst, it does not say that random other elements (in this case F:ProjFiles) are allowed. 2.8.4 PUT I don't understand the language here. I can't see from what's written here why a PUT could not work for a property. A GET returns text/xml specifying the name and value. So if I did a PUT that contained a name and value, what would the problem be? Seems like PUT would be very convenient, so why outlaw it? I'm sure there's a good reason, but I can't guess it from the spec. 2.8.5.4 PropFindResult The purpose mentions a SEARCH request, it should be PROPFIND Also the example uses the PROPFIND element not PropFIndResult. 3.10.6.1 In the example is the HREF for the Namespace PI *really* supposed to be Shttp:? If there's a reason to need to use SHTTP here, what is it? Otherwise it's just a confusing distraction. Minor notational inconsistency. At different places in the document listing reply codes, the descriptive strings are written in three different ways: * in parenthesis (e.g. 3.10.5) (possibly followed by a hyphen) * in string quotes (e.g. 5.2.7) * with no delimiter other than a hyphen This is really trivial I know but it makes the document just ever so slightly harder to follow. Finally, is case significant in XML? I hope not, because some elements are used as if was not, for example "Level". Even if case is not significant, the document should be written as if it were. (And, as I mentioned in a previous letter, we should use one consistent choice for writing elements whose names are compounds, e.g. MultiResponse.) Best regards ------------------------------------ http://www.parc.xerox.com/jdavis/ 650-812-4301
Received on Sunday, 19 October 1997 22:53:09 UTC