Re: Anomaly in the DAV:prop element usage

At 10:53 PM 1/6/00 -0500, Geoffrey M. Clemm wrote:
>
>   From: "Kaelin Colclasure" <kaelin@everest.com>
>
>   The use of a validating parser makes little sense in the context of
>   WebDAV properties -- because it is intended to expose a storage mechnism
>   for arbitrary element definitions (in the form of property values),
>   which themselves may or may not have a defined DTD or schema somewhere.
>
>Yes, that makes sense (although you do have to at least specify the
>namespace of those elements, even if you don't have the DTD's).  But
>in those cases where you do have the DTD's available (such as for the
>DAV:xxx properties), it does provide some cognitive dissonance to
>see an element like <DAV:creationdate/> in a PROPFIND request,

I agree that it is dissonant.   But there may be no better choice.

Let me make a suggestion - let he who thinks WebDAv should validate create
a DTD that will work.  The constraint is that it has to work even if there
is no DTD for some properties, and whatever mechanism you use should also
not have the bad network effect of needing to fetch the DTD with each
attempt to validate.  otherwise performance would be bad.  I am no XML
expert, so I can't tell whether this is easy or even possible.

I would certainly be willing to see changes to the DTD that would *allow*
webDAV XMl to be validating, if I didn't have to give up on expressivity
(ability to use properties without a DTD)

All things being equal, this would be better for DASL as well.  As one of
the authors I can say that I did spend some time trying to write a DTD for
DASL, mostly because I wanted to be able to check for dumb mistakes that *I
made.

Note also that RFC 2518 says that WebDAV's xml is not valid.  (but only in
passing in 17.7)  i think, given the amount of discussion that this has
brought up over history, it needs a passage in the WebDAV book of Why, if
not in the next Specification.

Received on Friday, 7 January 2000 04:24:22 UTC