W3C home > Mailing lists > Public > www-webdav-dasl@w3.org > January to March 1998

comments on first draft

From: Jim Davis <jdavis@parc.xerox.com>
Date: Wed, 11 Mar 1998 21:09:11 PST
Message-Id: <3.0.3.32.19980311210911.0081d5c0@mailback.parc.xerox.com>
To: www-webdav-dasl@w3.org
This is great so far.  There are two major things missing: sorting and
discovery.
There should be a way to ask for a sort at least by property value and by
'rank' (which should be a property that is defined only as result of doing
a search.)  Sorting should also be optional.

To enable interoperability, it is essential to be able to ask a resource
not only what grammar it supports, but what are the searchable and
selectable properties, what operators are allowed on each searchable
property, and what sorts are allowed.  This can best be done by defining
some DASL properties on the arbiter

I will shortly send out an edited version of this draft with proposals for
how add these.

In addition:

1) in 5.3 the scope's syntax is inconsistent with definition in 5.5

2) Why do we use the uri tag and not href

3) 5.6 redefines prop (defined by dav already)

4) 5.8 I find it problematic that contains mandates case-insensitive
comparision.  I would rather have this specified by e.g. an atttibute.
What about stemming?  soundex?  

5) 5.8 Rather than having contains take an optional prop element I would
rather define an explicit tag that refers to the text rendition content.

6) There are other operators widely supported on full text systems
including 'begins with' (right truncation) and 'adjacency'

7) I don't understand that 5.10 is trying to say about searching XML as
plain-text.

8) I don't understand what 5.11 is saying about variants.  First of all,
can someone tell me where variants are actually defined?  I don't see them
in RFC 2068 nor in WebDAV.  Second, is paragraph three saying that the
server should return separate records for each variant that matches, or one
record if ANY variant matches?

9) Need a note (at least a promissary note) about DASL and access control.
Are all readable properties also selectable?  Searchable?

10) Are we going to define any new status codes?

11)  I think literals should be be typed, either by a tag
 <literal><number>7</number></literal>
or perhaps by an attribute:
 <literal type="number">7</literal>

12) phrase is undefined in the DTD

13) We need literals at least for the 'unknown' value and probably also for
TRUE and FALSE.  You need a way to say e.g. 'where author is either unknown
or contains 'een'.

Again, this is a great start.
Received on Thursday, 12 March 1998 00:30:34 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Sunday, 22 March 2009 03:38:03 GMT