RE: Issue: ALLPROP_AND_COMPUTED

I am adamantly opposed to DAV:allprop.  In the context of
computed live properties, a client should never blindly ask
for all property values ... it should always first ask for
DAV:propname, and then use the subset that it can use.
The WebDAV versioning extensions explicitly allows a server to
*not* return the versioning properties in response to a
DAV:allprop request, so DAV:propname will be the only reliable
way of obtaining all the properties.  Finally, the fact that
PROPFIND/DAV:allprop is trivially replaceable with two PROPFIND
calls (the first being PROPFIND/DAV:propname) makes DAV:allprop
superfluous (in addition to being inadvisable).

I spoke to Yaron Goland earlier this week (about the ACL spec),
and he believes DAV:allprop should be removed/deprecated as well.
Protocol's are complex enough with important/required features.
A feature that is prone to misuse and is redundant is
a serious error, and should be removed in the next draft.

Cheers,
Geoff

-----Original Message-----
From: Jason Crawford [mailto:ccjason@us.ibm.com]
Sent: Wednesday, May 02, 2001 12:45 AM
To: w3c-dist-auth@w3.org
Subject: Re: Issue: ALLPROP_AND_COMPUTED




Hey everyone.  I didn't mean to silence everyone with my note below.   I
think Jim still needs opinions to figure out the disposition of this issue.
Do we have any more comments?

I'll propose that we put this off until we know if there is an actual
problem.  Perhaps someone who felt there was a problem could speak up and
add an opinion.  Do you still think there is a problem that we need to
solve?

J.

------------------------------------------
Phone: 914-784-7569,   ccjason@us.ibm.com


Jason Crawford/Watson/IBM@IBMUS@w3.org on 04/28/2001 02:03:04 PM

Sent by:  w3c-dist-auth-request@w3.org


To:   Greg Stein <gstein@lyra.org>
cc:   WebDAV WG <w3c-dist-auth@w3.org>
Subject:  Re: Issue: ALLPROP_AND_COMPUTED





<<
The issue isn't tossing allprop, but providing a warning to clients about
dealing with expensive computed properties.

Removing allprop does not solve the computed problem. Since allprop *is*
useful, then there is no reason to remove it.
>>
I agree with Greg.  We can still have this size problem with depth one and
depth zero.  And FTP's MGET can have a similar problem and we don't hear
people complaining about that.  Are we being overly concerned?  We as
authors of the spec can't know if this is a problem in advance either.
Sometimes it will be and other times it won't be.  Let's just deal with
what a client or server can do in situations where they feel it's a
problem.

So...

Is it sufficient for a client to simply disconnect/timeout?  Should we have
an error code that the server can use if it deems a request too expensive?
Or is there some other simple approach?  Or should we defer the issue until
version 2.0 if it then looks like this truly is a problem?

J.

Received on Wednesday, 2 May 2001 23:51:20 UTC