RE: REPORTS

> From: John Hall [mailto:johnhall@evergo.net] 
> 
>> This was discussed in the London IETF meeting and reported in
>> the minutes 
>>
http://lists.w3.org/Archives/Public/ietf-dav-versioning/2001JulSep/0193.html
>> As I recall, there was little objection in the room to this
>> report as it was considered a relatively simple recursion for 
>> the client and/or server to implement.

> This is NOT a simple recursion, at least not on my server.
> Version-tree is just a propfind -- for client and server.  
> Expand-property is a complete rewrite -- for client and 
> server.  I probably spent an hour or two on version-tree, and 
> a client would need less.  Expand-property is at least 2 
> orders of magnitude more difficult (at least on my server). 

The expand-property report is precisely a recursive PROPFIND
(i.e. when the server hits a specified property value that contains an
href, it recursively invoke PROPFIND on that href (just dumping the
results to the same output stream as nested XML). How is this not a
simple recursion?

And note that no rewriting of your client is necessary ...  your
client does not have to invoke expand-property unless it choses to do
so, since it is just a performance optimization (i.e. decreases the
number of roundtrips).

> I do know you will find several refereneces to 
> expand-property by me -- all indicating that there was 
> absolutely no plans for implementation of that optional 
> report in my server.  I don't see how moving this can be 
> considered 'by consensus'.

Note that "consensus" does not mean "unanimous".  Note also that
"SHOULD" (unlike "MUST") tells the client that it needs to be prepared
for servers that chose not to implement it, so a server is in
compliance even if it does not implement a SHOULD.

Perhaps you could provide more explanation about why you can't just
invoke your PROPFIND code recursively on your server to produce the
expand-property report?

Cheers,
Geoff

Received on Wednesday, 5 September 2001 18:10:44 UTC