W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > April to June 2004

Re: Re (2): Status of RFC2518Bis

From: Stefan Eissing <stefan.eissing@greenbytes.de>
Date: Wed, 7 Apr 2004 10:18:32 +0200
Message-Id: <28C79960-886C-11D8-BCE0-00039384827E@greenbytes.de>
Cc: w3c-dist-auth@w3c.org
To: edgar@edgarschwarz.de


Am 06.04.2004 um 22:58 schrieb edgar@edgarschwarz.de:
> Julian Reschke <julian.reschke@gmx.de> schrieb:
>> Locking has been optional in RFC2518, so there shouldn't be any 
>> problem
>> whatsoever having a RFC2518bis-minus-locking going to draft. In fact
>> it'll be easier because locking is the area that needs most attention.
>>
>> That being said, what *is* your position regarding separating locking
>> into a separate document?
> I wonder at the moment what we win by locking ?

Well, it's certainly a much needed feature for editing in-place. Almost 
all servers
implement it, not the least because MS-Office clients use it.

Although DeltaV has much to offer in this area, there is currently a 
lack of
deltav-aware clients for general editing purposes. So, locking in WebDAV
will keep its usefullness.

As to the difficulty of implementation: the obstacles are there for new
implementors as the specification in RFC 2518 is weak. So most issues
are discovered and fixed during interop(s). There is a general consensus
among implementors how to handle them on the server side. It's
documenting this consensus that's lacking.

> Roughly speaking it's for avoiding collaborative workers to damage 
> other peoples
> work.
> But it seems defining and implementing locks successfully is very 
> difficult.
> OTOH there is DeltaV which also is a means (Albeit more complex, but 
> servers and
> clients are coming) to avoid collaborative conflicts.
> So who needs locking ? Me definitely not.
> I would be happy to implement DeltaV and the underlying RFC2518 stuff 
> without having
> to implement locks.
> So doing a RFC2518bis-minus-locking would be fine with me.
> And whoever wants locks because he doesn't like DeltaV can work on a 
> lock spec.
> This also would help BIND, which wouldn't need to say anything about 
> locks
> anymore.
> Locking has been a pain in the ass for years. So let's get rid of it 
> in RFC2518bis !
> This really could help us to make progress because many disagreements 
> will
> disappear.
> Just my 2 cent without giving it a lot of thought.
>
> Cheers, Edgar
>
>
>
>
Received on Wednesday, 7 April 2004 04:19:46 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:44:06 GMT