Re: Target-Selector value
jamsden@us.ibm.com
Tue, 5 Oct 1999 10:17:37 -0400
From: jamsden@us.ibm.com
To: Tim_Ellison@oti.com (Tim Ellison OTT)
cc: ietf-dav-versioning@w3.org
Message-ID: <85256801.004E8D83.00@d54mta03.raleigh.ibm.com>
Date: Tue, 5 Oct 1999 10:17:37 -0400
Subject: Re: Target-Selector value
The revision id and label namespaces don't overlap, so there's no need to
distinguish them. The ambiguity would come from a label (or revision id,
although this is less likely) that is the same as a workspace name. It could
also result if we allow the Target-Selector to be any revision selector (my
preferred name) including activity, configuration, etc. I'm not suggesting that
we should do this now, but I would like to make sure we don't rule it out.
Basically I would like any element of a workspace RSR to be able to be used as a
revision selector so we can eliminate special cases/rules. Ambiguity prevented
us from doing this, but Tim's idea could be used to resolve that ambiguity.
It seems a little funny to use the protocol part of a URL as a namespace
qualifier though.
Tim_Ellison@oti.com (Tim Ellison OTT) on 10/04/99 03:00:40 PM
To: ietf-dav-versioning@w3.org (ietf-dav-versioning)
cc:
Subject: Target-Selector value
The Target-Selector is used to select a particular revision of a versioned
resource. It can be:
(a) a workspace URL
(b) a revision identifier
(c) a label
since the value could be one of three types, and it may be ambiguous which
is being sent by a client, I propose that the value is always a URI and that
we reserve two new schemes
revisionid: and
label:
where the protocol specific part of revisionid: is an opaque variable length
sequence of octets, and the protocol specific part of label: is a UTF-8
encoded String.
Examples of valid target selectors would, therefore, be:-
(a) http://foo.com/mywksp
(b) revisionid:3
(c) label:pick me
ref:
http://www.w3.org/Addressing/schemes.html