Next message: Clemm, Geoff: "RE: Versioning TeleConf Agenda, 4/17/00 (Monday) 2-3pmEST"
From: Edgar@EdgarSchwarz.de
Date: Wed, 19 Apr 2000 17:33:20 +0200
Message-Id: <200004191533.RAA17664@msheas02.msh.de>
To: ietf-dav-versioning@w3.org
Cc: Edgar@EdgarSchwarz.de
Subject: Re (2): DAV:revisions property for a workspace resource
> <jra>
> Well, if the information were calculated on the clinet, it would still have
> to be sent to the server to be useful. If the server had to decode the
> deltas in order to respond to requests for document revisions, then the
'requests for document revisions' is very generic. I was just talking
about deltas for CHECKIN. But if there was a single delta format naturally
it also could be extended to CHECKOUT. Interesting idea :-)
> server would have to know the format of the deltas and the contents of the
> resources it was required to reconstruct. This is the problem. Different
> resource types will require different mechanisms for expressing deltas. For
> example, C source code, XML documents, and binary files would all have
> different deltas. So introducing a new resource type on the web would
> likely require a new delta which would require the WebDAV server to be
> updated.
> </jra>
Why do you need different mechanisms to express deltas ? C source code,
XML documents, Word documents, HTML documents all consist of bytes like
'binary files'.
So why can't you use one delta format for all these resource types ?
This format can be based on byte differences and support deltas for
resources of unlimited size.
BTW, I agree that mechanisms to show differences in a way that make sense
would be different for different resource types. But that's not the point.
Here only generating the new revision is important.
Edgar
--
edgar@edgarschwarz.de http://www.edgarschwarz.de
* DOSenfreie Zone. Running Native Oberon. *
Make it as simple as possible, but not simpler. Albert Einstein