W3C home > Mailing lists > Public > w3c-dist-auth@w3.org > October to December 1996

Re: Prelim. DAV spec.

From: Henrik Frystyk Nielsen <frystyk@w3.org>
Date: Thu, 31 Oct 1996 17:31:57 -0500
Message-Id: <3.0b36.32.19961031173152.00975100@pop.w3.org>
To: Larry Masinter <masinter@parc.xerox.com>, slein@wrc.xerox.com
Cc: w3c-dist-auth@w3.org
At 12:34 PM 10/31/96 PST, Larry Masinter wrote:
># If there are cases where variants are NOT computed on the fly, what should
># happen when someone creates a new version of the resource (by editing
one of
># the variants)?  I assume the user would normally prefer for the other
># variants to be brought into synch automatically -- for new versions of them
># to be generated from the one he submitted.  Do we want to do anything to
># support this?
>It depends on how the variants are computed. If they can be
>automatically generated, you might want to update them automatically,
>but the state of language translation software, for example, isn't
>really up to automatically regenerating the French page from a new
>version of the English page, for example.
>Since there's not a general solution, the question is whether we need
>or want to do anything in the protocol. I don't think so.

Yes indeed. Whether representations are generated on the fly or not is an
implementation detail - not subject for protocol definition. The same is
the case for CGI-script outputs - you can't see on a URL what it produces
when you poke it (except from the cases where you peek into the URL and
find tokens like "cgi-bin" or "htbin" :-))

Not in response particularly to this mail but on the discussion on whether
a resource has a name or not, it goes beyond my mind to understand how I
can see something on my screen but I can't give it a name. There is no such
thing as nameless resources. HTTP allows two situations to occur:

1) The server does send a Content-Location with the response in which case
the client _may_ use this location for accessing this resource, for example
to replace it with a new version.

2) The server does not send a Content-Location in which case for all the
client knows the Request-URI _is_ the name of the resource. The client
_may_ want to use any entity tag to make sure that when it uploads a new
version in order to make sure that the server can identify the previous


Henrik Frystyk Nielsen, <frystyk@w3.org>
World Wide Web Consortium, MIT/LCS NE43-356
545 Technology Square, Cambridge MA 02139, USA
Received on Thursday, 31 October 1996 17:34:25 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:01:09 UTC