Re: Auto-Download tags/function?

lilley@v5.cgu.mcc.ac.uk (Chris Lilley, Computer Graphics Unit)
Errors-To: listmaster@www0.cern.ch
Date: Thu, 19 May 1994 23:32:43 +0200
Errors-To: listmaster@www0.cern.ch
Message-id: <94051922301963@cguv5.cgu.mcc.ac.uk>
Errors-To: listmaster@www0.cern.ch
Reply-To: lilley@v5.cgu.mcc.ac.uk
Originator: www-talk@info.cern.ch
Sender: www-talk@www0.cern.ch
Precedence: bulk
From: lilley@v5.cgu.mcc.ac.uk (Chris Lilley, Computer Graphics Unit)
To: Multiple recipients of list <www-talk@www0.cern.ch>
Subject: Re: Auto-Download tags/function? 
X-Listprocessor-Version: 6.0c -- ListProcessor by Anastasios Kotsikonas
sanders@BSDI.COM said:

> "Daniel W. Connolly" writes:

>> I don't understand how that allows me to represent, in a document,
>> a link to a specific format of a document.

> First of all, you don't want people creating links to specific formats,

Isn't that a bit prescriptive? I would prefer to see this statement expressed 
thus:

"You don't want to force authors to link to specific formats every time, 
although that should be possible if that is what the author wants to do."

> that defeats the whole point of format negotiation.

In the particular case that started this thread (check the subject line) 
defeating format negotiation was precisely what the original poster wished to 
achieve, in a particular situation.

Also, I would say that format negotiation is not in place at the present time. 

OK, contentious I know, but: browsers do not support it, users cannot configure 
it, servers don't support it and information providers don't work that way.

(Just an observation. I am dealing with the majority here - if you happen to 
know of a proof of concept browser/server pair that does, my point still 
stands).

Furthermore, discussion on this list has shown that the issue of automatic 
format conversion, were it to be implemented, is complex. There are lots of 
special cases we don't want to debar just so that whgat is percieved as the 
common type of transfer will work correctly.

> Stored links should
> be to the most varient form of the URI except in very special cases

I believe that it is those very special cases that are the present topic of 
discussion.

> Right now
> a cacheing proxy must use some kind of heuristic and that should possibly
> be addressed.  I suppose the proxy would need some way to ask the
> server about it's conversions. 

I would refer you back to the recent discussion on image format conversion in 
the context of cacheing proxies, where this is discussed.

> That could get real nasty though
> (how good, how fast, from what to what, etc).  Ugh!

That was the general consensus, I believe, along with the realisation that 
'good' was a vector not a scalar. There also seemed to be some agreement that 
the best person to know what conversions and degradations were acceptable, on a 
case by case basis, was the author of the document who understands the actual 
meaning of each document, image, etc.

I can readily see that linking to non-specific versions would be the right thing 
in the default case, where a well chosen set of defaults will provide adequate 
results most of the time, and allowing the author to specify a particular file 
in cases where it matters should be a supported option. Equally, allowing an 
author to override the type of a known file for particular reasons is also 
catered for in Dan's suggestion.

Perhaps someone could post an example where the URI spec helps clear up the 
various example situations that have been posted in previous threads, such as 
medical images, prepress images, etc while still doing the right thing in 
typical situations where converting, say, an iris RGB to a Q=50 JPEG is a good 
and proper thing to do.

--
Chris Lilley
+-----------------------------------------------------------------------------+
| Technical Author, ITTI Computer Graphics and Visualisation Training Project |
+-----------------------------------------------------------------------------+
| Computer Graphics Unit,        |  Internet: C.C.Lilley@mcc.ac.uk            |
| Manchester Computing Centre,   |     Janet: C.C.Lilley@uk.ac.mcc            |
| Oxford Road,                   |     Voice: +44 61 275 6045                 |
| Manchester, UK.  M13 9PL       |       Fax: +44 61 275 6040                 |
| X400:  /I=c/S=lilley/O=manchester-computing-centre/PRMD=UK.AC/ADMD= /C=GB/  |
|  <A HREF="http://info.mcc.ac.uk/CGU/staff/lilley/lilley.html">my page</A>   | 
+-----------------------------------------------------------------------------+