Re: Wish List for New Spec

Robert Hazeltine (robert@janis.virago.org.au)
Sun, 7 Jan 1996 13:41:28 +1100 (EST)


From: Robert Hazeltine <robert@janis.virago.org.au>
Message-Id: <199601070241.NAA12193@janis.virago.org.au>
Subject: Re: Wish List for New Spec
To: lilley@afs.mcc.ac.uk (lilley)
Date: Sun, 7 Jan 1996 13:41:28 +1100 (EST)
Cc: rhazltin@zeppo.nepean.uws.edu.au, www-html@w3.org
In-Reply-To: <279.9601021050@afs.mcc.ac.uk> from "lilley" at Jan 2, 96 10:50:05 am

Chris

lilley wrote:
> 
> Robert Hazeltine wrote:
> 
> > 1.  A reaonable means of adequately identifying a text document giving it a 
> > "front page" (meta information seems to be the best canditate);
> 
> Does that mean the META element fulfills your requirements, or not?

There is a gap in the specification of adequate identification of a 
file.  There is no equivalent of a title-page and bibliographic details 
(on the reverse of the title-page) which usually specifies imprint 
information, or even a half-title that is part of printed material.

I believe that this information is even more important with electronic 
publications than traditional ones because of the lack of visual cues 
from a book or magazine.  Without it, it like picking up a book with the 
cover and title pages missing.

Programmers have RCS and the like, include version and title info in the 
file title to identify application distribution files.  There is no 
equivalent for text files.

I think that it would obviate a lot of problems if there were _minimal_ 
requirements to identify the document title, author, imprint and 
language.  The LINK element does tie files together and 
language identification is used in other elements.  The TITLE element 
does not necessarily reflect the publication title and, I would hazard a 
guess, it nearly always identifies a component file.  However, they do not 
provide that missing piece of information.

Other than that, the author should optionally use META element to 
demonstrate her/her own ingenuity. :-)

The need for such a mechanism is more apparent at a larger site such as a 
library and making jumps to external resources.  I certainly do not go along 
with the idea espoused by some librarians that META information is a 
technique "to catalogue on line" - it should be simple and limited to what 
the author/editor decides, not what is added later by some library or IT 
specialist.

I have noticed that Dan has started to add RCS information as part of 
the citation of files/documents and this does illusatrate the problem 
precisely.

What I am suggesting is a modest enhancement for HTML3.  Perhaps the 
inclusion of this would prompt enriched editors that also manage document 
revision information.

> Yes! Plus, the preservation of some semantics - the LaTeX method of 
> describing pictures of equations is only slightly more useful than
> the current HTML method of GIF/XBM pictures of equations. 

Any suggestions for the interim on putting up papers with equations, etc?

> Not only asked for (over the last couple of years) but worked on and 
> close to consensus. See:
> 
>   http://www.ics.uci.edu/pub/ietf/html/
>  
Thanks for this information.  The examples I gave originally were for an 
academic environment, but the need for bi-lingual or multi-lingual 
documents is just a applicable to everyday affairs, in Australia at least.

Rob...
-- 
Robert Hazeltine                    http://www.zeta.org.au/~grove/tvhome.html
grove@zeta.org.au                   http://www.nepean.uws.au/library/
r.hazeltine@nepean.uws.edu.au       http://www.nepean.uws.au/dimps/