OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help



   RE: [xml-dev] SGML on the Web

[ Lists Home | Date Index | Thread Index ]

OTOH, I was very happy to see the transformational 
approach.  Doctor Goldfarb had listed that as a 
sine qua non in one of his presentations very early 
in the game.  I've never used XSL-FO.  I don't do 
high quality print systems and have been someone 
who campaigned for simpler printing requirements 
since well before the web.  One had to be around 
for the 1000dpi DoD requirements.  It cost a lot 
and made no sense in a world of cheap laser printers.  
300 dpi was just fine, and even less when the 
common screen resolution was 72-75 dpi.  In short, 
we were campaigning for hypertext systems even 
when we were building raster page turners based on 
them simplifying the information delivery and 
bringing down costs.   Eventually, I had to break 
down and admit HTML was the best of the worst 
solutions, and that almost everything else was worse.

BTW:  before it starts, I don't claim that tranforms were  
Goldfarb's idea, just that he had it in a presentation 
I attended and it was one of those AHA moments for 
me because the problems of remapping were well 
understood but I had yet to see the need for DSSSL. 
In other words, in a toolkit, it made enormous sense 
to anyone who had to write scripts for conversion 
work, and I'd done a lot of that.

I actually liked HyTime because a standard way to 
express dynamic linking (linking with a timing 
component) was on my mind.  Then every version 
got more complex and obscure and the pain started.
I finally had to buy Durand and deRose's book to 
understand what was there.   Otherwise, it has 
some very elegant ideas of which, probably 10% 
are useful in every day work, but the rest are 
there when the ordinary suddenly turns extraordinary.

Also, I am with Tim Bray.  <A href= works as is. 
Leave it alone.  When and if we start seeing 
floating text menus and more retrograde GUI 
controls in our systems, then we need to dust 
off these old designs and see what is useful.
I don't know if it is useful to design yet 
another presentation language.  That feels 
like MID IV.  Nice design but it undoes too 
much extant work just to get corner cases 
into the mainstream.  Well maybe.  MID was 
designed because the Navy wanted a notional 
presentation system that met their requirements. 
If you hear the term "common delivery platform", 
it is the same idea as yet unsatisfied ten 
years later. (Sure, they could use IE, but 
that is a non-starter in a world that must 
have inspected and provably correct code.)

To me, it seems fitful to make a link mean 
goto or get AND is-a or has-a.   It feels 
like a zero abstraction:  useful in the notation 
but meaningless otherwise.


-----Original Message-----
From: Simon St.Laurent [mailto:simonstl@simonstl.com]

I'm not questioning that XML is useful, but I am questioning whether
"SGML on the Web" ever really figured out the Web.  Given that we still
have friction between HTML and XML, manifested most recently as friction
between HLink and XLink, it seems worth asking.
> In hindsight it's easy to say "Oh yeah, of course you use XSLT for
> HTML output, and XSL-FO for print formats." But in, say, 1997 or -8
> there was no concept yet of XSLT as a separate transformation
> language. And considering what Len said about the old-timers'
> experience with web browsers, there may have been a degree of naivete
> about browser vendors' ability and willingness to support a new
> approach to document rendering.

Those are my memories as well.


News | XML in Industry | Calendar | XML Registry
Marketplace | Resources | MyXML.org | Sponsors | Privacy Statement

Copyright 2001 XML.org. This site is hosted by OASIS