Re: An Enhanced Web (was Re: [Fleckenstein] The Story of Bubblenomics)

From: Eugene Leitl (eugene.leitl@lrz.uni-muenchen.de)
Date: Mon Dec 27 1999 - 00:13:23 MST


Harvey Newstrom writes:

> You are no longer discussing what is currently supported. Your argument has
> degenerated to the claim that since postscript is a programming language, it
> can do anything you program it to do. I agree with you at this level, but
> think the claim could be made for almost any language.

Point is: 1) Some languages are much better at it than others 2) HTML
is not a programming language, but a hypertext markup, with hooks to a
pointless Babel of ad hoc languages added as an afterthought. Totally
different ballpark altogether.

> Yes, but the size of postscript rendered pages make it too slow for an
> interactive webpage over a modem. I also am familiar with Display

PostScript can be pretty compact. Look at what dvips tends to produce,
and it does font embedding (another massive shortcoming of almost
(almost? almost??!) any other system I've seen. Groping for fonts
under X, straining to read something? Trying to make a reproducible
hardcopy under Winblows?

Manually concocted PostScript documents (especially considering the
opportunity for compression and the threaded nature of the PostScript
code) could be very compact indeed.

Furthermore, if you could render PostScript incrementally, instead of
downloading multi-megaton machine-dumped-crap documents with braindead
"compression" ("compression" as in RLE) standards (Yaypegs, wavelets
in PostScript? Where art they?) and launching a (crappy, since I don't
use AkroBot) external viewer it would render just fine.

> Postscript, such as on AIX or NeXT systems, and am aware that it is a
> wonderful display format language. In theory, I agree with your idea that
> postscript would have been a great standard for the web. I just don't think
> that it was an obvious choice and that there was no reasons for not using
> it.

Well, it was very obvious to me. The only obvious minus was that
decent OpenSource PostScripts weren't very available back then, so
you'd have to pick up a decent Forth (better write it from scratch),
and implement PostScript on top of it.
 
> > Can't you see my point? We wouldn't have the current web debacle if we
> > did it right from the start. And it could not have been done by the
> > industry, because it is profitable to erode standards.
>
> I agree with this, totally. But I am pessimistic enough to think that
> postscript would have been eroded by Netscape and then Microsoft just as
> HTML was. Modern businesses seem to go for short-term goals and destroy
> long-term goals. I think this bad management would have happened no matter
> what standard was chosen.
 
Solution? Easy. Minimalistic, trivial-to-define-and-possible-to-verify
standards. Preferably done by a single person, or two. Expandable
standards, with well-defined expansion hooks. Trusted source (one or
few persons vs. Godzillaesk groups of persons, particularly rabidly
mercantile ones. Linus vs. Korporate Amurrika, Inc.). Really mean
badass BDSM torture test suites. Proof if identity by authentication.
(Would it require breaking a full-blown cryptosystem to pervert a
standard instead of just a few legalese-spouting sharks in suits, we
would have a lot less Microshafts).

As long as you can claim anything without having the burden of proof
you're just screwed.

You see, I've just devined by own Java. It's totally incompatible to
anything, but it's true genuine organic 110% Java. You better believe
it, d00d.

> In summary, I think I agree with all your points. I just am more
> pessimistic in my vision of what could have happened had things been
> different.

We have a rich set of datapoints of what has been running wrong with
the industry/market. While certainly a minority, it should be doable
to define nonprofit longer-lived standards. In fact, survival of
OpenSource coder community will be dependant on that.



This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:06:12 MST