From: Eugene Leitl (eugene.leitl@lrz.uni-muenchen.de)
Date: Sun Dec 26 1999 - 20:19:35 MST
Harvey Newstrom writes:
> > What's so magical about hypertext? It is not very hard to do.
>
> I didn't say it was hard. I said that HTML does things that postscript does
> not currently support.
PostScript can support anything, because (as Forth and Lisp and some
more modern but fatter languages (Python would come to mind)) it is an
all-purpose metalanguage, tailored for defining problem-specific
languages. (By the way a very different mindset than nonincremental
compiling languages, but you can't make their users see that because
they don't even know what they're missing). If combined with
authentication mechanisms allowing you define trust hierarchies it
could for instance fill the niches of HTML, JavaScript, Java, XML and
VRML and streaming multimedia while (in theory) requiring a single
browser running a standard (and when I say standard I mean _standard_)
VM (which at some point would be cast in silicon (while requiring a
fraction of the transistors of Javur) and migrate into video
hardware), by embedding code within documents, or allowing you to
pluck missing methods off the net. (The high-octane stuff (rendering
multimedia) would sure suck molasses, unless implemented efficiently
(so there _is_ a motivation for upgrading browsers and making
profits), but this at least would allow people with obsolete browsers
to have a peek at the high-octane content regardless of the browser
version and the OS they happen to be running). GAR. I'm in a bad mood,
but you already knew that.
> > HTML is an ASCII stream. PostScript is an ASCII stream. PostScript is
> > a powerful, yet simple to implement interactively expandable
> > all-purpose programming language with a focus on device-independent
> > graphics and typography, including real-time GUI rendering.
>
> I know what postscript is. I know what HTML is. You cannot currently do
> HTML type stuff in postscript. That is the answer as to why HTML was
You have never used Adobe Acrobat, have you?
> invented and postscript wasn't used for the web.
The reason a beefed-up PostScript wasn't used for the web is because
the people who pioneered HTML were turkeys, and couldn't extrapolate
two steps ahead nor did take a good look at prior art.
I have a right to say this because I thought HTML was total crapola
when it first came out, for above reasons. No benefit of hindsight
then. Of course the computer industry never made much sense to me
either. I never understood why the Lisp machines never catched on and
why Forth is still a niche. It takes the ach-so-rational computer
industry and the market decades to deliver decent technologies to the
user, while suffering throwbacks all the time.
We all know the reasons, but this doesn't mean I fucking have to like
it.
> > HTML is just a simple markup language. PostScript with a couple
> > screens of code can do everything HTML do, and More.
>
> Yes, with proper programming, postscript could be made to do what HTML does.
> But you would have to add all the mark-up capabilities of SGML to postscript
> to create a mark-up language. Sure, it could have been done. The original
You make it sound as if this is rocket science.
> HTML designers felt that a creating a subset of SGML was easier than
> modifying a printer/graphics language for the task.
All you have to do is give a PostScript interpreter access to the
input stream. Clickety-click, typety-type. (Remember NeWS? Black
hardware?) Everything else is a single screenfull of PostScript
code. Which can be a part of the document you're currently
downloading.
If you want to be real fancy, throw in networking, file access and
hardwire strong crypto into the interpreter. The need for it was
fairly obvious a decade ago.
> People have created hundreds of web languages. Another one made out of
> extended postscript is quite possible to do.
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.
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 15:06:12 MST