Re: LISP for Complex URC Sytax [WAS: Re URC func spec ]

Daniel W. Connolly ([email protected])
Thu, 24 Mar 1994 20:20:18 --100


In message <[email protected]>, Dirk Herr-Hoyman writes:
>At 8:32 PM 3/23/94 -0500, Simon E Spero wrote:
>>Using a lisp syntax is a great idea (hell, my next home system is going to be
>>a symbolics 3645); however it might be more appropriate to use SGML to
>>do structured text. If you compare the number of HTML pages out there with
>>the number of completed IAFA templates, it's clear that using RFC822 style
>>layout isn't always a big win.
>
>And with SGML there is at least some way of handling the character set issue.

Would you care to expand on that? I've been studying the expression of
character set issues in SGML for about two years, and I'm almost
completely confused at this point.

As far as I know, the character set issues that SGML deals with are
along the lines of "Ok... I've got this hardcopy of fred's sgml document,
so I'll key it in. Now... when fred writes &#65;, what character is that?
Ahh... I see in the SGML declaration that he's using ISO-646:1983, so
that's an 'A' character."

There is only ONE character set used in an SGML document -- the
document character set. You can play games with NOTATIONS, i.e. ways
of interpreting the characters that differ from the document character
set. I guess that would do for the purposes of using different languages
for different elements, ala

<Title notation="Kanji">&#134;&#125;...</Title>
(apologies to folks who actually know how to express Kanji using octets...)

>urc.dtd, why not?

Well, mostly because I think HTML is something of a disaster. My group
is trying to build a product that deals with HTML, and it's nearly
impossible. HTML is, in practice, not defined by any DTD. It's defined
by the NCSA mosaic source code.

But I'll have to admit I have been playing around with something you
could call urc.dtd. It's part of the "whole ball of wax" sort of
Integrated Open Hypermedia formalism I'm trying to figure out...

Some folks define hypertext as "text that's not constrained to be read
linearly." I think the more salient part about hypertext is that the
computer helps you nagivate and read it.

So the object of the game, in my view, is to allow the computer to
help us express/understand the messages we send daily. Things like:

"The lcs.mit.edu:/pub/contrib directory is mirrored at
gatekeeper.dec.com:/pub/mirror/mit/contrib"

"On Feb 27, Fred Flinstone said 'lakjsdlfkj' (and you
can read the rest of what he said in ...)"

"I think section 2 of that draft is bogus (here's how
you can read the draft yourself...)"

The reason that MIME is such a breakthrough for IOH is that there is
now a reliable way for the machine to peek into the body of the
message and understand it. As long as messages were plain text, you
were constrained to heuristic strategies. And if messages are programs
(like TeX and nroff) then there's only one way to get information out
of them -- run them. SGML offers an unprecedented level of structure
and reliability that we can use to enhance communications.