Re: Lout to HTML translator ?

[rodrigo vanegas]

Your message dated: Thu, 15 Sep 1994 09:35:16 -0000

>I have been using Lout for some time now as my preferred text formatter. I
>would like to now place some of my documents on the WWW and wondered if 
>anyone has already developed a lout to HTML translator.

The short answer is: not that i know of and i doubt it could be done
well anyway.  But you may want to find another simpler language which
can be processed to produce either ASCII, TeX, or HTML.  See below...


The long answer:

It's unclear that this is exactly what you want.  Lout, like TeX and
your typical WP for PCs, is software that let's you unambiguously
describe pages of text with a degree of precision captured only by
something like PostScript or a bitmap for each page.

For example, if you write "This sentence is false." in your document,
Lout will unambiguously decide what font family and point size to use
and the precise location where this setence belongs.  If your printer,
which understands Lout's output in PostScript, were to display the
sentence in a different location or with a different font, it would
be considered a bug.  Lout is a high-level page description language.

Contrast this with HTML (which is a derivative of SGML).  HTML is a
markup logical document description language.  This means that when
you say "</em>emphasized text<em>", you've described the enclosed text
as having the logical attribute of "emphasis".  Lout, on the other
hand, allows you to "@I italicize" text.  There is no pretense of
purpose (logical structure) but instead merely a promise that it will
be interpreted as having the physical attribute of italics.

Why is this important?  Because the output of Lout is PostScript and
deterministic.  That is, for any given Lout document, there is one
expected PS output.  With HTML, however, the interpretation is subject
to the client software at the other end of the net.  It is true that
printers, too, are "interpreting" the PS Lout produces but the
manufacturer's intent is to produce a bitmap which matches the
expectations of the PS definition as closely as possible.  HTML
viewers, on the other hand, can do anything they wish!  This is why
HTML does not have commands such as "gimme 2 inches of whitespace" or
"select the Helvetica font".


>On the same lines has anyone developed a Lout to anything else translator - so
>that I could go Lout -> something else -> HTML !!

So given the above distinction, the correct translation is actually in
the reverse direction: HTML -> Lout.  Since this isn't quite what you
want, your best bet is to use a third language which is an appropriate
subset of both HTML and Lout and produce source for each in turn.

Neither of these is exactly what you asked for, but they are good
leads.  In the free software world, there are two that i know of

  GNU's TeXinfo.  It has three output formats.
    * Info files (this is quasi-ascii hypertext within emacs)
    * TeX -> DVI -> PostScript
    * HTML (the program that does this is available separately)
  prep.ai.mit.edu:pub/gnu/texinfo

  Ian Jackson's Linux FAQ.  This is a real hack but it works for at
  least one person.  It also has three output formats.
    * Info files
    * plain 7bit ASCII
    * Lout -> PostScript
  You should ask Ian personally.  His email is: address@hidden


rodrigo vanegas
address@hidden