Re: comments re draft version 2.0

David Carlisle:

> > I believe that the technique of special-loaded DVI's knows
> > no bounds other than an author's reasonable patience (and the danger
> > of overloading the design of DVI).
> I think you need to keep two things distinct,

I do, and I agree with almost everything that you said.

>                                               ...  or a radical SGML
> declaration to make an SGML system understand tex like syntax (about
> which you could say more than me:-)


Well, I finally decided to quit my odd ball reliance on '/' with
SHORTTAG to mark empties.  (It did work without a problem with my elisp
output.)  So now "\foo;" generates "<foo/>".  The use of the ';' is
optional in most circumstances with the particular exception of when
"\foo" is followed without white space by PCDATA provided that "foo"
is declared empty.

I continue to prefer to handle "\alpha;" as "<alpha/>", rather than
"&alpha;" in most cases, because I want to have the option of writing
output formats that do not handle "&alpha;".  Beyond that there are times
when delayed evaluation is needed.  ("&alpha;" is not in yet, but it will
be, at the very least its code will be, when unicode is brought in.)

Aside from these things and, of course, the name space, the SGML set up
is just like that of HTML 2.  Hardly radical, any more.

                                     -- Bill