Re: Toward Closure on HTML

Daniel W. Connolly ([email protected])
Tue, 05 Apr 1994 10:36:03 -0500


In message <m0po9a4-000062C@monolith>, William M. Perry writes:
>
> The great thing about the HTML+ DTD is ... get ready... it is actually
>parseable by commercial (and free) SGML editors/validators. Unlike the old
>html spec.

Which html spec lacked a DTD? I have several versions, including
the one from info.cern.ch that work fine (except that documents
from NCSA rarely parse :-)

> So, hopefully more and more people will start using these
>tools, and they won't have to worry about the 'oh-so-inconvenient' tags.

Here's hoping!

>
> How's this for a proposal:
>
>1. Get rid of <p> tag and replace it with double-\n
>2. Get rid of <li> tag and replace it with \t*
>3. Get rid of <dd> tag and replace it with \n
>4. Get rid of <dt> tag and replace it with \t*
>5. etc
>6. etc
>
> That would increase the readability of the text immensely.

Now if we do this, folks will _have_ to start using SGML parsers
to validate documents when they write them, and they'll _have_ to
use real SGML parsers at runtime. This stuff _can_ be represented
with SGML SHORTREFs, but I don't think you'd like the situation
you'd end up with -- one misplaced '*' character could screw up
the whole parse.

But the biggest problem with introducing shortrefs at this point is
that we'd really want to wait until they were implemented and used for
a while before we could standardize them. Now is not the time for
that, I'd say.

Dan