Re: Hyphenation (was Re: A suggested tag)

F. E. Potts (fepotts@fepco.com)
Sat, 19 Apr 1997 14:03:24 -0600


Date: Sat, 19 Apr 1997 14:03:24 -0600
From: fepotts@fepco.com (F. E. Potts)
Message-Id: <97Apr19.132508mdt.18434@gw2.fepco.com>
To: msftrncs@htcnet.com
Subject: Re: Hyphenation (was Re: A suggested tag)
Cc: www-html@w3.org

On Sat, 19 Apr 1997 11:42:17 -0600, Carl Morris responded to:
> > Hyphenation is a stylistic attribute not a structural change. It
> > follows then that an attribute could be added to CSS and there is
> > no need to extend HTML.

> Maybe I am, late to this punch line, in some language somewhere,
> PARAGRAPHS, HEADINGS, BLOCKQUOTES, TABLES are all stylistic...
>
> Lets get off the damned issue, HTML is stylistic, ASCII is
> content...  There is no way to describe one without the other, even a
> TEXT file is stylistic.

We get into issues of style vs. structure a lot on this list, and
mostly the arguments are fairly silly.  Certainly we try to keep
structure separated as much as possible from presentation issues, but
there are fuzzy areas to take into consideration.

Let me give an example.  An editor of mine, working from a FOSI
printout, was having difficulties determining where untitled sections
began and ended under certain conditions (for example, when one section
ended at the bottom of a page, and the next began on the following
page).

My solution to this problem--something that made the purist in me
shudder--was to add to the 12083:1994 FEPCO variant DTD I am using a
new element called separator.  This element is used as follows:

	<separator size="12">&bull; &bull; &bull;</separator>

This scandalous element, placed between sections, would, in the FOSI,
cause it to put a centered line of three bullets between the sections,
solving my editor's problem (generated text, my first choice, was too
complex, for some sections have titles, and others don't).

The point of all this is that in "real-world" situations we do the best
we can, and that's it.

CSS and other style sheets are, in effect (like the FOSIs I use),
basically batch processors (or at least that's the way I think of
them).  And HTML is mostly a presentation DTD.  There are many
occasions when we will have a small item to post, and it is most
appropriate to just write it in 3.2 and let that be the end of it.

Beyond that, when, say, we might be posting an article or a book, then
style sheets come into their own, for we only have to design it once,
and then all the following pages will be properly formatted (and, if at
some future date, we wish to make a change, we only have to do it once
to the style sheet to make the change globally effective).

I personally would like to see HTML stay pretty much at the 3.2 level,
and have most other additions focused on the style sheets.  HTML is
already getting to be far too complicated, with far too many "tags", to
be easily used by the general public.  And if HTML is made too
complicated, too difficult to use, for the general public, they will
either not use it, or will use it badly (as most of them do now, which
surely should say something).

The pros, of course, will move on to XML (or SGML), where they can have
the power and control they need.

I might note that for a book or article, I can come close to
duplicating my FOSI's output using 3.2, so HTML at that level can be
quite useful.  It is just that style sheets make many things easier.

We really do get too anal on this list over the issue of structure vs.
presentation, and should focus instead on (1) keeping HTML simple for
the masses, (2) adding a simple style sheet (CSS), and (3) be working
towards the future (which seems at this point to be XML and DSSSL-o).

And that brings up another issue: should not W3C be thinking about a
public list similar to this one for XML?  It seems to be about time for
it.

-fep

--
fepotts@fepco.com
http://www.fepco.com/