RE: Tag Soup (was: FW: XHTML)

On Sun, 5 Dec 1999, Kjetil Kjernsmo wrote:
> On Sun, 5 Dec 1999, Arjun Ray wrote:
> >

> The problem is all the very bad authoring tools out there, and all
> those who have earn big money by writing poor pages and teaching
> others how to write poor pages. I guess we would have to educate the
> educators...

Yes, re-educate them, if possible.  Assuming they would even want to
learn.  The vast swath of lemmetellya gibberish out there is *not*
causeless.  Look closely, and as a corpus, it indeed has an internal
consistency.  These people *learned* Tag Soup from their software, and
they are simply passing on their "knowledge" to others.

> As TimBL points out, we are in a desperate need of good authoring
> tools,

I'd say "good authoring habits", too.  

> But for people to use authoring tools like that (as opposed to those
> outputting tag soup), we would have to teach people to adopt a
> different mindset.

Indeed.  Writing structured markup is more like filling in the blanks of a
template than stream-of-consciousness composition ("Uh, I guess I need a
<FOO> here now...")  At a minimum, it involves a conscious decision to use
a *document type*.  In other words, the focus is on preparing an entire
document of a particular class, and not on cranking out a one-off, sui
generis and irreproducible.

> I think we would have to tell people how they can do what they first
> want to do "You wan't to write meeting minutes? Here's how!" "Personal
> Homepage? Start like this" "Scientific Texts? This is what you need to
> know." Once they've gotten into it that way, they may have understood
> the power of HTML, and may continue to write structured HTML for other
> purposes as well.

Not HTML.  SGML.  More than one document type.  Most of the time, HTML can
profitably be the result of a transformation process, from documents in
tagsets more attuned to the particular information domain.  With all its
cryptic initialisms, it's very hard to overcome the impression that it's
all voodoo. 

> >For one thing, emphasize the *distinction*.  Let Tag Soup have its day in
> >the sun.  Give it the legitimacy of a formal spec.  
> 
> Perhaps... My first reaction is that I'd rather say to people "folks, this
> stuff is dead  'E's passed on!  This parrot is no more! He has ceased to
> be!  'E's expired and gone to meet 'is maker! [...] THIS IS AN
> EX-PARROT!!" s/parrot/tag soup/i, :-) and that would be hard if tag soup
> gets the legitimacy of a formal spec, I guess.

You may be right.  But, to my mind, the insidious problem is in the
implicit premise - or "persuasive" suggestion - that Tag Soup and
Generalized Markup *can* be reconciled.  The value of a spec for Tag Soup
is that it will be explicit, and thus can be made to stand out in
contradistinction.  It makes no sense to pile exception on "application
convention" on special case on friendly prose, all to come up with
something that claims to be an "SGML application" and in the same breath
even begin to suggest that it has absolutely *anything* to do with Tag
Soup processors like Netploder.  This is humbug on a truly monstrous
scale.  And, if instead, this "SGML application" has nothing to do with
Netploder, then why in &deity;'s name *isn't* there something that does?


Arjun

Received on Sunday, 5 December 1999 12:45:27 UTC