Re: why TITLE, not TITLE?

Stephanos Piperoglou (spip@hol.gr)
Sat, 10 May 1997 20:40:25 +0300 (EET DST)


Date: Sat, 10 May 1997 20:40:25 +0300 (EET DST)
From: Stephanos Piperoglou <spip@hol.gr>
To: Paul Prescod <papresco@calum.csclub.uwaterloo.ca>
cc: www-html@w3.org
Subject: Re: why TITLE, not TITLE?
In-Reply-To: <3373480D.70C248E6@calum.csclub.uwaterloo.ca>
Message-ID: <Pine.LNX.3.95.970510201541.308C-100000@fenchurch.hol.gr>

On Fri, 9 May 1997, Paul Prescod wrote:

> Since browsers tend to be used as part of the authoring
> process, they should also support validation and error reporting.

A problem much discussed on this list, as you know. Perhaps if NN had a very
strict parser then all these people "validating" their documents against it
would have learned to write good HTML.

Again, the problem lies with educating document authors. The argument is
simple: a very, very, VERY small percentage of people use the same browser,
browser version, OS, screen resolution, window size, image loading settings,
font settings, appearance settings, and language settings as the author
does, so there's no point in "validating" against your browser. Offshoots:

- Well, how DOES one validate a WYSIWYG-created document? Even if the HTML
is TECHNICALY valid, nobody can guarantee that the document won't degrade
into something ludicrous because the author used <BR>'s to align the text in
his tables or something. Unfortunately, one DOES have to learn HTML syntax
in order to create good (not just valid) documents.

- What is NOT required is that authors learn SGML. I haven't, for that
matter. I have no interest in SGML other than HTML authoring, BUT a lot of
my documents don't conform to an active DTD for various reasons. I generally
use the 3.2 tagset, but I also use CLASS because I think CSS1 is pointless
without it, and often I use ISO-8859-7 (Greek) characters, which will not
validate with the 3.2 DTD. I don't know how to make my own DTD, and can't
afford the rather incredible effort it will take to learn how to, since I
don't require knowledge of any other SGML application. My documents are
still useable with the vast majority of user agents, so how can I validate
them?

What is needed is:

a) Some education about what validation means and why it is necessary (all
those HTML books that are around should include at least something on this)

b) A wide variety of DTDs or a DTD that is adaptable enough (HTML Pro, for
instance, though a good idea, is comprised mainly of things that I don't
want and don't want to consider valid)

c) Authoring tools that validate, and browsers that validate NOT as a
process of parsing the document, but merely as an informative function so
the viewer (potentially the author) can view his document but can also be
aware that not everyone can.

--
Stephanos "Pippis" Piperoglou - http://users.hol.gr/~spip/index.html
I've never finished anything I began, but this time I'm