Re: draft-ietf-html-style-00.txt & class as a general selector

lilley (
Fri, 8 Dec 1995 14:19:25 +0000 (GMT)

Scott E. Preece> said:

> | From: lilley <>
> | As you say, the mental model counts. Why is para 14 in strike through mode?
> | Perhaps because it is text from an earlier draft that must be present, but
> | shown as ready for removal or replacement?

> Perhaps it's strike-through to simulate writing something, crossing it
> out, and writing something else after it.

Or indeed some other thing. I did not claim to know your intent; just
said that *you* would know your intent. At least, I would hope so ;-)

> It's a weak example anyway

Yes. So, I am afraid, have been your others. Each example has used stylistic
change to express some meaningful information.

> | Note that the in-line, ad-hoc style usage you seem so keen to promote
> | denies the ability for either the author or the reader to apply a
> | different style.

> Allowing is not the same as promoting, just as facilitating is not the
> same as catering to the lazy.

Unfortunately, when the current browser-specific alternatives look so similar
to your in-line un-named style attribute method, promoting is exactly what
it will do.

> | Using your method, people using a speech synth or a
> | braille device or a device that can't do strike-through has no option
> | but to miss the distinction between para 14 and the rest. You have removed
> | the ability of print-disabled users to configure their browser and
> | extract the most meaning from your document.

> That's true.

Glad to hear you say that.

> In this case it happens to be a visual joke anyway,

Blind people don't have a sense of humour?

Think again about what you are saying - HTML is not visual. It is a bunch of
information, Several (common) presentations are visual. Here is some (Dilbert
inspired) information:

<p> I will need to talk to those <c class="blunder">idiots</c>
<c class="recovery">people</c> in marketing.</p>

Here is a visual joke using this information

I will need to talk to those idiots^H^H^H^H^H^H people in marketing

Here is an audio joke using it:

I will need to talk to those
idiots .. erm, *people* in marketing
^^^^^^------------------------------------ say louder and quicker
^^^^-------------------- embarrased, stifled giggle
^^^^^^----------tail off, short pause

> Other instances are probably a wash on this issue -
> "column group class emphasis 1" is not noticeably better than "column
> group style color red",

Because you carefully omitted details that would have helped form a
better class name. You spoke of tabular data from a scientific study and
the need to discuss certain groups of the columns. In practice they would
have meaning - "all those treated with amoxicillin" or "all those from
higher income groups or "the early 1980s" or whatever; a name could be
readilly produced.

> | Someone who personally hates strike-through but their device supports it
> | has no option but to live with it. Their style sheet cannot over-ride
> | yours. You have removed freedom of choice from the reader by not naming
> | the distinctiveness of your para 14

> That's true - if I have specifically chosen strike-through for a
> semantic meaning (as in the cited, hackneyed example), the reader has no
> choice but to live with it, anymore than if the reader prefers the word
> "structure" to the word "building" and I happen to have used the latter.

OK, fine. Let me try a different example:

<p>You must do <c style="{color: green}">these things</c> but never
do <c style="{color: red}">these things</c>

Someone who is red-green 'colour blind' (like, for example, 8% of the
caucasian male population) has to live with it, because you have removed
their freedom of choice by not naming the distinctiveness and hard coding
a single visual presentation into the HTML. And as Glenn said: if the
information happensd to be on some read-only device you can't go and edit it.

> | That is what "the separation of form and content" is about. It's not
> | some religious principle that is nodded to for no good reason. It is
> | a design goal that has real, practical benefits; benefits that you seem
> | desparate to throw away.

> As mentioned elsewhere, I'm not desperate and I have no desire to throw
> away anything.


> *I'm* the one who wants to include something that *you*
> want to throw away.

I don't want to throw it away; I want to not add it. I want this so as to
maintain the existing design goals of HTML. I trust you see the

Scott, this has been an interesting debate and it has helped clarify
and explore all the issues around using stylesheets with HTML. However,
it has become a bit of a two-person show and I am concerned that we
are just continuing the argument for the sake of it. Now, you have said
at various times in this and related recent threads::

> Of course a good stylesheet makes writing easier. It would be stupid to
> use in-place styling where a class is appropriate.

> I don't really care about the typing; tools really should take care of
> that.

> I would expect direct styling to be used so infrequently that
> the novice would be unlikely to be confused.

It has also been shown that inline style information prevents users
with other output media from being able to take full advantage of
the information in HTML. You have indicated that it is not your
wish to disenfranchise the disabled.

So, in the interests of wrapping up - and knowing that you are unable to
attend WWW4, which is a pity as this issue will surely be discussed -
I would like to ask your opinion. If you could be there, and it came
to a vote, would you raise a blocking objection if the proposed STYLE
attribute was dropped, or constrained to point to a named style only?

Chris Lilley, Technical Author and JISC representative to W3C 
|  Manchester and North Training & Education Centre   ( MAN T&EC )  |
| Computer Graphics Unit,             Email: |
| Manchester Computing Centre,        Voice: +44 161 275 6045       |
| Oxford Road, Manchester, UK.          Fax: +44 161 275 6040       |
| M13 9PL                            BioMOO: ChrisL                 |
| Timezone: UTC        URI: |