W3C home > Mailing lists > Public > www-style@w3.org > August 2005

Re: [CSS21] Wider variety of (non-junk) examples requested

From: Ian Hickson <ian@hixie.ch>
Date: Fri, 26 Aug 2005 22:18:38 +0000 (UTC)
To: Chris Lilley <chris@w3.org>
Cc: Adam Kuehn <akuehn@nc.rr.com>, Bjoern Hoehrmann <derhoermi@gmx.net>, www-style@w3.org
Message-ID: <Pine.LNX.4.62.0508262042110.26016@dhalsim.dreamhost.com>

On Fri, 26 Aug 2005, Chris Lilley wrote:
>>>> I don't understand why adding "</P>" in that example is good.
>>> Its good because it then becomes well formed xml, thus conforming to 
>>> minimal levels of quality for that particular document language.
>> It's not XML, it's HTML,
> OK cool, so conform to that if you want.

We are... As far as I can tell, all the examples are valid, or extracts 
from valid documents (i.e. they do not have anything wrong or misleading 
that would cause authors to learn back techniques).

> It is, as you noted, harder to conform to than XML but if you want to go 
> for the harder option, feel free as long as the majority of examples use 
> XML.

I'm not sure where you thought I said it was harder to comply to HTML than 
XML, and, I have no idea why you think there should be more XML examples 
than HTML, especially given that HTML is the more mature and more widely 
used technology when it comes to CSS UAs (by several orders of magnitude).

>> and it's a perfectly fine extract of a valid HTML document.
> How do we know that? 

By inspection?

> Its *potentially* an extract of a valid HTML 4.01 document. Its 
> "feasibly valid". But if, for example, it was a child of head, or title, 
> or img, or P, then it would not be valid.

I'm not even going to dignify that with an answer.

> I would have thought that the value of conforming to whatever 
> specification defines the document you are trying to style is trivially 
> obvious. You have in this thread pointed out subtleties of how many 
> nodes there are, exactly what the parse tree is, etc which affect the 
> examples. Thus, precision is vital to testability.

These are *examples*. The only concern is teaching people how to use the 
technology. It has nothing to do with testability -- you'll note our test 
suite uses XHTML as the primary format. In any case, HTML4 is just as 
precise as XHTML1, and just as unambiguous when you have a valid document.

There is no value in using full documents with DOCTYPEs and so forth, and 
you yourself have said this is not necessary. There is similarly no value 
in using XML over HTML for fragments, especially since most authors are 
using HTML, so that is what would be most helpful for them.

> The idea that ordinary mortals will be so perturbed by a closing </P> or 
> whatever that their heads explode rang hollow when I first heard it in 
> 1993 or thereabouts, and sounds even less likely now.

Similarly, the lack of one will have no effect either...

>>> Its rather sad to see this sort of allergic reaction to the mere 
>>> suggestion of making something be conforming XML, in this day and age; 
>>> see the related comment about unclear applicability to XML.
>> It has nothing to do with using XML or not
> Good, I am glad to see that the request to use mainly XML examples has 
> been accepted.

Chris, you have an annoying habit (seen both here and in other lists) of 
putting words in my mouth or drawing conclusions from statements I've made 
that have no logical link. I would appreciate it if you could stop it.

Just because the reason we refuse to change the examples has nothing to do 
with XML does not mean we have agreed to use mainly XML.

> See, its phrases like "latest syntax bandwagon" that give the impression 
> the CSS WG does not care for XML. Since CSS 2.1 claims to be suitable 
> for XML in general, a little more effort making such a claim plausible 
> would be much appreciated.

I think it's probably fair to say that the CSS working group does not 
"care for XML" any more than it "cares for HTML". Which is not to say it 
doesn't care for XML -- it's just that XML and HTML are considered equally 
important. CSS is language agnostic, and is widely applied to both HTML 
and XML, as well as other tree-based formats, such as SGML-based languages 
other than HTML. There are implementations that apply CSS to languages 
unrelated to HTML, such as XUL and DocBook.

At this point I really don't know what you're asking for. I'll certainly 
be happy to go through the spec checking that each example has no syntax 
errors and so forth, but requests to "use more XML" and suggestions that 
three line examples of simple document fragments that could be cut and 
pasted into HTML documents with no problems should have DOCTYPEs seem to 
be unhelpful, as they would either require dangerous changes (in terms of 
potential introduction of errors) or would make the document more 
confusing to the very people the examples are aimed at.

Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'
Received on Friday, 26 August 2005 22:18:51 UTC

This archive was generated by hypermail 2.3.1 : Monday, 2 May 2016 14:27:20 UTC