|
Posted by Jukka K. Korpela on September 14, 2007, 6:11 am
Please log in for more thread options
Scripsit Stefan Ram:
> You seem to know better than what is written down and are
> therefore able to judge which parts of the specification
> are to be ignored as »false« and which are to be obeyed,
> while I choose to take the specification and its wording
> for what it says.
No, I am just accustomed to reading standards and specifications, and
writing them too, so I can tell the difference between normative and
non-normative content. Examples are, even without any explicit statement
about their status, just examples, not normative.
When the HTML 4.01 specification _defines_ something normatively, then
presents examples that contradict the normative part, we can take this as a
symptom of a sloppy process. But if we want to argue in terms of what is
semantically or otherwise correct HTML 4.01, then the normative parts are
what matters.
Quite obviously, some people who participated in writing that stuff wanted
to define a "description list" element, but other people (or the same
people's other ideas, far from being crystal clear) made the specification
say "definition" and "term".
The conclusion is, not surprisingly, that <dl> is basically good for
nothing. There is nothing to be won by using it.
--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/
|
|
Posted by Stefan Ram on September 14, 2007, 10:20 am
Please log in for more thread options
>The conclusion is, not surprisingly, that <dl>
>is basically good for nothing.
When one finds that one's interpretation of a concept renders
it useless, sometimes this is an indication to question this
interpretation.
|
|
Posted by Jukka K. Korpela on September 14, 2007, 3:40 pm
Please log in for more thread options
Scripsit Stefan Ram:
>> The conclusion is, not surprisingly, that <dl>
>> is basically good for nothing.
>
> When one finds that one's interpretation of a concept renders
> it useless, sometimes this is an indication to question this
> interpretation.
In this case, the one whose interpretation has been proved to be illogical
and inconsistent is the W3C, the authority that defined <dl>. In browsers
and other software that processes HTML markup, <dl> elements are just
something to be rendered in a particular manner (by default at least), with
no processing that would be based on _any_ interpretation of any semantics
of <dl>.
And there is no useful processing that _could_ be based on the
interpretation of <dl> as "description list", whereas the normative
definition as "definition list" _could_ be used (e.g., in searches for
defining occurrences of terms, or in searches that should ignore such
occurrences. Thus, if you use <dl> for something else than a list of
definitions of terms, you take a risk of having your document processed all
wrong. The risk is small, but why take it when there is nothing to be won?
The burden of proof lies on those who suggest that <dl> is useful for any
other purpose than getting the particular rendering style that browsers use
for it. And naturally, we are supposed to use CSS and not presentational
markup for styling.
--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/
|
|
Posted by Stefan Ram on September 14, 2007, 4:12 pm
Please log in for more thread options
>In browsers and other software that processes HTML markup, <dl>
>elements are just something to be rendered in a particular
>manner
The page
http://www.informatik.uni-freiburg.de/Java/manpages/glossary.html
contains
<dt>definition
<dd>A declaration that reserves storage (for data) or provides
. When one searches for a definition of »definition«
http://google.to/search?q=define%3Adefinition
, this snippet is actually found.
>And there is no useful processing that _could_ be based on the
>interpretation of <dl> as "description list",
According to the HTML 4.01 specification, it is a list of
mappings (pairs) belonging to a certain relation, like
product-and-advantages, speaker-and-text, or term-and-definition.
The definition does not use my wording »mapping« or
»relation«, but this is what best fits the text and examples.
>whereas the normative definition as "definition list" _could_
>be used (e.g., in searches for defining occurrences of terms,
>or in searches that should ignore such occurrences.
This would be useful, but is not the normative definition.
>Thus, if you use <dl> for something else than a list of
>definitions of terms, you take a risk of having your document
>processed all wrong.
This is correct. Therefore, the above Google search also
might give some misleading hits.
>The burden of proof lies on those who suggest that <dl> is
>useful for any other purpose than getting the particular
>rendering style that browsers use for it.
I tried to report, what the HTML specification specifies
about DL.
This does not include the claim that this DL-specification is
actually useful or that I would applaud to it or would have
written it in the same way. These are different questions.
I am just the messenger of this DL-specification, not its
author. Therefore, I do not have to proof that it is useful.
>And naturally, we are supposed to use CSS
>and not presentational markup for styling.
A document marked up according to the specification should be
rendered in a reasonable and pleasent way, even when no CSS is
used. This is the responsibility of the user agent.
|
|
Posted by Jukka K. Korpela on September 14, 2007, 5:34 pm
Please log in for more thread options
Scripsit Stefan Ram:
> <dt>definition
> <dd>A declaration that reserves storage (for data) or provides
>
> . When one searches for a definition of »definition«
>
> http://google.to/search?q=define%3Adefinition
>
> , this snippet is actually found.
What makes you think it has anything to do with the _markup_ used?
_If_ it had something to do with it, then it would be a strong argument
against using <dl> for anything but lists of definitions for terms, since
such usage would imply seriously misleading search results.
> According to the HTML 4.01 specification, it is a list of
> mappings (pairs) belonging to a certain relation, like
> product-and-advantages, speaker-and-text, or term-and-definition.
You still haven't checked what the specification says _normatively_ - how it
_defines_ the semantics of <dl>, as opposite to examples and notes about how
it could be used bla bla bla.
> I tried to report, what the HTML specification specifies
> about DL.
It does not help, since you have not paid attention to what it _specifies_:
"Definition lists vary only slightly from other types of lists in that list
items consist of two parts: a term and a description."
What it later says about what the element could be used for does not change
the _meaning_, any more than the defined _meaning_ of <blockquote> is
changed by notes like "some authors have used BLOCKQUOTE merely as a
mechanism to indent text".
--
Jukka K. Korpela ("Yucca")
http://www.cs.tut.fi/~jkorpela/
|
| Similar Threads | Posted | | Problem with descriptive lists in CSS | July 6, 2004, 2:54 am |
| A rendering problem with left aligned image and lists | August 13, 2004, 9:34 am |
| lists in tables | October 9, 2004, 5:43 am |
| Definition lists | March 3, 2008, 9:01 am |
| CSS display blocks and select lists in IE | July 16, 2004, 9:59 pm |
| the canonical structure of links in lists? | July 26, 2005, 4:31 am |
| Gracefully degrading expandable lists | January 24, 2006, 1:18 am |
| Good practice? Title attributes on lists of links | August 13, 2005, 4:04 pm |
| Need Help With Semantic Markup | October 12, 2004, 5:47 am |
| Markup Validator | April 5, 2007, 11:32 am |
|
XML Sitemap