W3C home > Mailing lists > Public > public-html@w3.org > September 2009

Re: Change back the semantics of <cite>

From: Smylers <Smylers@stripey.com>
Date: Fri, 11 Sep 2009 01:03:56 +0100
To: HTMLWG <public-html@w3.org>
Message-ID: <20090911000356.GA5051@stripey.com>
Leif Halvard Silli writes:

> Smylers On 09-09-11 00.16:
> 
> > Leif Halvard Silli writes:
> >
> > > HTML 5 draft ... explains ... <cite> may only  be used for the title
> > > of a work.  I have just stumbled upon a problem in this regard
> > > when looking at the <dialog> element, where <cite> would be the
> > > right element for singling out the acting object:
> > >
> > >  <dialog> <dt> <cite> Leif</cite> , HTMLwg member
> > >                 <dd> I propose <dl dialog> instead
> > > </dialog> 
> >
> > Why would <cite> be the right element for that?
> 
> Because the current definition (aka HTML 4 definion) of <cite> is
> "name of a source" and includes people source. And source is a good
> name for each participant in a dialog.

That's a circular argument: you're saying that <cite> needs to encompass
this use because it's the right element, and it's the right element
because it's within what you believe the definition of <cite> should
say!

In general folks' names aren't marked up in running text, and users
don't need a special 'this is a name' semantic to be pointed out to them
in any way.  If you wish to include multiple pieces of information in
the <dt> of a <dialog> then the comma you have used above is sufficient
for users to work out what's what.  If you wish to style the pieces
separately then <b> or <span class=speaker> (or, for the other part,
<span class=affiliation>) could be used.

Why do you think this particular semantic requires support from the
HTML5 language?

> > Why does your name need marking up at all there?
> 
> You'll find the answer in the thread and in bug 7508 and 7509. But
> I'll recap: The answer is the same the answer to why a it is necessary
> to use a <dfn> inside the <dt>

Why is it the same?  Defining instances are generally marked up in
running text, so that they may be distinguished to users in some way and
so they can be linked to.  If a term is defined in a list those reasons
may still apply, and an author can choose to continue to use <dfn> for
the same reason it's used elsewhere.  (Note that HTML5 gives an example
of using <dfn> like this; it doesn't say it's "necessary" to do so.)

Conversely, names are normally not marked up at all.  They can continue
not to be marked up at all when in the <dt> of a <dialog>.

> - despite the fact that it is a _definition list_.

HTML5 calls it a "description list", and gives examples of its use where
the terms are not definitions in the <dfn> sense.  For example it shows
using <dt> in a <dl> to declare the authors of a document; those people
are not the definition of the word "authors", so <dfn> would be
inappropriate in that <dt>.

> Namely: If you stuff the <dt> with extra information - such as the
> draft gives examples for w.r.t. glossary lists - then it is necessary
> to single out which of the words in in <dt> that is being defined.
>
> Actually, the draft says that in a glossary one should _always_ use
> <dfn>

No it doesn't.  It has examples which use it, but there's no "should" or
"always" in there.

> - regardless of whether the <dt> contains only the exact, defined term
> or not. But otherwise it is the same idea that I am talking about.

Not really.  It shows that purpose of the <dfn> is not to single out
part of the <dt> (because there isn't always something to single it out
from).  The purpose of <dfn> there is exactly the same as <dfn> would
have anywhere else in the document.

> > > Hence I propose the definition to change. The new text should say
> > > that <cite> is not a mark-up for "name" or "person". However, it
> > > is an element for marking up a source. And if the source is a
> > > person, then <cite> may indeed be used to mark up the reference to
> > > that person.
> >
> > Why is having such an element useful?
> 
> No one has so far suggested obsoleting <cite> .

Right.  <cite> as defined in HTML5 is useful for marking up titles of
works, since such are usually distinguished to users in some way.  That
use can be met by leaving the definition of <cite> as it already is in
HTML5.

You haven't answered why having an element with your definition would be
useful.

> > Having an element for marking up the titles of works is useful
> > because they are usual formatted distinctly (typically italicized)
> > in text, to convey to readers that the title is not 'normal' text.
> > Conveying that requires _some_ element.
> >
> > Whereas people, even when sources, do not typically have their names
> > distinguished.  So using a <cite> for both prevents it from being
> > able to convey anything useful.
> 
> I feel that you had this answer ready made - I'm not sure if you have
> looked into why I asked for <cite> in this round.

I looked, but couldn't see it -- hence my "why" quesions!

> I arrived at this from an actual use case.  I have explained above why
> it can be necessary to single out who - in the <dt> element of the
> dialog container element - is the source.

I'm afraid you haven't.  You've asserted that you want to mark up the
source, but haven't explained why that is useful.

> But I can say more:
> 
> One will often, in a dialog, need to have different styles for  
> participant versus other information.

Not everything which is styled differently needs to have semantics in
HTML5.  Where the styling is merely presentational classes can be used.

> I have merely chosen the same approach as Ian suggests for glossaries
> in the draft: Single out the center word - the "defined term" so to
> speak - instead of marking up what is /not/ the center word.

No you haven't.  Ian suggests using <dfn> to mark up defining terms
consistently wherever they are in the document (in running text, or are
part of a <dl> <dt>, or as the entirety of a <dl> <dt>).

You suggest a need to mark up names just in the case that they form part
of a <dialog> <dt> (but not when they are in running text or an entire
<dialog> <dt>), and to mark them up with an element which elsewhere is
used to convey titles of works.

That is hardly "the same approach".  But even if it were, it doesn't
explain _why_ doing so is a good idea.

Smylers
Received on Friday, 11 September 2009 00:04:36 UTC

This archive was generated by hypermail 2.3.1 : Monday, 29 September 2014 09:39:08 UTC