Re: Key/Item data

Jim wrote on Monday, March 17, 2003 at 11:49:41 AM:

> On Wednesday 12 March 2003 9:18 pm, John Lewis wrote:

> [renaming <dl>]

>> Would it make a lick of difference? Yes. Authors of HTML/XHTML1
>> would have to learn new element names instead of using the old
>> ones,

> They have to do this already, xhtml 2.0 is not meant to be
> backwards-compatible. What's the point in breaking compatibility if
> you aren't going to take advantage of it?

What's the point of breaking backwards compatibility for no tangible
gain? My position is that the names are fine as they are, so changing
them would just be changing them for the sake of changing them. I
don't think changing dl/dt/dd's names is an advantage. I think it's
the definition that should be clarified.

I think you can agree there is also a difference between learning (for
example) five new elements and eight new elements. Taking out three
elements and replacing them with three new elements is a different
thing entirely from adding a number of new elements. Replacing three
elements is a big deal in a small language like XHTML2.

(On the other hand, I think replacing h1-h6 with section/h is a
fantastic idea. It breaks backwards compatibility, but there are real
gains to be had. h1-h6 have always felt limiting and clumsy to
style--I think the new system will be twice as good, and not very hard
for (X)HTML authors to adapt to, which is a nice bonus. As well, we
will be replacing six elements limited to six levels of headings with
two elements unlimited in the number of levels of headings.)

>> implementors would presumably need to duplicate their default
>> dl/dt/dd styling for three identical but differently named
>> elements,

> A couple of lines in a stylesheet would do this, unless I'm missing
> something.

The point is that it's unnecessary unless you rename the elements. I
said it was a bad thing. I never said it was an *infinitely* bad
thing. The world will not crumble if we replace dl/dt/dd. It's my
opinion that we don't need to, and that we don't gain much of anything
by doing so.

>> and authors (new and old) would have the dubious benefit of a more
>> appropriate element name.

> There's nothing dubious about it. The benefits of good names are
> widely accepted to apply to functions and variables, why not
> elements and attributes?

dl/dt/dd are good names. You haven't provided me with a good reason
why we need new names, or why dl/dt/dd are bad names. (You say the
benefits of good names are "widely accepted," but you don't list any.
You also haven't explained why dl/dt/dd are bad names--unless being
unintuitive means bad.) I'm not the one that wants the spec
changed--the burden of proof lies with you. I'm defending the current
names because I think changing them would be basically pointless.

>> I think a clearly more beneficial course of action is to clarify
>> that dl/dt/dd are generic elements. The example in HTML4 implies
>> that they are, and a rewritten definition would make that
>> implication explicit.

> I would be happier with this than leaving it ambiguous, but I do
> feel that naming and describing something as a definition list
> should count for something, it's very unintuitive to twist the
> meaning like that.

The XHTML specification defines what the elements mean. The names of
dl/dt/dd already don't match their definition perfectly in (X)HTML.
Neither do span or div, the inline and block generic structure
elements. Surely we should rename them as well to prevent confusion? I
think all of these elements should keep their names for the reasons
given above.

> [...] it's very unintuitive to twist the meaning like that.

True, it can be unintuitive. I've yet to see an author learn perfect
(X)HTML through intuition. XHTML is already unintuitive, and renaming
dl/dt/dd won't change that. That's what tutorials and the like are
for. I agree that intuitive names are a good thing, but I don't agree
that we should bother a great amount changing element names so they
are intuitive.

I'm sure my opinion is partly colored by my own use of (X)HTML. I've
used dl/dt/dd for appropriate non-definition list uses before. I've
seen others do so as well. So I already know that the names will not
hinder dl/dt/dd's use greatly even with an *unclear* specification.
Right now someone can look at the example and claim it's a fluke (or
even a mistake!). I think 99% of the usefulness of renaming the
elements, and none of the disadvantages, can by had simply by
clarifying the definitions of dl/dt/dd in the prose.

-- 
John Lewis

Received on Monday, 17 March 2003 15:10:34 UTC