[whatwg] clear naming for WHAT work

Dean Jackson dean at w3.org
Mon Jul 19 02:17:37 PDT 2004


On Wed 14 Jul 2004, Matthew Raymond wrote:

> Dean Jackson wrote:
> >I suggested privately that you call this effort HTML 5. I
> >still think this is a good idea. I notice that your DOCTYPE includes
> >"HTML5".
> 
>    I think that "HTML 5.0" would be the best name for the standardized 
> version of the drafts we're working on, and I wouldn't mind seeing 
> "HTML5" used in the doctype for the combination of all the drafts.

Great! I'm not completely crazy then (or at least I have company).

> >Dave Hyatt responded that HTML 5 was a bad name because the new
> >features work in XHTML as well.
> 
>    I don't see why. XHTML 1.1 doesn't have the same version number as 
> HTML 4.01, and no one complains about that. Then again, perhaps I don't 
> understand his position.

I'd like to hear more about it as well.

> > Is this acceptable? One of the
> >benefits of XHTML is that you (are supposed to :) know what is
> >happening. If it is an extension to XHTML, it probably should
> >use the XHTML extension mechanisms. 
> 
>    As far as I know, it does, in the form of an XHTML module called 
> "XHTML5B".

Cool. I did see some references to that flying past since I posted.

> >I really don't think you should add new elements like <output> to the
> >XHTML namespace. The reason the namespace is there is precisely so you
> >don't have to do this.
> 
>    I'm not convinced of this. If that were so, you'd have to create a 
> new namespace every time you wanted to expand functionality. Over time, 
> that would create a massive number of namespaces and cause significant 
> confusion.

I don't suggest you create a new namespace for every rev.
Just one to identify the elements added by WHAT.
I assume that WHAT will try to avoid any conflicts with itself.

(and note that I'm not sure that WHAT's extended <input> element
should be in the XHTML ns or something else)

> To me, namespaces are for separation by category or type, not 
> by revision or extension.

Exactly. The category and type of <output> is different from
elements in XHTML. It is also defined by someone else, and owned
by someone else. I don't think you should add elements into
a namespace that someone else defines.

This is another reason for strong branding around HTML 5.

> > Tim Bray describes this better than me.  (For
> >full disclosure, the W3C also broke this rule when adding Ruby to the
> >XHTML namespace - but IMO it's still wrong).
> 
>    So really, there isn't a consistent standard for not using 
> namespaces in this situation anyway...

No, I was saying that the W3C goofed up and regret the mistake.
That's partly why I suggest a new namespace.

If you don't make your own namespace, then you might as well 
have no namespaces. You have the root tag (<html>), what else
do you need? (I also think namespaces are not so friendly for
authors, but I'd like to see some way to know if my content is
W3C+WHAT HTML, W3C HTML or W3C+WHO HTML).

> >Alternatively, declare that your HTML 5 is never XHTML, and define a way
> >to make an XML version of HTML 5 (if it is needed), with a different
> >name (and ns). Say XHTML-WHAT-5?  Note that I haven't looked into the
> >trademarking of HTML and XHTML - maybe all this is impossible anyway.
> 
>    There's no real reason to remake XHTML when we can simply add a new 
> module or set of modules to it. After all, we aren't removing existing 
> markup or changing its function in the context of legacy content.

OK. My suggestion here was again more about branding. It seems
weird to have an HTML 5 and then an XML version that is XHTML with
WHAT-ext-mod. However, this is why XHTML is designed that way, so 
I'll accept it.

> 
> >Tim Bray also suggests that you fake the namespaces in HTML
> >(ie <what:output>). I'm with him on this. 
> 
>    What (no pun intended), are we going to do about the people who are 
> switching from HTML5 to XHTML? Do you expect them to have memorized what 
> namespace every element is in?

Why do they have to memorize a namespace for every element?
If their content uses <what:output> then all they need to
do is add xmlns:what on the root element.

I'm suggesting you prefix all your new HTML elements with "what:".
As Tim said, Microsoft have done this without problem in IE for 5 years.

Yeah, it's really ugly. That's why I suggest the alternative: WHAT
should clearly specify HTML 5 (or HTML WHAT) as well as an XML
version.  I think kludging into existing specifications is going to
confuse more people (more than the number of people hand-converting to
XHTML).

> This will significantly add to the XHTML 
> learning curve and make hand conversion of HTML to XHTML more complicated.

I really doubt it. They are going to have to put an XHTML ns
declaration in there anyway. I don't think this will be any harder
than fixing unbalanced content. Do you really think this is a bigger
complication?

Remember that hand converters are going to cut and paste a bit
of text (because no one can remember namespace URIs). Having
one or two NS declarations in that text isn't going to make
the task any harder.

Moving from HTML to XHTML isn't easy, and I really don't think much of
the existing web will ever move (in fact, I doubt much of the existing
web will change to anything -- be it to valid HTML, WHAT HTML or
XAML). What we can hopefully influence is new web content. If it
is produced by tools (eg blogger, typepad or MT), then it should be
possible to get it right most of the time.

> >A final point. How open are these specifications?  The reason I ask is
> >that it *may* be the case that some things here are useful in W3C
> >work. Can we use it, with attribution of course?   At the moment it
> >says (c) Opera, which is fine but I suggest you have some licensing
> >agreement in place as soon as possible.
> 
>    Ian has already asked Opera to put it in public domain. The only 
> reason it hasn't been already is that lawyers are involved.

That's great. Like I said, the main reason I ask is that I think
W3C should be able to use WHAT work if it proves useful.

Dean


More information about the whatwg-whatwg.org mailing list