W3C home > Mailing lists > Public > www-html@w3.org > February 2000

RE: naming custom/extended tags

From: Arjun Ray <aray@q2.net>
Date: Fri, 11 Feb 2000 19:22:43 -0500 (EST)
To: www-html@w3.org
Message-ID: <Pine.LNX.4.10.10002111856220.31536-100000@mail.q2.net>

On Fri, 11 Feb 2000 JOrendorff@ixl.com wrote:

> Michael Blow trolled:

That's a bit extreme, don't you think?

> > Are there any W3C recommendations on how to name a custom tag?

No recommendation, but some helpful hints (i.e. ways to think about
the problem):


> > We need to define a few custom tags, and want to following any
> > naming standards that may be in place, to minimize potential
> > conflict.

Naming is the least of your problems.

> This is going to be a controversial question.

Possibly.  There's a lot of NIH involved.

> HTML wasn't meant to be extended in this way.  So whatever you do,
> you might have conflicts later.  

Yes and no.  It's possible to avoid conflicts down the road, but the
groundwork for this hasn't been covered yet (and, with current trends,
may not be for quite a while.) 

> XML has a feature (XML Namespaces -
> http://www.w3.org/TR/REC-xml-names) that eliminates the problem.

It's a bogosity, and it doesn't eliminate the problem.

> You've probably heard:  the gurus are now recommending that you use
> XHTML rather than HTML anyway.

The gurus?  All I see is W3C advocacy.

> If you just can't do XML, for whatever reason, then you're stuck
> with hacking traditional HTML (or "Tag Soup" is it is called in
> this venue.) The best thing you can do is pick a prefix and stick
> to it.  The prefix should be a few letters (or whatever) followed
> by a colon or hyphen.

Good advice.  But avoid the colon.  It's "reserved for namespace
experiments".  Emphasis on 'experiments' optional but advisable:)

> So if your product is called the Greyhound Media Blaster, then you
> have:
>   <gmb-video ... >
>   <gmb-noise ... >
>   <gmb-odor  ... >

It may not mean much now, but I'd suggest adding two attributes to
each such homegrown element type:

    html  -  the value should be a HTML element type, to indicate
             a "fallback" approximation, usually something generic
             like 'p', 'div', 'span', etc.
    htign -  with one of two values (htign|empty): to indicate whether
             and how the content is treatable in fallback HTML
             fashion. The default here is non-assertion, which means
             that the content is treatable regularly.

I've tried before to explain the point of a convention like this, so I
won't go over it in detail again.


Received on Friday, 11 February 2000 19:02:52 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:05:52 UTC