W3C home > Mailing lists > Public > www-style@w3.org > September 2010

Re: Extend use of namespaces

From: Paul Duffin <pduffin@volantis.com>
Date: Sun, 19 Sep 2010 10:23:09 -0600 (MDT)
To: Anne van Kesteren <annevk@opera.com>
Cc: Sylvain Galineau <sylvaing@microsoft.com>, www-style <www-style@w3.org>, "Tab Atkins Jr." <jackalmage@gmail.com>
Message-ID: <975642887.20210.1284913389355.JavaMail.root@zimbra.volantis.com>

----- Original Message -----
> On Sun, 19 Sep 2010 12:47:14 +0200, Paul Duffin <pduffin@volantis.com>
> wrote:
> I'm not advocating adding special cases, I just have not seen the
> need.

What is :svg except a special case where you add a special feature into CSS that only applies to one language. I can't think of one other CSS feature that is specific to a markup language.

> And so far all your arguments are from theory. From theory it is very
> easy
> to imagine the most complicated systems. (Which is one of the reasons
> we
> ended up with namespaces in the first place. There would be hundreds
> to
> millions of markup languages requiring a URL because otherwise the
> identifier would not be unique enough, yadayada.)

So HTML + MathML + SVG is a theory, what about XHTML 2 + XForms.

This is not an argument from theory, I create and maintain maybe a dozen schemas with their own namespaces. They consist of standards based schemas (over which I have no control) and proprietary ones. There would be no way I could combine them together without namespaces. Or style them without namespace support in CSS. As far as I know none of our customers have had any serious problems with our use of namespaces, sure they have forgotten to add one, or used the wrong URL but that is just people being people. What causes me pain, day to day is not the namespaces but the schema definitions in XML Schema but that is another issue.

> > Namespaces are not as complicated as you make out, neither from an
> > author's perspective or from an implementation perspective.
> I think you are wrong.
> http://wiki.whatwg.org/wiki/Namespace_confusion

I did a brief scan of them and the majority of them were to do with bugs in implementations, the syntax used, lack of documentation, or abuse / misuse / errors in how they are used. Most of the bugs referenced are old and have been fixed. I don't think that the way XML Namespaces were added to XML was the best, in particular how they have been exposed in the various APIs. However, I think that page contains more FUD than providing useful information. Just because someone happened to forget to put a prefix in some sample code is no argument against namespaces.

I saw an argument against HTML using namespaces, that was not driven by some fundamental problem with them but simply that they had been misused in existing HTML and XHTML pages so that they would break if they were run on browsers that did implement them properly. That is not an argument against namespaces as such but just the syntax used to represent them.

If namespaces are so bad then how come they have been used in just about every single XML based specification that the W3C have created, from XHTML 2 to XBL, from XSL to XQuery. Most of those specifications would be very very different and far less elegant without namespaces.

> More complex is a disadvantage I have pointed out. Maybe you do not
> accept
> it, but please...

Complexity is a subjective measurement. Personally, I think that something like :nth-of-type() is far more complex, and way more confusing. But then again as I said I live and breathe namespaces and schema so maybe I am an exception.

> > [...]
> I have no interest in repeating this discussion. It has already been
> had
> here in the past, recently even:
> http://lists.w3.org/Archives/Public/www-style/2010Jul/thread.html#msg48

I have just read the whole discussion and it seemed to cover all the issues I raised but as far as I can tell no consensus was reached, no changes were made, it just fizzled out. It did cover all of the issues I raised and probably more besides.

Is the process described in http://www.alistapart.com/articles/prefix-or-posthack/ going to be followed or not? 

If not then how is the problem going to be resolved without discussion.

> Except there is no need to prefix it with the whole draft identifier.
> It
> could just be -wd1-, -wd2-, etc. as different drafts would not define
> the
> same property twice. And if they would we'd just increase the number.

That would work too.

> Anyway, we are not going in that direction.

Received on Sunday, 19 September 2010 16:23:43 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:49:47 UTC