[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: HTML5 and almost no namespaces

• From: Mukul Gandhi <gandhi.mukul@gmail.com>
• To: Michael Kay <mike@saxonica.com>
• Date: Fri, 3 Jun 2011 14:24:32 +0530

On Thu, Jun 2, 2011 at 4:37 PM, Michael Kay <mike@saxonica.com> wrote:
> You're saying you like namespaces because they are a solution to a real
> problem.
>
> But the objection to namespaces is not that there is no problem to be
> solved, it is that namespaces (as currently defined) are a bad solution to
> that problem; they introduce far too much complexity in relation to the size
> of the problem. Far simpler solutions are available to allow names to be
> disambiguated.


Are you suggesting that current XML namespaces are bad for every known
uses of XML?

The current XML namespaces to define language vocabularies (for e.g as follows)

<xsl:stylesheet xmlns:xsl="..">
    ...
</xsl:stylesheet>

or

<xs:schema xmlns:xs="..">
  ...
</xs:schema>

looks just fine to me. I think, it'll be very difficult for users to
adjust to any other style than above for XML vocabularies. The other
issue is carrying on with the legacy vocabularies and the
corresponding instances in design & which are already deployed.
Therefore backward compatibility would be quite important, if we want
to define any new extensions in this area.

But I think, it may be good if optionally we may have XML namespace
values to be able to use any string, or perhaps any typed simpleType
(in XSD terminology) value.

For e.g, the namespace value could be a UUID value, as follows.

<x:root xmlns:x="7570b77c-8a79-419b-bed2-2c3c99203be5">
   ...
</x:root>

(because sometimes, the namespace names may not have something to do
with URIs or HTTP etc).

The advantage with URIs for namespace values I think, is that they can
easily be available as unique (the owners of the XML vocabularies need
a [globally] unique namespace value for their [XML] languages).

But I think, that the namespace scoping rules and concepts of prefixes
and "xmlns" bindings should largely be retained (in a new version, or
a profile of the current one if they are ever designed).




-- 
Regards,
Mukul Gandhi

• Follow-Ups:
  • Re: HTML5 and almost no namespaces
    • From: Michael Kay <mike@saxonica.com>

• References:
  • HTML5 and almost no namespaces
    • From: Jesper Tverskov <jesper.tverskov@gmail.com>
  • Re: HTML5 and almost no namespaces
    • From: Stephen D Green <stephengreenubl@gmail.com>
  • Re: HTML5 and almost no namespaces
    • From: Mukul Gandhi <gandhi.mukul@gmail.com>
  • Re: HTML5 and almost no namespaces
    • From: Michael Kay <mike@saxonica.com>