W3C home > Mailing lists > Public > xml-names-editor@w3.org > October 2002

Re: IRIs everywhere (including XML namespaces)

From: Elliotte Rusty Harold <elharo@metalab.unc.edu>
Date: Sun, 13 Oct 2002 09:16:29 -0400
Message-Id: <p04330101b9cf1f2179f7@[192.168.254.4]>
To: Misha.Wolf@reuters.com
Cc: www-tag@w3.org, xml-names-editor@w3.org, w3c-i18n-ig@w3.org

At 9:08 AM +0100 10/13/02, Misha.Wolf@reuters.com wrote:


>You write about developers who will be made to jump through various
>hoops as a consequence of XML namespaces following other W3C
>specifications and using IRIs.  What software components are these folks
>likely to be developing?  I would estimate that each OS would need one
>component which translates IRIs to URIs.  You seem to be suggesting many
>such components.

I am pointing out that OS's and APIs do not now have such components; 
or, worse yet, they have them but they're broken. If each OS did have 
such a component and it was reliable, this would be a lot less 
problematic. Perhaps some OSs and APIs do have such a component. I 
know Java does not. A standard component or code from the W3C that 
developers could copy and paste would help a lot here. Doing the 
right thing should be easier than doing the wrong thing.

As to software components these developers are likely to be 
developing, these include anything that uses IRIs, including:

browsers with XLink support
XInclude processors
RDDL processors
schema validators
XML parsers
web servers
web spiders

>In an earlier mail [1], you wrote:
>>  >- XML 1.0 system identifiers [2]
>>
>>  This references an erratum, not the actual spec. The original 1st
>>  edition spec is clear though that what we now call an IRI is allowed;
>>  that the escaping is performed by the processor as necessary, not the
>>  author.
>
>I was puzzled then, as you seemed to be suggesting that, in contrast
>with XML's usage of IRIs, in the case of XML namespaces, the author
>would need to perform the IRI-to-URI conversion.  At the time I didn't
>respond, but your latest mail suggests that we have very different
>mental models of the way in which XML namespaces would use IRIs.  In my
>model, the IRIs are not converted to URIs unless and until they are to
>be dereferenced.  Please explain your model.

I agree. However, in the case of XML, most users are relying on a 
parser to do the conversion for them. They are not themselves writing 
the code to retrieve an external entity at a particular URI. In 
namespaces, they might well be. However, namespaces are probably one 
of the more sanguine uses of IRIs since resolution is the exception, 
not the rule. Mostly, this question led me to look at the deeper 
implications for IRIs across the spectrum of web technologies. It 
seems to me that there are some architectural issues here that have 
not previously been hashed out, and that affect a lot of 
technologies, not just namespaces.

That said, I am worried that at least some existing parsers may not 
be making the proper IRI-to-URI conversion. I need to write some test 
cases. Does anyone know if the existing test cases for XML check this 
requirement? Any parser vendors want to chime in and say "Yes, we do 
handle this," or "No, we don't"?
-- 

+-----------------------+------------------------+-------------------+
| Elliotte Rusty Harold | elharo@metalab.unc.edu | Writer/Programmer |
+-----------------------+------------------------+-------------------+
|          XML in a  Nutshell, 2nd Edition (O'Reilly, 2002)          |
|              http://www.cafeconleche.org/books/xian2/              |
|  http://www.amazon.com/exec/obidos/ISBN%3D0596002920/cafeaulaitA/  |
+----------------------------------+---------------------------------+
|  Read Cafe au Lait for Java News:  http://www.cafeaulait.org/      |
|  Read Cafe con Leche for XML News: http://www.cafeconleche.org/    |
+----------------------------------+---------------------------------+
Received on Sunday, 13 October 2002 09:21:07 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 19:56:48 UTC