RE: IRIs everywhere (including XML namespaces)

> From: www-tag-request@w3.org [mailto:www-tag-request@w3.org]On Behalf Of
> Chris Lilley
> Sent: Wednesday, October 30, 2002 9:40 PM
> To: www-tag@w3.org; Julian Reschke
> Subject: Re: IRIs everywhere (including XML namespaces)
>
>
>
> On Wednesday, October 30, 2002, 9:27:57 AM, Julian wrote:
>
>
> JR> Hi.
>
> JR> If XML 1.1 namespaces allows IRIs as namespace names, and
> IRIs may contain
> JR> unescaped space characters, applications that assume that they can use
> JR> whitespace to delimit namespace names will break, right?
>
> JR> For instance:
>
> JR> http://www.w3.org/TR/xmlschema-1/#schema-loc
>
> Correct (I brought that one up on the TAG call a couple of days ago to
> general cries of 'ouch' etc).

I confess that I mentioned it here because I remembered having seen it
somewhere, but couldn't remember where :-). This explains it.

> Further, if you escape the space and then compare it to an IRI where
> the space was not escaped, they compare as not identical.
>
> If you escape the space between the two items, then its no longer a
> list.
>
> Of course, if XML Schema allowed a choice of list delimiters instead
> of just space (commas, semicolons, etc) then unescaped spaces would be
> ok because you could use some character that was not in that
> particular IRI to delimit the list (same as the " ' " and ' " '
> technique for attribute values).

Sure, it's a specific case. Not every XML application will be affected in
the same way. The main point being, replacing URI refs by IRI  refs *may*
break existing applications, and I'm not sure that the benefits are worth
it.

BTW: up until recently, I thought that IRIs are just URIs that allow
"arbitrary" Unicode for the sake of I18N. Why allow the space character
then?

--
<green/>bytes GmbH -- http://www.greenbytes.de -- tel:+492512807760

Received on Wednesday, 30 October 2002 16:07:42 UTC