W3C home > Mailing lists > Public > public-iri@w3.org > May 2004

Re: draft-duerst-iri-07.txt: 2 week mailing list last call

From: Martin Duerst <duerst@w3.org>
Date: Thu, 20 May 2004 12:03:51 +0900
Message-Id: <>
To: John Cowan <cowan@ccil.org>
Cc: Graham Klyne <gk@ninebynine.org>, public-iri@w3.org

Hello John,

At 07:08 04/05/19 -0400, John Cowan wrote:

>Martin Duerst scripsit:
> > Good point. This has been brought up in various discussions e.g. related
> > to namespaces. In terms of namespaces, it reads: Namespaces compare
> > character-by-character, so don't chage escaping. But don't try to use
> > two namespaces that only differ by escaping, because that would just
> > be stupid and would lead to confusion.
>I have proposed that the XML Core WG add the following motherhood note
>to both NS 1.0 and NS 1.1:

I think it is a very good idea to add such a note. Another, related
thing is that namespace documents should actually contain the
namespace URI/IRI literally, rather than just rely on where they
are, because they might be found with a slightly different URI/IRI.
But I guess that's too much of implementation advice that probably
doesn't belong in the spec.

>         Note:  Although namespace names are only identical if they
>         consist of the same characters in the same order, other uses of
>         URIs provide different rules for identity.

I would probably try to avoid the word 'identity'.

>Therefore, namespace
>         names SHOULD NOT:

I think it is better to refer to the relevant section of RFC 2396bis,
and to wait for that to become an RFC. Several of your points are
not alligned with RFC 2396bis, and you risk to be out of sync if
there are any changes.

>         a) use upper case characters in the scheme name;
>         b) percent-encode any character that does not require it;
>         c) use upper case letters in the host name;
>         d) contain the sequences "/../" or "/./";
>         e) specify an explicit port number that is equal to the default;
>         f) end in "/" or "#" (namespace names used by RDF are an exception);

I think the '/' end is the default for some things in RFC 2396bis.

>         g) make use of the "file:" scheme, since its meaning is not absolute.

This is the only point that is not discussed in section 6 of RFC 2396bis.
But it should be more general, there may be other schemes that are similar
to "file:".

>Of course, I'm not allowed to tell you what they think of the idea.  ;-)
>Improvements are solicited.

You should send this also to the URI list (uri@w3.org) for feedback.

One might also be able to argue that with the new URI spec and IRI spec,
the namespace spec(s) wouldn't need to say things like this anymore.
But it's probably still better to mention it.

Regards,    Martin.

>Cowan  jcowan@reutershealth.com  www.ccil.org/~cowan  www.reutershealth.com
>"If I have seen farther than others, it is because I was standing on
>the shoulders of giants."
>         --Isaac Newton
Received on Wednesday, 19 May 2004 23:12:47 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:14:31 UTC