RE: UDDI's UUIDs issue

Michael,

I agree completely on using existing schemes, hence why I raised this issue.
Interestingly, my work on the TAG has led me down the path of watching
closely for URI schemes and usages.

I took a quick look through their works, and it seems there are basically 3
interesting schemes.  1) is for the domain name, which maybe should be an
http scheme.  2) is for a uuid, which probably should be a uuid scheme. 3) a
structured scheme with a uuid and a key, which I'm not sure about.

I don't think all the facts are on the table yet, certainly I haven't had
the time to track all this down.

Cheers,
Dave

> -----Original Message-----
> From: [email protected] [mailto:[email protected]]On
> Behalf Of Michael Mealling
> Sent: Sunday, November 24, 2002 5:46 PM
> To: Ugo Corda
> Cc: David Orchard; [email protected]
> Subject: RE: UDDI's UUIDs issue
>
>
>
> FYI:
>
> If you plan on using UUIDs in a URI form then you should use the
> developing ISO/ITU standard for how to do it [1]. Its still in process
> due to some issues with the 'variant' field and Microsoft's
> usage of it
> but that's not a huge issue.
>
> IMNSHO, you should only request a new 'uddi:' URI scheme if there's no
> other scheme that can handle what you want. The IESG is taking a much
> harder look at URI registration requests these days and they will
> explicitly ask you why you aren't using the one recognized by
> ISO/ITU...
>
> -MM
>
> [1] http://www.ietf.org/internet-drafts/draft-mealling-uuid-urn-00.txt
>
> On Thu, 2002-11-21 at 18:33, Ugo Corda wrote:
> > >Could you elaborate on why the scheme uddi will be
> registered?  This
> > seems a little bit strange to me.
> >
> > The reason is to make it an official URI scheme maintained by IANA.
> > Why do you think it is strange?
> >
> > >And why would they use the uuid: scheme for uuids rather than
> > creating a new uddi?
> >
> > I am not sure I follow. The scheme used is still the uddi one. For
> > example, a key that used to be represented as UUID
> > "4CD7E4BC-648B-426D-9936-443EAAC8AE23" in version 2, can now be
> > represented in version 3 as
> > "uddi:4CD7E4BC-648B-426D-9936-443EAAC8AE23". (The
> UUID-based format is
> > just one of the possible formats - see [1] and [2]). The
> rationale for
> > this wrapped UUID representation is, I believe, easy migration from
> > version 2 keys (see [3]).
>
>
>
> --
> Michael Mealling <[email protected]>
> VeriSign Labs
>
>