RE: [uddi-spec] First draft of CR-036 (changes to key hashing) uploaded

["John Colgrave" <colgrave@hursley.ibm.com>]

Luc,

I saved a business to my local UDDI registry and that is the key that was
assigned to it.  As such it is a random UUID, as the original byte-sequence
was, but this time it is in network byte order.  I don't think we need text
to say that this should be put in network byte order, the existing text is
OK suggesting that the byte sequence be used exactly as given.  The only
problem was that the byte sequence was not given in the correct order.  The
new one is.

I deliberately avoided the issue of whether the V1/V2 keys presented in the
WSDL TN were derived or evolved and given that the hashing scheme is only a
RECOMMENDED approach then I think that is reasonable.  If a registry
implements some other scheme then it will have to treat the keys as evolved.
This will also be the case for registries that implement the old scheme.

John Colgrave
IBM


-----Original Message-----
From: Luc Clement [mailto:lclement@windows.microsoft.com] 
Sent: 23 June 2003 22:28
To: John Colgrave; uddi-spec@lists.oasis-open.org
Subject: RE: [uddi-spec] First draft of CR-036 (changes to key hashing)
uploaded

Thanks John for this posting. Could you please explain how you arrived to
"14 e3 a2 b1 3b d8 4c f5 af a6 0d 14 1b f3 20 76" (i.e. how was it
generated, what order is it being represented in?)

Secondly, given the confusion, I think that the text of the CR and the
errata needs to be clear that this key needs to put in network byte order.
Your CR does not reflect this update to the spec as I think it should.

FYI, please also note that once CR-032 and CR-036 are approved that we'll be
in a position to put to vote the WSDL v2 TN. If approved, this TN will be
posted before the first v3 errata and thus, we'll need some supporting text
on the web page explaining that the keys presented in the TN are those based
on a corrected version of the hashing algo. 

Luc Clément 
Microsoft 
Co-chair, OASIS UDDI Spec TC 



-----Original Message-----
From: John Colgrave [mailto:colgrave@hursley.ibm.com] 
Sent: Wednesday, June 18, 2003 09:20
To: uddi-spec@lists.oasis-open.org
Subject: [uddi-spec] First draft of CR-036 (changes to key hashing) uploaded

I have uploaded the first draft of CR-036.

As agreed, I have not changed the table.

I have not updated the reference to the Leach draft as I believe that was
incorporated into CR-002.

An updated version of the Leach draft formed the basis of the registration
document for a UUID URN Namespace, but that document expired on April 1,
2003.  If you want to look at it anyway, you can find it at
http://www.ietf.org/internet-drafts/draft-mealling-uuid-urn-00.txt

There is an online version of the description of the original DCE UUIDs
available at http://www.opengroup.org

I ended up regenerating keys for 41 entities, all but one of them tModels.
Of these, only 9 or 10 would not have been affected by the case-folding
change so I think we did the right thing is changing the name space ID used
in the algorithm.

I think it would be a good idea for someone, ideally with a "little-endian"
implementation, to update it with the new name space ID and the
case-folding, and check that they produce the same keys.

John Colgrave
IBM




You may leave a Technical Committee at any time by visiting
http://www.oasis-open.org/apps/org/workgroup/uddi-spec/members/leave_workgro
up.php