Thanks Farrukh,
Sounds good to me.
--
Regards,
Ivan
-----Message d'origine-----
De : Farrukh S. Najmi [mailto:]
Envoyé : jeudi 25 janvier 2007 17:11
À : BEDINI Ivan RD-BIZZ-CAE
Cc : ebXML Regrep
Objet : Re: [regrep] regrep 1/24/2007: Question on RS Section 3.1 Errata(PublicComment)
BEDINI Ivan RD-BIZZ-CAE wrote:
> Exactly,
> I agree with the latest modification.
>
> Another concern now is about the attribute LID definition, as follow:
> <ebRIM3.0_Section_2.5.6 Attribute lid> ...If the submitter assigns the
> lid attribute, she must guarantee that it is a globally unique URN...
> </ebRIM3.0_Section_2.5.6 Attribute lid>
>
> This could seem inconsistent with the new text and not clear about the action to do. Not sure about what means *globally unique*.
> If my interpretation is correct, in reality it is not so because physically two differents UUID imply two diferents RegistryObjects (R01 and RO2) with the same LID => the LID is not globally unique.
> What must be guaranteed is that in this case => Twoor more RO with the
> same LID must have (at least) a different version number (i.e. for the
> same LID => RO1 version must be different from RO2 version)
>
>
The complete text says:
"The lid attribute MAY be specified by the submitter when creating the original version of a RegistryObject. If the submitter assigns the lid attribute, she must guarantee that it is a globally unique URN."
The text applies to specifying lid when submitting the "original version of a RegistryObject". In this case it MUST be globally unique.
Perhaps a clarity could be:
"The lid attribute MAY be specified by the submitter when creating the original version of a RegistryObject. If the submitter assigns the lid attribute *when submitting the original version of a RegistryObject*, she must guarantee that it is a globally unique URN."
--
Regards,
Farrukh
Web: http://www.wellfleetsoftware.com
