Re: [regrep] Membership requirement for RegistryPackage

[Farrukh Najmi <Farrukh.Najmi@Sun.COM>]

Len,

Thanks for pointing this error in Figure 1. I agree with you that the intent is
for a RegistryPackage to have
0 or more members and to not require at least 1 member. The same is true for
ExternalLinks. An ExternalLink
may exist without any RegistryObjects that it serves as an external link for.

I have noted your issue in the issues log as:

Issue 177: Fig 1 shows that a RegistryPackage must have at least 1 member (1..*).
It should show 0..* instead. Same applies to ExternalLinks

Barring any differeing opinions, we will fix this typographical error in the next
version of the specification.

--
Regards,
Farrukh


Len Gallagher wrote:

> Kathryn and all Registry members,
>
> NIST is preparing some test cases for Registry implementations, and we've
> come upon the following unexpected requirement:
>
> In RIM Section 6, the UML diagram in Figure 1 requires that every
> RegistryPackage have 1 or more members.
>
> The subsequent RIM Sections 6.9 and 7.8 discuss packages further, but these
> sections do not say anything to confirm or deny the requirement that every
> package instance have at least one member.
>
> RIM Section 9.9.2.1 clarifies that package membership is represented by
> Association instances with associationType equal to "HasMember". But there
> is no further discussion to confirm or deny the requirement that every
> package instance have at least one member.
>
> The XML Schema referenced by Registry Services, is also silent on whether
> or not a RegistryPackage instance is required to have at least one member.
>
> We think there is a strong possibility that the requirement of Figure 1 is
> NOT intended. We need for the Registry techncial committee to confirm that
> every package should have at least one member, or state that there is an
> error in Figure 1 of the RIM specification.
>
> We need an answer because it would be very difficult to write the tests to
> cover both situations. The tests will look quite different depending on
> whether or not empty packages are possible.
>
> If we assume that the at least one member requirement is intended, then
> we'll need tests to make sure that a RegistryPackage instance cannot be
> created without simultaneous creation of at least one HasMember Association
> instance. And we'll need tests to make sure that an error is raised if a
> user attempts to delete the last member of a package. Alternatively, if
> empty packages are legal, then it would be possible to create an empty
> package first and later populate it with members.
>
> -- Len
>
> At 09:24 PM 4/3/02, Breininger, Kathryn R wrote:
> >resending...
> >
> > >  -----Original Message-----
> > > From:         Breininger, Kathryn R
> > > Sent: Wednesday, April 03, 2002 8:07 AM
> > > To:   'ebXML Regrep'
> > > Subject:      Comments and resolutions
> > >
> > > I will have the most current Comments and Resolutions log posted on the
> > > Registry TC page near where the Version 2.0 specs are currently located.
> > > This way anyone curious about comments that have come in during the review
> > > period, and how they were resolved, can easily find the information.
> > >
> > > Having a special meeting for commenters is not advisable at this point.
> > > Anyone interested can attend our telecons at any time, since the agenda
> > > and the call in information is published in the e-mail.
> > >
> > > The call-in details and agenda for our meeting tomorrow will follow
> > > shortly.
> > >
> > > Kathryn
> > >
> > >
> > > Kathryn Breininger
> > > CENTRAL Project Manager
> > > Emerging Technologies
> > > Boeing Library Services
> > >
> > > 425-965-0182 phone
> > > 425-237-3491 fax
> > > kathryn.r.breininger@boeing.com
> > >
> >
> >----------------------------------------------------------------
> >To subscribe or unsubscribe from this elist use the subscription
> >manager: <http://lists.oasis-open.org/ob/adm.pl>
>
> **************************************************************
> Len Gallagher                             LGallagher@nist.gov
> NIST                                      Work: 301-975-3251
> Bldg 820  Room 562                        Home: 301-424-1928
> Gaithersburg, MD 20899-8970 USA           Fax: 301-948-6213
> **************************************************************
>
> ----------------------------------------------------------------
> To subscribe or unsubscribe from this elist use the subscription
> manager: <http://lists.oasis-open.org/ob/adm.pl>