[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [regrep] Issue with UUID's for Core Components and BIE's
Farrukh, OK. Then some further suggestions - define a default classification for a LID - that is used if one is not provided - something like "registry-system-reference" or such (this is probably handy anyway for a bunch of other stuff!). I've actually found that RR should ship with some default classifications built-in for system admin' purposes - this could be part of those - and you could give non-normative notes on these for implementers. In a similar vein - if you are going to autogenerate LIDs for people - they should be able to associate a prefix alias with their user profile - to be used when creating those default LIDs. This also works great for people working for a customer - (say USPS) - then they will simply get the next available LID in that series - by association. When they setup an account this should be part of the configuration dialogue required. Thanks, DW ----- Original Message ----- From: "Farrukh Najmi" <Farrukh.Najmi@Sun.COM> To: "David RR Webber" <david@drrw.info> Cc: <regrep@lists.oasis-open.org> Sent: Friday, April 09, 2004 4:17 PM Subject: Re: [regrep] Issue with UUID's for Core Components and BIE's > David RR Webber wrote: > > >Farrukh, > > > >I still think its better to upgrade the External ID. > > > >And less confusing -if you have two things doing > >related and overlapping things - which do people > >pick - or do they put the same value in both > >places - probably - and therefore its a clue > >it should be just one field. > > > > > I agree that similar but different things are confusing and bad. > I have found that ExternalIdentifiers are the least used feature of > RIM. I personally do not like them the way they are. I think it would > have been better to have their type be Slot. In other words a RegistryObject > has an externalIdentifiers attribute that is a Collection of Slots and also > a Slots attribute that is a Collection of Slots. The first has a > specific meaning > (external identiofiers) while the later could have any meaning wahtsoever. > > I absolutely do not like the fact that to have a simple value such as a > LID you need > to define a LID ClassificationScheme. > > Also ExternalIdentifier is an extensibility mechanism similar to Slots > that can be > just about anything. While the LID is a much tighter first class part of > the model > similar to id. It is just so fundemental and basic taht we want it to be > reachable by the most direct path. > > >Not sure this should be mandatory - blanks > >should be allowed - since people may insert > >content that does not need a LID - or may > >choose to assign one at some later point > >in the process. > > > > > According to the proposal a LID is required but if the client does not > provide it the registry generates one. > Things like IDs logical or otherwise should not be late bound in my > opinion. There are too many issues > that crop up when you do that. > > >I'm completely baffled as to why you think a > >separate LID field is more efficient - queries > >should be just as fast either way. > > > > > Because it is part of the object and does not require a join of two > objects to assemble it (say in a relational model). > It is fetched in the same access as teh rest of the object. That said I > am not a big fanb of premature optimizations. > > >Also - surely everything inserted into the registry > >requires a classification so - I'm not seeing any > >difference here? > > > > > Actually my experience is that most things in the registry are not > classified. > > >Thanks, DW. > > > >----- Original Message ----- > >From: "Farrukh Najmi" <Farrukh.Najmi@Sun.COM> > >To: "David RR Webber" <david@drrw.info> > >Cc: <regrep@lists.oasis-open.org> > >Sent: Friday, April 09, 2004 3:56 PM > >Subject: Re: [regrep] Issue with UUID's for Core Components and BIE's > > > > > > > > > >>David RR Webber wrote: > >> > >> > >> > >>>Farrukh, > >>> > >>>Why not overlay the LID functionality on the External ID? > >>> > >>>No need to change the RIM then - and this is in any case > >>>the behaviour the External ID was intending to have all > >>>along ; -) > >>> > >>> > >>> > >>> > >>Good observation David. > >> > >>I considered that very point but decided against it because every object > >>needs to have a LID (like an id) and given > >>that ExternalId requires a ClassificationScheme it seemed better for > >>queries to > >>introduce a new attribute. It would also be more efficient for storage > >>and retrieval. > >> > >> > >> > >>>Thanks, DW. > >>> > >>>----- Original Message ----- > > >>> > >>> > >>> > >>> > >>>>Please see the latest Versioning proposal posted after Nikola and I > >>>>worked on the last one. The Logical ID (LID) addresses the > >>>>need for a human friendly ID. Check it out at: > >>>> > >>>> > >>>> > >>>> > >>>> > >>>> > >>http://www.oasis-open.org/apps/org/workgroup/regrep/download.php/6326/Vers i > >> > >> > >onControlProposal-03.doc > > > > > >>> > >>> > >>>>-- > >>>>Regards, > >>>>Farrukh > >>>> > >>>> > >>>> > >>>>To unsubscribe from this mailing list (and be removed from the roster of > >>>> > >>>> > >>>> > >>>> > >>>the OASIS TC), go to > >>> > >>> > >>http://www.oasis-open.org/apps/org/workgroup/regrep/members/leave_workgrou p > >> > >> > >.php. > > > > > >>> > >>> > >>>> > >>>> > >>> > >>> > >>> > >>-- > >>Regards, > >>Farrukh > >> > >> > >> > >> > >> > > > > > > > > > -- > Regards, > Farrukh > > >
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]