[vodml] Attribute multiplicity

[Gerard Lemson]

Hi Arnold and rest
After the change mentioned by Mark, there is currently no reason for supporting [0..<some other attribute>] multiplicities anymore.
As it would make the models even uglier (no offense intended, just stating my opinion), I will keep poo-pooing (though this may be dangerous, https://www.youtube.com/watch?v=QeF1JO7Ki8E&t=29) the idea.
Cheers
Gerard


> -----Original Message-----
> From: Arnold Rots [mailto:arots at cfa.harvard.edu]
> Sent: Friday, February 19, 2016 4:08 PM
> To: CresitelloDittmar, Mark <mdittmar at cfa.harvard.edu>
> Cc: Gerard Lemson <glemson1 at jhu.edu>; Data Models mailing list
> <dm at ivoa.net>
> Subject: Re: [vodml] Attribute multiplicity
> 
> I want to repeat a suggestion that I made earlier. even though I know
> 
> that it was pooh-poohed by Gerard. But I still think it could ameliorate
> 
> teh concerns and it did not really receive serious consideration.
> 
> 
> Right now (i.e., before the message Mark just posted) we have the following
> options for attribute multiplicity:
> 
> n
> 
> 0,n
> 
> where n is a positive integer.
> 
> Mark's message would also allow:
> 
> 0..n
> 0..*
> 
> What I had suggested is to restrict this somewhat by requiring that
> 
> any indeterminate (that is, in the model) attribute multiplicity be
> 
> explicitly stated through a non-negative attribute, replacing
> 
> 0..n and 0..* simply by:
> 
> n
> 
> as is currently allowed, but with n being either a literal or a non-negative integer
> attribute, like
> 
> n:nonnegativeInteger
> 
> a:real{multiplicity=n}
> 
> 
> So, the question is: would this help in preventing bad modeling?
> 
> I thought it might, since it makes the model less indeterminate.
> 
> 
> Cheers,
> 
> 
>   - Arnold
> 
> 
> 
> --------------------------------------------------------------------------------------------------
> -----------
> Arnold H. Rots                                          Chandra X-ray Science Center
> Smithsonian Astrophysical Observatory                   tel:  +1 617 496 7701
> 60 Garden Street, MS 67                                      fax:  +1 617 495 7356
> Cambridge, MA 02138                                         arots at cfa.harvard.edu
> <mailto:arots at cfa.harvard.edu>
> USA                                                   http://hea-www.harvard.edu/~arots/
> --------------------------------------------------------------------------------------------------
> ------------
> 
> 
> 
> On Fri, Feb 19, 2016 at 3:47 PM, CresitelloDittmar, Mark
> <mdittmar at cfa.harvard.edu <mailto:mdittmar at cfa.harvard.edu> > wrote:
> 
> 
> 	All,
> 
> 
> 	Update on this topic:
> 
> 	  In a recent focus-meeting in Baltimore, this topic was heavily discussed
> and explored.
> 
> 
> 
> 	The primary reason for the restriction is that it typically will catch an
> instance of 'bad' modeling,
> 
> 	where a concept has been denormalized to a simple type.  It is not that
> there is something
> 
> 	inherently wrong about the open multiplicity.  It was also noted, that
> the same 'bad' modeling
> 
> 	is not caught when the multiplicity is a fixed length.
> 
> 
> 	However, there seems to be areas where it MAY be useful to have such
> multiplicity.
> 
> 	In reviewing the concequences of needing to allow this condition later,
> it seems that
> 
> 	it may be large.. effecting vo-dml spec, and possibly votable spec as
> well.
> 
> 
> 	In the end, it was agreed that vo-dml should be modified to allow the
> open multiplicity on attribute types.
> 
> 	It was agreed that the document should be be very clear about the
> modeling concern and that the
> 
> 	condition should be strongly discouraged.  As a procedural matter,
> occurances should be discussed
> 
> 	within the group for alternative representations.
> 
> 
> 	Actions:
> 
> 	  GL - update the specification accordingly.
> 
> 	         update xslt scripts to issue WARNING when the condition is
> detected as a reminder to
> 
> 	         review the modeling.
> 
> 
> 
> 	Mark
> 
> 
>