WG: Issue: Attribute Multiplicity

["Zimmermann, Peter" <Peter.Zimmermann@m.dasa.de>]

Hi all,

some little comments/questions on Lofton's proposal:
*	Why should name be multiple? Is it intended to hold several natural
language variants?
*	Sure, multiple attribute occurrences can't be expressed in XML as
such, but there are two other alternatives:
	- use elements in the companion metadata (which I would prefer), or
	- use attribute type NMTOKENS (limited char set as you need to have
separators, e.g. spaces).

Regards,
Peter Zimmermann
EADS Military Aircraft
Customer & Product Support
D-81663 Munich
Tel.: +49-(0)89-607-21738
Fax: +49-(0)89-607-21875
Email: Peter.Zimmermann@m.dasa.de
=============================================

> -----Ursprüngliche Nachricht-----
> Von:	Lofton Henderson [SMTP:lofton@rockynet.com]
> Gesendet am:	Freitag, 19. Januar 2001 00:48
> An:	cgmopen-members@lists.oasis-open.org
> Betreff:	Issue:  Attribute Multiplicity
> 
> CGM open members --
> 
> Here is a WebCGM clarification issue which might be a little more 
> controversial than the set of editorial/routine comments in my previous 
> mail.   WebCGM 1.0 Release 2 ought to clarify these points.
> 
> CLARIFICATION/ISSUE.  Multiplicity of APS attributes.  Which ones can be 
> multiple?  I thought we went through all of them and indicated in the
> spec, 
> single/multiple in each subsection 3.2.2.1 - 3.2.2.8.  But the following 
> five APS attributes do not state explicitly.  In parentheses, I have 
> indicated what I recall from earlier discussion, and/or recommend.
> 
> Region (recall/recommend:  single).
> ViewContext (recall/recommend:  single)
> LayerName (recall/recommend:  single).
> LayerDescription (recall/recommend:  single).
> Name (recall/recommend:  multiple).
> 
> For reference, the spec is explicit about:
> 
> single:  ScreenTip, Content.
> multiple:  linkURI
> 
> Note:  I understand the arguments against multiple attribute occurrences 
> (such a CGM is not expressible in XML).  But I don't think we can can fix 
> this in Release 2, certainly not if ProfileEd remains as 1.0.  There is 
> legacy content using it.  Rather, we need to address it in ProfileEd 2.0
> or 
> 1.1 or ...  For example, we could allow multiplicity as a list within the 
> SDR, and deprecate multiple occurrences.
> 
> Regards,
> Lofton.
> 
> *******************
> Lofton Henderson
> 1919 Fourteenth St., #604
> Boulder, CO   80302
> 
> Phone:  303-449-8728
> Email:  lofton@rockynet.com
> *******************