Issue on modules - mapping

[Julian Fowler]

Rogerio et al

Apologies for the delay in replying to your message.

I have three comments on this issue:

1. From a QC perspective I see no problem with a module having an 
ABSTRACT SUPERTYPE in its ARM that has either an incomplete mapping 
or no mapping at all -- as long as the module documentation is clear 
and indicates how the ARM is to be completed.

2. The revised mapping specification guidelines do cover the case 
where the mapping of an ARM ABSTRACT SUPERTYPE is specified in its 
subtypes; however, currently this deals only with references to 
subtype mappings within the same mapping specification. I believe 
that this therefore requires an extension to the mapping 
specification capability, intended for use in modules only, that 
states explicitly that a mapping is incomplete and is to be found in 
other (unspecified) documents.

3. If the ARM ABSTRACT SUPERTYPE has any attributes or relationships 
that map consistently for all *possible* subtypes then these should 
be mapped in the "declaring" module. This is consistent with the 
approach to subtype/supertype mapping agreed at the QC meetings in 
Charleston. This implies that mappings of subtypes in other modules 
cover only the aspects of the mapping that are specific to that 
subtype. QC has agreed that mappings of "inherited" 
attributes/relationships may be repeated in subtypes (using the 
/SUPERTYPE/ template). However, this must be done consistently -- 
i.e., if this approach is taken then mappings for *all* inherited 
attributes/relationships shall be repeated.

regards
Julian

On 2000-11-08 at 9:43, you wrote:

> 
> Julian,
> 
> During the shape appearance and layers modules ballot (1001-1009) the
> following issue was submitted:
> 
> Issue Reference No.: GER-1001-14
>  Originator:	Ullrich Pfeifer - pfeifer at dik.tu-darmstadt.de
> Date:	
>  Urgency/Type:	major, technical	Page:	clause 5.1
>  Sentence/abstract/keywords:
>  Issue Description:
>  there is no mapping for appearance
> 
>  Proposed Solutions/Remarks:
>  add mapping or explanation
> 
>  Resolution:
> Entity appearance is an abstract supertype whose subtypes are only
> known in modules that use 1001 (e.g., part 1009 - Shape Appearance and
> Layers). As a result, it is not mapped in part 1001. This issue will
> be forwarded to the Quality Committee.
> 
> 
> I have been told to get input from QC on this issue. My point of view
> is that, in modules, we may have abstract types that are placeholders
> and therefore the mapping is not known. These abstract types will have
> subtypes in modules that reference the module being specified. The
> mapping therefore should be defined in the subtypes.
> 
> I would appreciate your input on this issue.
> 
> Regards,
> 
> Rogerio
> 
> 
> ----------------------------------------------------------------------
> ------ -------------------
>  Rogerio Barra, Ph.D.			mailto:barra at aticorp.org
>  Advanced Technology Institute		Voice : (843) 760-3378
>  5300 International Blvd.		Fax    : (843)760-3349
>  N. Charleston, SC  29418		http://www.aticorp.org/
> 
> 


-- 
Julian Fowler, PDT Solutions
Telephone: +44 15242 63389 Mobile: +44 7939 276005
Fax: +44 870 052 3414 Email: jfowler at pdtsolutions.co.uk
http://www.pdtsolutions.co.uk