Re: [xtm-wg] 1.0 Spec Comments

[Steve Pepper <pepper@ontopia.net>]

At 10:47 07/02/01 -0600, Eric Freese wrote:
>Section 2.1 - para starting with "Because associations express..." -
>Last sentence states "Relationships may involve one, two or more roles. ".
>However the DTD says that the <roleSpec> subelement is optional.  Are these
>in conflict?

And Daniel Rivers-Moore replied:
>In response to Eric's question on the optionality of roleSpec ...
>
>roleSpec was introduced when we were discussing the association templating
>mechanism. It was a pointer to the role object in a template, that
>determined the "type of involvement" played by a member in the instance
>association conforming to the template. Given that an association was not
>required to have a template, it was an optional element.
>
>At the time I argued that we needed an additional element, for the role
>itself (in the instance association),  as well as the pointer to the
>"role-constraints" in the template. However, with the decision not to
>include templating in version 1.0, the need that I felt for two elements
>(one for the instance role, one for the template "role-constraining-topic",
>has gone away. We only have Role (in the conceptual model), and <roleSpec>
>(in the interchange syntax). The <roleSpec> element references a topic that
>"specifies the role" that all the topics referenced from within that
><member> element play. It therefore does indeed make sense that it be
>required. I think this is another "erratum" on the DTD... (the other being
>the <baseName> subelement of <occurrence>, which has been removed, as agreed
>in Paris).

The GI "roleSpec" (and the concept) are not bound to the concept
of templates as discussed in Dallas.

It also comes from the discussion in Swindon regarding the different
nature of [type(s)] in ISO13250 and the feeling that the term "role"
was being overloaded (in occurrence role and association role).

It seems to me that it is just as admissible to have associations
in which we know there are involvements but not their nature, as it
is to have association in which we know the roles but not who plays
them. Knowledge *is* sometimes incomplete, and topic maps should be
able to reflect this.

Therefore, I think we are justified in keeping <roleSpec> as
optional.

Steve

--
Steve Pepper, Chief Technology Officer <pepper@ontopia.net>
Convenor, ISO/IEC JTC1/SC34/WG3  Editor, XTM (XML Topic Maps)
Ontopia AS, Maridalsveien 99B, N-0461 Oslo, Norway.
http://www.ontopia.net/  phone: +47-22805465  GSM: +47-90827246


------------------------ Yahoo! Groups Sponsor ---------------------~-~>
eGroups is now Yahoo! Groups
Click here for more details
http://click.egroups.com/1/11231/0/_/337252/_/981635323/
---------------------------------------------------------------------_->

To Post a message, send it to:   xtm-wg@eGroups.com

To Unsubscribe, send a blank message to: xtm-wg-unsubscribe@eGroups.com