Yes, I agree it is not consistent.
I seem to recall that EDIFACT uses a comma separator in its definitions of elements represented by a list of 'codes' (e.g., "Country, Coded"). I like that comma! I believe they also define 'Country' (e.g., Spain). Note that they do not define that as 'Country, Identifier'. ) Note also that 'Country, Coded' and 'Country' each identify the country.
I've no problem with making a distinction among collections of values represented in our standards by code list values. In particular, I accept the 'instance' concept of an identifier. Because identifiers represent instances, and instances tend to be volatile, many (perhaps most) identifiers are not (or can not practically be) coded. But a few, like country and currency codes, are coded.
Clearly, some identifiers can and are represented by codes, and conversely, some code lists are identifiers. Clearly also, not all code lists are identifiers.
From: Peter Denno [mailto:peter.denno@xxxxxxxx]
Sent: Wednesday, October 27, 2004 8:56 PM
Subject: Re: [ontolog-forum] CCTONT work progress
> Robert Miller said:
> Ah, but we DO use codes to identify objects in the real world.
> The realm of code sets spans adjectives, nouns, and I suppose verbs and
Then you agree with me that the statement in section 6.1.1 contradicts the
statement in section 8.1?
On Wednesday 27 October 2004 18:01, Monica J. Martin wrote:
> >>I've also attached a paper from Mike Adcock, now deceased, who wrote
> >>about the quandry about code and identifier. A lot has been done and
> >>debated since this was produced in June 2002. As a final note, if it is
> >>a real world object, the representation term is used instead (if memory
> >>serves me correctly) not the Code. Type. Thanks.
> >Wallace: I had a look at the attached document. The best that can be said
> > of it is that the examples may help us infer the distinction that the
> > authors failed to describe. From the examples it seems that Codes can be
> > aliases for attribute values whereas Identifiers are attribute values
> > which can be used to denote object instances (or occurences, but NONE of
> > the identified things in the examples are occurences).
> >Of course, Codes are attribute values as well. My guess is that they wish
> > to think of these aliases as identifiers for the Identifiers, although
> > one could think of them as Identifiers where some naming context is
> > needed to uniquely identify objects using them. The only "real"
> > difference between this and what they call Identifier is that the context
> > is explicitly defined and the values and their mappings should be in a
> > repository somewhere.
> mm1: It may sound simple and one may reason that the types are the same
> or directly similar....but you've not seen the debate over the last
> three years that would indicate otherwise. It is a religion of sorts
> that they are different even though often misused and assumed the same.
Then the specification editor needs to record the definitions provided by the
victors. Otherwise the rest of us can't get our jobs done, and nobody will be
able to implement this correctly.
National Institute of Standards and Technology,
Manufacturing System Integration Division,
100 Bureau Drive, Mail Stop 8260 Tel: +1 301-975-3595
Gaithersburg, MD, USA 20899-8260 FAX: +1 301-975-4694
Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/
To Post: mailto:ontolog-forum@xxxxxxxxxxxxxxxx