[tm-pubsubj-comment] Re: paradigmatic PSIs

[Bernard Vatant <bernard.vatant@mondeca.com>]

*Steven R. Newcomb
> >| Each and every PSI, in the standard and at each site, will list all
> >| of its addresses -- its address at each host's domain name -- as its
> >| canonical, utterly valid addresses.  The use of any one of them as a
> >| subject indicator reference unequivocally establishes the subject as
> >| the same one that appears in the standard.

* Lars Marius Garshol
> >A little note: the best is probably if there could be one canonical
> >URI for each subject. That would make them a lot more practical in
> >actual use, especially in software, as the software could then check
> >for one URI rather than X different ones.

*Eric Miller
> This I understand a bit more.  Recognizing (yes?) that many different
> communities may indeed be providing such URI's.  e.g. ISO providing a
> unique identifier for representing (country x), Getty providing a unique
> identifier representing (country x), the USPO's providing a unique
> identifier representing the (contry x)...  All of them "valid", yes?

What "valid" means here is a real issue. For example it has been well stated in GeoLang TC
that if the TC makes available (publishes) different sources of countries codes as PSIs,
it should not make any assertions about the equivalence or not between those sources,
other that the ones declared by the sources themselves. And publishing a PSI is not an
assertion of "validity", just an assertion of availability and stability. It's up to the
users to refer to the PSIs they trust and consider as valid, and/or to the source
publishers themselves to assert equivalence.

*Steve
> >The isotopicmaps.org domain seems ideal for this purpose.

*Eric
> hmm... this is something i don't understand.  It seems more cost effective
> (from both the societal trust and maintenance standpoint) to help
> organizations establish persistent policies and procedures for making
> available these PSI's on their respective sites. no?

You're right generally speaking, for any specific legacy of subjects (e.g. Countries or LC
Subject Headings ...). But I think Steve was there speaking about very generic
(paradigmatic) PSIs, e.g. for "topic" "association" "occurrence" ... the ones that are
used by the standard itself. It makes sense that those be defined in some authoritative
and unique space linked to the standard. Although "heretic" PSI for those same concepts
could be available elsewhere :))

*Eric
> I was under the impression that's what the OASIS PSI working groups were doing?
> To identify existing controlled vocabularies associated with a variety of descriptive
> facets (In this case starting with ISO re place, language, etc.), define
> such persistent, descriptive and evolution policies and in essence help
> these groups make this information available to various supporting
> applications and services.

> If my assessment of what the OASIS PSI working groups are attempting to
> accomplish is incorrect please let me know.

Your understanding seems perfectly correct to me.

> Suggesting these PSI's are located on a different, centralized site,
> maintained by a different maintenance organization seems unnecessary and
> counter to what I think (hope) we're collectively after.

Generally speaking, yes. Concerning the specific PSIs for concepts used by standards, it
makes sense that they come along with the standard itself.

Bernard

PS: I forward that thread to tm-pubsubj-comment@lists.oasis-open.org,
 which seems to me the right forum.
For those who are not yet subscribed, go to http://lists.oasis-open.org/archives/

Bernard Vatant
Chair - OASIS TM PubSubj Technical Committee
www.oasis-open.org/committees/tm-pubsubj/