[tm-pubsubj-comment] Re: [topicmapmail] ANN: UNSPSC topic map available

[Kal Ahmed]

On Tuesday 15 October 2002 02:51, Mary Nishikawa wrote:
> At 10:01 PM 10/14/2002 +0100, Kal Ahmed wrote:
> >On Monday 14 October 2002 21:15, Lars Marius Garshol wrote:
> > > * Kal Ahmed
> > >
> > > | For no particular reason other than "because it is there", I've
> > > | converted the Universal Standard Products and Services
> > > | Classification into XTM form and uploaded the results to [1].
> > >
> > > Oh. You too? :-)
> >
> >Great minds...
>
> Kal, glad you see the importance of this too :-)
>
> > > | For those that don't know the UNSPSC is a standard hierarchical
> > > | classification of commercial products and services and, according=
 to
> > > | the publishers "is used in sell side and buy side catalogs and as=
 a
> > > | standardized account code in analyzing expenditure (Spend
> > > | Analysis)."  [2].
> > >
> > > Mary Nishikawa uses it in her paper from Extreme 2002:
> > >   <URL:
> > > http://www.idealliance.org/papers/extreme02/html/2002/Nishikawa01/E=
ML20
> > >02Ni shikawa01-toc.html >
> >
> >Great minds...
>
> Many "great" minds together =3D progress?  ;-)
>
> This talk developed  from countless discussions with Lars Marius and a =
year
> of discussions with the OASIS TM Published subjects members (see the
> credits in my paper).
> Kal, we welcome you :-)
>

Thanks! I feel embarrassed that I didn't find that paper. My excuse is th=
at I=20
didn't get to go to Montreal this year. Not much of an excuse, I know :-)

> Actually Jack Park came to my talk, and he asked that I write a chapter=
 for
> the next Advances in Topic Maps book -    (this is what he meant by "po=
rtal
> integration" with topic maps -- Jack's description of what I will be
> writing on). You will be writing a chapter too, right? Great minds toge=
ther
> :-)
>
> I am hoping that my Late Breaking News talk for XML 2002  is accepted
> because I am expanding on the ideas presented in my Extreme Markup pape=
r.
> The UNSPSC was only the beginning and the point was to show that a stan=
dard
> "upper level" and generalized scheme can be used a central location for
> others to "pick up" binding points to use in their own maps of internal=
 use
> subjects; I want to have fixed published subjects for the UNSPSC and
> publish them in a fixed place. It would be a good idea for the owners o=
f
> the classification, the ECCMA,  to do this, as I said in my talk, but w=
e
> may need to first publish it in http://psi.oasis-open.org
>

Mary, I think that we are both thinking along exactly the same lines. I w=
ould=20
be interested to hear see your presentation as I'm not going to get to ma=
ke=20
it to XML 2002 either (the Atlantic really sucks sometimes ;-). Do you ha=
ve=20
other vocabularies in mind ?

One thing that I think is interesting is that you and I took different=20
approaches to topic naming - I left out the numbers from the name. You ha=
ve=20
added sort names with the numbers to allow a different sort order of the=20
topics. Another difference is that my topic map contains version informat=
ion=20
(which may or may not be of use to topic map users). Yet with a common PS=
I we=20
would be able to merge both views of the same data. In fact, it should be=
=20
possible to create a mediating topic map with the sole purpose of causing=
 the=20
two different views of the same data to be merged. Thats something I thin=
k=20
I'll have a go at.

Some organisation should be involved in the bootstrap process. I think OA=
SIS=20
is ideally placed to do this if neither ECCMA nor the UNDP nor UN-SPSC wo=
uld=20
go for it. But thats three big ifs in a row.=20

> The UNSPSC is used by web services, so I am sure you see the importance=
 of
> this.
>

Absolutely - I really can't imagine directory services like UDDI working=20
without topic maps to be honest.

> The Library of Congress could very well do the same thing.
>
> > > | Lacking any form of suitable subject indicator I have opted to
> > > | create a private urn namespace x-unspsc. Perhaps someone on this
> > > | list could suggest a better way to provide useful subject indicat=
ors
> > > | for this map ?
> > >
> > > What about the old tired-and-true method of http URIs that point to
> > > nowhere? I feel that is better than making new URI schemes
> >
> >I've done that an awful lot in the past and I feel guilty about it. My
> >feeling
> >(now) is that at least one knows where one stands with a made-up URN
> >namespace. There is no way of resolving it, so there is no possibility
> > that by resolving it a subject descriptor would appear. In other word=
s it
> > serves as a simple identifier (unfortunately not necessarily a unique
> > identifier, but I suppose I could use urn:x-techquila.com:unspsc:  on=
 the
> > basis that fewer people are going to use the x-techquila.com namespac=
e in
> > their made-up URNs).
>
> Kal, see my paper and the work of the committee.
>

Thanks for the pointers - I'll read through this stuff in more detail=20
tonight/tomorrow and hopefully have something more constructive to add a=20
little later.

Cheers,

Kal