W3C home > Mailing lists > Public > public-esw-thes@w3.org > November 2004

Global concept identification and reference: Is PSI issuing really flawed due to copyright restrictions?

From: Alexander Sigel <sigel@wim.uni-koeln.de>
Date: Wed, 10 Nov 2004 20:07:50 +0100
Message-ID: <F18091DBF62C0849A69BC306A7BE9FBB0CBA4B@mailserv.wim.uni-koeln.de>
To: <public-esw-thes@w3.org>

Yes, there are copyright issues, but the argument that anyone can issue
PSIs is not fundamentally flawed because of that, IMHO.

Let's take as an example name authority data (not a thesaurus, but a
referencing instrument for concepts), e.g. from the German national
library (DDB PND) and from LoC. 

You are right that a knowledge worker may not be allowed to publish a
certain _full_ entry.
But in order to identify and later reference that PSI, we need only
_enough identifying_ information (as attributes), not the full entry.
This may also be a legal abridged paraphrase, or just the text
"Please buy the content from the publisher (link, cost), open their
ebook on p. 6290 and have a look at the third entry on that page. This
is what I, the issuer of this PSI did. Now let's talk about the person
referenced there". 
Nobody can prohibit me issuing this PSI and add information from other
public sources.

E.g. I can uniquely identify the Bezold, Friedrich von I am talking
about if I put into my published description the note that it is the
same person that is discussed as a member of the academy with all his
publications in the open publication:
http://www.bbaw.de/bibliothek/digital/struktur/autoren/bezold/literatur.
pdf
(Indeed http://www.bbaw.de/bibliothek/digital/struktur/autoren/bezold/
could be made a PSI)

I can declare that the person identified by the DRA copy of LoC NA
APK-5424
http://lcmarc.dra.com/lcauth/APK-5424 (me, in this case), is the same
person as referenced by LoC NA n 2001009105, without breaking copyright.
(I hope ...)

Thus my PSI would maybe look like this:
-----------------
Entry
Sigel, Alexander

URI
http://purl.org/NET/psi/#Sigel_Alexander

Human-interpretable Published Subject Indicator
LoC NA: n 2001009105
copy @DRA: APK-5424 

Explicit statement that this unique URI is to be used as the Published
Subject Identifier for this Published Subject Indicator  
The unique URI http://purl.org/NET/psi/#Sigel_Alexander is to be used as
the Published Subject Identifier for the Published Subject Indicator for
"Sigel, Alexander"  
-----------------

I agree that some publishers might even prohibit using their number for
referencing purposes. But then we are not discussing discouraging
"unofficial" PSIs because of their proliferation, thus introducing
disorder into the system, but questions of open referencing systems in
which referencing systems for open content interlinking are to be
recreated.

What I meant was that in order to use the PSI for Item (in FRBR) I just
have to define a PSI and do not need some extra permission.

But I see the problem of KOSs publishers. DDB PND is used in the
Kalliope-Portal, but can - until now -not  be used as a referencing
system, because DDB has not allowed this.

The workaround is to publish some proxy subject, e.g.
http://141.20.126.79/~voj/kalliope-gateway.php?autor=Bezold%2C+Friedrich
+von
lists the PND number 118662880 if the text string "Bezold, Friedrich
von" is unique.

Currently DDB checks if this is legal ...

<topic id="12345"> 
  <subjectIdentity> 
    <subjectIndicatorRef
xlink:href="http://purl.org/NET/psi/#PND_118662880"/>
  </subjectIdentity>
  <!-- names and occurrences -->
 </topic>

In the end, we will be able to reference with decentrally provided PSIs.

[Andrew Houghton]
> I'm assuming it's not forbidden to allow KOS publishers to use PSI's
in a private service
> oriented context.  
> Some KOS publishers will not be willing to use PSI's due to
intellectual property concerns, but > might be willing to provide opaque
URI's to concepts in their KOS.
Yes.

Some KOS publishers might be willing to provide reference numbers (e.g.
WBI - World Bibliographic Index person numbers) and a name, but not much
more (e.g. dates, profession). Commercial KOS publishers might want to
add some free content (like free chapters of books) for
marketing/promotional purposes and to drive traffic to their site. The
full content will cost money.

The interesting question is: How much information in attributes is
necessary to identify the concept (here: person) and how can this work
with commercial publishers in conflict between their need for click
streams and revenues and their legal interest in protecting their
intellectual property. But this is not a technical issue and does not
flaw the concept of decentrally issued PSIs in P2P fashion. 

Regards
Alex
-----
Alexander Sigel, M.A., Researcher in Semantic Knowledge Networking
sigel@wim.uni-koeln.de, +49 221 470-5322, http://kpeer.wim.uni-koeln.de/
U Cologne, Dept. of Information Systems & Information Management
office: Pohligstr. 1, Room 406, 50969 Cologne, GERMANY

-----Original Message-----
From: public-esw-thes-request@w3.org
[mailto:public-esw-thes-request@w3.org] On Behalf Of Houghton,Andrew
Sent: Wednesday, November 10, 2004 7:30 PM
To: public-esw-thes@w3.org
Subject: RE: Global concept identification and reference: Published
Subjects and decentrally provided identification points for notions
Received on Wednesday, 10 November 2004 19:13:57 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:38:52 GMT