[sc34wg3] Next Steps?

Patrick Durusau sc34wg3@isotopicmaps.org
Thu, 21 Aug 2003 10:59:07 -0400


Well, XRI is out, at least in draft form. I won't comment on it here but 
suffice it to say if we want Published Subjects to gain some traction we 
need to move forward here so XMLVoc, and GeoLang can as well.

Suggested outline for a specification:

I. Introduction

II. List requirements and recommendations (see note below on forcing URLs)

III. Specify the details of the requirements and recommendations

IV. Conclusion

V. Boiler plate, etc.

A little more on III:

Requirement 1. Must be a URI, simply refer to the canonical document 
(RFC 2396). Not a lot we can add to that is there?

Requirement 2. must resolve..., must use URL (RFC1738, relative URL 1808)

Note can use URNs but not considered conforming PSIs for this 
specification. (I still think this is a serious and web-headed mistake. 
This requirement excludes all manner of things included by XRI, which 
despite their failure to read Pepper on web identity, at least avoids 
this mistake. People are going to want to use URNs and to exclude just 
makes PSIs less attractive, with no gain. Sorry, really think we need to 
fix this.)

Requirement 3. explicitly state the unique URI


"The unique URI for this PSI is: (insert content)."

Recommendations: 1, 2 and 3:

Use unqualified Dublin Core metadata. Already in use by the OAI (Open 
Archives Initiative, http://www.openarchives.org). No point re-inventing 
the wheel. (for unqualified Dublin Core metadata, see: 

Note that it is by definition machine and human readable, and if you use 
the same metadata for both, by definition have meet #3.

Recommendations 4 and 5, declaration that it is a PSI and its publisher.


Dublin Core metadata <creator>

PSI Declaration:

Dublin Core metadata <type>, which is defined as: "The nature or genre 
of the content of the resource."

Thus, <type>Published Subject Indicator</type>

Note that other than providing the specification, there should be NO 
tutorial or introductory material in the specification. That should be 
in a separate document that issues along with the specification.

The separate document with examples/tutorials, etc., should show how to 
construct PSIs in a vaiety of ways, HTML page, HTML fragment, XTM, OAI 
(the latter is one I am particularly interested in, see the following.)

OAI permits the construction of sets of metadata and I suspect that one 
could build a very robust set of PSIs using strictly the OAI schema. 
Note that OAI already has hosting and harvesting software available, 
which would make use of OAI compliant PSIs quite easy.

OAI has a unique identifier element for each record (their terminology) 
so one could include the unique URI there.

Comments on either the outline of the specification or perhaps even the 
OAI part of the tutorial material? (I am willing to volunteer to do at 
least a rough draft of both, perhaps more. I really think we need to 
move forward fairly quickly and am willing to devote time to that end. I 
might even be tempted into using IRC, but never having been in a chat 
room, no promises. ;-) )

Separate question: Given that XRI is allowing identifiers beyonds URLs, 
is there any support for revisiting the question of URLs? If not, that 
will lose a degree of support but I don't know how much.

Actually since it was never specified what "published" means I suppose 
closed systems like Jim Mason's could "resolve" to a local file or server.

Hope everyone is having a great day!


Patrick Durusau
Director of Research and Development
Society of Biblical Literature
Chair, V1 - Text Processing: Office and Publishing Systems Interface
Co-Editor, ISO 13250, Topic Maps -- Reference Model

Topic Maps: Human, not artificial, intelligence at work!