(unknown charset) Re: Minutes of June 22 HCLS Conference Call - Pointers to Work

> ACTION: Chimezie to send out pointers to Chimezie’s work

Per this action item, here is a summary of my research interests and 
pointers to my work - related to SW:

Pointers to Work and Areas of Interest

## XML and RDF Content Management (4Suite) ##

I've been heavily involved in the development of [1] 4Suite, an 
open-source library for XML and RDF processing.  My particular focus has 
been the RDF [2] library (which is currently being ported [3] to RDFLib) 
as well as 4Suite's content management system (the 4Suite [4] repository), 
which has an abstracted persistence layer, integrated XML services, 
ACL-based security and a [5] framework which uses RDF to represent system 
provenance of resources in the repository.

## RDFLib ##

Ever since the decision to port 4Suite's RDF library to RDFLib, I've been 
contributing to RDFLib rather frequently.  In particular, we (RDFLib 
developers) have been working towards a very expressive set of APIs for 
RDF processing and [6] persistence we hope will motivate a drive to 
formulate some language-agnostic best practices for RDF APIs.  RDFLib is 
the RDF library we are currently using for our research efforts here in 
our department.

## RDF Querying Languages: Versa and SPARQL ##

I was involved in the conception and original implementation of the [7] 
Versa querying language.  It still remains my query language of choice.
  Recently, I [8] wound up an ongoing effort to write a stand-alone SPARQL 
parser and query preprocessor which works in tandem with Ivan Herman's 
sparql-p to provide SPARQL querying capabilities to RDFLib.

## Automated extraction of RDF from XML in closed systems ##

My early involvement with development on the 4Suite repository exposed me 
to some mechanisms for automated, but controlled extraction of RDF from XML 
that [9] contrast the intent of GRDDL.  I think there is some 
merit in exploring how RDF can be extracted from closed, single-purpose 
systems where XML is the representation language of choice.  This is 
especially relevant with HL7 RIM where a single, static XSLT transform can 
be written to convert HL7 messages to an RDF equivalent.  Such a transform
would only need to change to accomodate modifications to the HL7 XML 
schema.

## N3-based Expert Systems ##

[10] FuXi is an open-source framework which is meant to evolve into a 
N3-based expert system which allows for registering rules, automated 
inference triggering and prospective querying (querying over a graph 
expanded by inferred facts using a forward-chaining mechanism).  FuXi 
currently uses Pychinko for rule production.  We've been using FuXi here 
within our research project to investigate how N3 rules can be used to 
derive variables specific to clinical trials and research projects for our 
fellows as well as how N3 can be used as a framework for [11] temporal 
reasoning.

## RDF Scalability on SQL Stores ##

I've been [12] working on a MySQL schema for N3 / FOL persistence for 
RDFLib  with an eye for [13] scalability and efficiency.  The schema is 
based off a general relational [14] model for N3 / FOL.

## RDF Graph Visualization Tools ##

I've had a long standing interest in user interfaces for RDF querying and 
have been maintaining both a 4Suite [15] library for generating diagrams 
from RDF documents as well as a user interface [16] motif for managing 
results as triples or navigable diagrams  both called Triclops.

## XML-based Form Entry ##

Before the advent of XForms I was [17] interested in common patterns with 
form-based data collection using HTML forms and XUpdate alone.

## Scutters and Mechanisms for Agents that Dereference URIs##

Redfoot's kernel follows a [18] suggested protocol for recursive dereferencing of URIs as a network of RDF graphs.

## RDF and Natural Language Processors ##

My initial interest with SW technologies was in the [19] possibility of 
using RDF as a knowledge representation framework for existing natural 
language processors.  The NLP I investigated at that time was called 
Thought Treasure.

[1]  http://4suite.org
[2]  http://www.xml.com/pub/a/2000/10/11/rdf/index.html
[3]  http://lists.fourthought.com/pipermail/4suite-dev/2005-September/002020.html
[4]  http://www-128.ibm.com/developerworks/xml/edu/x-dw-x4suite5-i.html
[5]  http://copia.ogbuji.net/blog/2006-03-28/Optimizing
[6]  http://rdflib.net/store/
[7]  http://en.wikipedia.org/wiki/Versa
[8]  http://copia.ogbuji.net/blog/2006-06-23/rdf_query_apis_for_rdflib
[9]  http://copia.ogbuji.net/blog/2005-08-20/extracting_rdf_from_controlled_vocabularies
[10] http://metacognition.info/FuXi/
[11] http://copia.ogbuji.net/blog/2006-04-24/practical_temporal_reasoning_with_N3
[12] http://rdflib.net/pipermail/dev/2006-April/000003.html
[13] http://copia.ogbuji.net/blog/2005-10-28/Addressing
[14] http://copia.ogbuji.net/files/N3RelationalModel.xml
[15] http://copia.ogbuji.net/blog/2005-05-15/Using_4RDF
[16] http://metacognition.info/Triclops/help.doc
[17] http://www.xml.com/pub/a/2002/06/12/xupdate.html
[18] http://copia.ogbuji.net/blog/2006-01-29/A_RESTful_
[19] http://www.itworld.com/AppDev/916/UIR001229ontology/

Chimezie Ogbuji
Lead Systems Analyst
Thoracic and Cardiovascular Surgery
Cleveland Clinic Foundation
9500 Euclid Avenue/ W26
Cleveland, Ohio 44195
Office: (216)444-8593
ogbujic@ccf.org