About Me
I am a sanitation engineer employed by the World Wide Web Consortium (W3C), working in Cambridge MA at MIT's fabulous Stata Center.
I am a sanitation engineer employed by the World Wide Web Consortium (W3C), working in Cambridge MA at MIT's fabulous Stata Center.
In my time at W3C, I have worked at all three hosts: MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) in the USA, the European Research Consortium for Informatics and Mathematics (ERCIM) headquartered in France and Keio University in Japan. My major work responsibilities have included libwww and the client applications, a PEP model library, perl modules for RDF databases, SQL databases, RDF visualizers, annotations support, apache, mozilla,web services and lots of systems stuff. My interests lie in machine interpretation of data have lead me down the Semantic Web road, though that path is oddly ambiguous.
I am currently a team contact for the W3C Semantic Web Health Care and Life Sciences Interest Group (HCLSIG) and working part time on SPASQL.
SPQL with remote execution: SPDL presentation to the TAG
SPDL with Choreography: SPARQL in WSDL
(SPDL 1st technology presentation to EC: W3C WS2 Web Semantic Web Services Demonstrator)
(SPDL 2nd technology presentation to EC: WS2 Project Review - Annotated Service Invocation)
introduction at MySQL User's Conference: MySQL in the Semantic Web a SPARQL parser in mysqld
-intro, +costs and expressivity comparison: Adding SPARQL Support to MySQL
comparing XML, Relational and RDF: SPARQL Support in MySQL
Clinical Observations Interoperability Use Cases Lightening Talk
W3C Work on Semantic Web in Health Care and Life Sciences
SPARQL Overview for the BioRDF Task Force
And here's stuff I need to get around to:
Here's a bit of stuff I remember coming up on radar...
Given a BNF grammar, yacker writes parsers in C, C++, perl and python. I use it for tracking evolution in languages that I've implemented. Also nice to settle LL(1)/LALR(1) issues in real time.
WSDL (and its embedded XML Schema) have annotation mechansims. Adding WSDL and XML Schema annotations that associate RDF GraphTerms with allows a mapping between SPARQL triples and a locations in a SOAP message. See the write-up and the write-up of the subsidiary SPAT language.
Having played with mapping SPARQL (and Algae2) queries to SQL, I decided to move that logic into mysqld, allowing arbitrary clients of the conventional MySQL protocol (perl, dbi, ODBC, JDBC) use of it with no modification. See the code, write-up and some earlier notes.
RDAL - RDF Annotations Language — another schema annotations, this time for Relax NG. RDAL associates semantic actions (stored in RNG Annotations) with call-outs to a generic API; implemented as a validating XSLT with callouts to another XSLT library.
As web protocols are are more and more stacked, the conventional two-tier media-type mechanism becomes insufficient to hint to applications that they may or may not have some utility in looking at the entity. The list of IANA media types is available in a general IANA registry.
Machine understanding implies machines are told what consitutes sufficient processing. I actually don't even remember writing this proposal.
Briefly describe and contrast a set of XML protocols. RPC and messaging, mostly.
The modularity of XML specifications deliniates a set of functionality boundries. Different documents will require differnet modules. This document is a plea for a taxonmy of modules names, ordering and optional processing directives.
provides an algorithmic transformation of WSDL document to an RDF model.
example of use of the RDF WSDL Mapping.
NOT DONE - barely even started - don't look here.
survey of web service research from the RDF perspective.
Adds RDF support for Creative Commons to a sample document.
Resource nodes to describe documents, analogous to i-nodes, implemented over http extensions.
Topic Maps describe associations of objects (identified by URI). RDF describes relationships between objects. This document describes a few algorithmic mappings between the two.
The above perl RDF parser uses an XML-centric grammer with scripted actions associated with each production. This workes to parser many forms of XML. I hope this work leads language requirements for a general schema annotation language.
This provides an alternative to the an XML-centric grammer approach used by the perllib parser.
A way to present a user with policies from a published P3P policy.
Many communities decide to use RDF or Topic Maps or UML without understanding the overlaps or constrasts in expressivity of those systems. This paper attempts to reduce FUD (fear, uncertainty and doubg) and promote some common understanding.
While rdf parsers allow us to access to RDF stored in XML or N3 documents, persistent triple stores provide more convenient access.
For many applications, using a triple store is less efficient than using a database optimized for the particular data. This RDF SQL Mapping paper describes using such a database as an RDF database.
This paper explores RDF access to an OrderTracking database and demonstrates practical queries to the database.
We can get much more efficient use of databases if we normalize our data into conventional relations. This paper demonstrates that this normalizations need not sacrifice flexibility of expression.
Many RDF database implementations support an attribution or context. There is no standard way to do this. This paper explores some alternatives and discusses their respective pros and cons.
Similar to the XML protocol matrix.
RDF SQL mapping paper also discusses this.
juncture between semantic web and the grid forum.
Implementations for Apache and Mozilla.
Protocol Extension Protocol
Abstrace API for access to RDF datastores. Has concrete bindings in SOAP and HTTP GET/PUT/POST/DELETE.
10 Jun 0800 UN173 BOSc→SFO 20 Jun 0810 AC563 SFO →YVR 25 Jun 1100 UA768 YVR →ORD (26 Jun) 26 Jun 0630 UA525 ORD →BOS
31 Jul 1730 AF337 BOSe→CDGe 0645 (20 Oct) 10 Aug 1315 AF332 CDGf→BOSe 1555
13-14 July
4-9 Oct