January 29, 2003

Bugzilla account for SWAD-Europe / RubyRDF

I've just got a Bugzilla account for SWAD-Europe work, as part of W3C's Bugzilla installation. To find out more about the facilities it offers, I am experimenting with a 'component' for RubyRDF, my Ruby software tools. If I can fix enough of the bugs to make it useful, I'll probably use RubyRDF as a basis for a number of demos through the SWAD-Europe project.

Bugzilla is interesting because it is used on quite a few projects that work in public, and so various people are working on add-ons (RSS feeds, RDF views etc., email interfaces...). I'll know more about it once I've used it a bit more...

Posted by danbri at 11:52 AM | Comments (0)

January 24, 2003

Summary of state of current RDF/XML parsers

I wrote this Summary of state of current RDF/XML parsers to the RDF Core WG list yesterday. It seemed appropriate to start collecting pointers to implementations and it looks encouraging.

Posted by dbeckett2 at 09:37 AM | Comments (0)

January 23, 2003

WP10 D10.2 - Mapping data from RDBMS

SWAD-Europe: Mapping data from RDBMS - report by Jan Grant and myself from D10.2 now done.

Posted by dbeckett2 at 04:43 PM | Comments (0)

January 21, 2003

Sesame choice of RDF Parser

I was sure Sesame had a parser of its own, not ARP and I just found it in the Configuring Sesame section of the (GFDL) User Guide for Sesame:

Sesame supports two RDF parsers: HP's ARP parser, and Sesame's own parser (which is so new that it doesn't even have a name yet). The ARP parser currently is the most standards compliant in that it checks almost everyting for compliance with the RDF specification. Sesame's RDF parser is by far the faster parser, but it doesn't check everything yet (e.g. it doesn't check whether URIs contain characters that they aren't allowed to contain).

The default parser is still ARP, as we would like to have some more feedback on the compliance of our own parser to the specs. You can switch parsers by changing the value of the rdfParser parameter in the web.xml file. Supported values are ARP and Sesame.

Posted by dbeckett2 at 10:11 AM | Comments (0)

January 20, 2003

MEG Registry Updates

Today I've been working on tidying the server side parts of the MEG Registry for educational metadata schemas. It is an RDF application both in the data model, client and server. The client is written in Java by Damian Steer and the server in Perl/C using Redland, by myself. The system is entirely based on the RDF graph model as well as using the rdf/xml format for interchange between the c-s, with a REST model for operations. See more at the MEG registry project web site with pointers to the sources (GPL or open source).

Posted by dbeckett2 at 06:01 PM | Comments (0)

SWAD-Europe Wiki

I have set up a SWAD-Europe Wiki installation, as a place for quick note-taking, FAQ drafting and link sharing. Like many Wikis (and mailing lists, though that's a bit different), it is currently configured as a publically writable space. It may prove useful as a tool to support discussions in the RDF Interest Group and associated spin-off task forces (eg. RDF calendaring). Time will tell. See my note to the public-esw list for more details on how the Wiki might be used.

Posted by danbri at 03:47 PM | Comments (0)

W3C Web Services Choreography Working Group created

On 14 January 2003, W3C created a W3C Web Services Choreography Working Group to address the description of the composition and relationships of Web services.

The charter recommends a mapping to the Semantic Web.

Posted by hugo at 11:54 AM | Comments (1)

Experiments with mapping XML to RDF: testing Edinburgh Schema Mapping Framework

I have been experimenting with some tools that map 'colloquial' XML into an RDF-like representation. See the online readme and nearby files for my work in progress. I've added an RDF (NTriple) output stylesheet, and am now looking for more scenarios / test cases that might help us evaluate such an approach. The original work was produced by Henry Thompson and Ari Krupnikov at the University of Edinburgh. In SWAD-Europe we are exploring whether this is a viable approach to deploying RDF in XML-centric environments.
Posted by danbri at 11:10 AM | Comments (0)

January 17, 2003

RDF Core agrees to publish last call Working Drafts

RDF Core agrees publishes last call Working Drafts
-- Minutes RDF Core telcon 2003-01-17 Update 2003-01-21 correction; the WG has agreed to publish them, out in a few days.

Posted by dbeckett2 at 07:07 PM | Comments (1)

MT inserting hyperlinks in IE

[Now playing with formatting buttons]

View source shows this script in the edit entry page downloaded by Movable Type:

[[
function insertLink () {
if (!document.selection) return;
var str = document.selection.createRange().text;
if (!str) return;
var my_link = prompt('Enter URL:', 'http://');
if (my_link != null)
document.selection.createRange().text = '' + str + '';
}

function mtShortCuts () {
if (event.ctrlKey != true) return;
if (event.keyCode == 1) insertLink();
if (event.keyCode == 2) formatStr('b');
if (event.keyCode == 9) formatStr('i');
if (event.keyCode == 21) formatStr('u');
}
]]

Posted by gklyne at 12:48 PM | Comments (0)

January 16, 2003

A little RDF parser testing

Not a lot to show today but I had a quick go at the new Java ARP2 alpha release parser and compared it to my Raptor parser, written in C and probably does a lot less checking. ARP2 is faster than ARP1 - I tried it out on a significant bit of rdf/xml with about 270K triples but Raptor remains up to 40x faster than ARP2. The figures:

Parser TimeRelative to Raptor
ARP1 1.0.5216s54x
ARP2 alpha
(JDK 1.4.1-01)
135s, 157s, 165s, 164s33-41x
Raptor (0.9.8 CVS)4s, 4s1x
Posted by dbeckett2 at 08:22 PM | Comments (0)

Graham Klyne - my ESW activity so far

I've been doing some work as part of Rutherford Appleton Laboratory's contribution, looking toward exploring the use of Semantic Web technologies to model and manipulate trust in open systems, looking initially at authorization and access control issues.

So far, I've prepared:


  • a security protocol surver document [1]

  • a trust scenarios document [2]

  • a simple experiment using non-specific RDF tools (cwm and my own report generation software) to generate network system configuration files from an access policy description in RDF. [3][4]

We're still talking about where it might go from here, but I have some ideas...

[1] http://www.ninebynine.org/SWAD-E/Security-formats.html

[2] http://www.ninebynine.org/SWAD-E/Trust-scenarios.html

[3] http://www.ninebynine.org/SWAD-E/Scenario-HomeNetwork/HomeNetworkConfig.html

[4] http://www.ninebynine.org/SWAD-E/Scenario-HomeNetwork/HomeNetworkAccessConfig.html

Posted by gklyne at 07:01 PM | Comments (0)

January 15, 2003

RDF Core work for today - RDF/XML Syntax last call candidate working draft

So the RDF Core work for today was updating the RDF/XML Syntax last call candidate working draft. Here's the message I sent to RDF-Core and the (still unapproved) draft

Posted by dbeckett2 at 09:25 PM | Comments (0)

Mapping data from RDBMS Report

This is mostly a test that this is working for me. But today I've partially been working on SWAD-Europe: Mapping data from RDBMS (DRAFT) merging in the bits Jan wrote and fleshing out the rest.
Posted by dbeckett2 at 09:04 PM | Comments (0) | TrackBack

Latest RDF calendar irc meeting

The latest follow-up to the SWAD-E calendar workshop was at 6pm GMT 2002-01-15 (agenda, logs, participants). Developers from Mozilla calendar, OpenCal and apple iCal, as well as W3C and lots of other interested parties came along. We talked about xcal, tests, interoperability, openCAP and CAP in general. The RDF calendar workspace is used for RDF tests and various links, and previous chats and logs. The next chat is Wednesday 22nd January 2003 at 5pm GMT.

Posted by lmiller2 at 08:15 PM | Comments (0)

SWAD-Europe weblog installed.

I have set up a weblog for SWAD-Europe using the Movable Type system widely used elsewhere. The idea is that SWAD-E project partners and collaborators might use this to share links and commentary, particulary across our inter-related workpackages.

Next job is to set up more accounts for others, and customise things a little more.

Posted by danbri at 07:14 PM | Comments (0)