If any of you are using NPCL to write your topic map ontologies, you should know that we have just released a new version of the NPCL Schema Editor. Includes MSBuild support, compiling multiple NPCL files into a single file, auto-generated documentation and much much more. Fun for all the family and competitively priced at $0. For more info see http://www.networkedplanet.com/ontopic/2009/03/npcl_schema_editor_for_visual_1.html
It is very interesting to see that Amazon have now made available over 1TB of public data. Its great that all of this data is now available in one place, ready shredded into queryable structures that allows developers to get to grips with it and start to do something really interesting. But wait a minute, if I want the DBPedia dump I have to go here…, if I want the Wikipedia english articles I need to go over there. If I want the US census data from 2000 its this place, if its the census data from 1990, its somewhere else. Oh and don’t even get me started on having to choose between Windows and Linux. What these data sets are are essentially separate database snapshots that you can load into your own EC2 instance in the Amazon cloud and then start processing.
…and thats kind of disappointing. Having lots of open data is a great start, but it is only a start. And here is the challenge – there are no consistent semantics acrosss these data sets, there is a great deal of wet-ware time that needs to be invested in working out the linkages between them and in getting hold of some consistent notions of identity that could assist in merging. The easy way out is to pick and choose and make a “mash-up”, but there is nothing reusable in a mash-up, and a million mash-ups do not make a viable platform for building the really cool apps of the next decade on. Topic maps on the other hand has a model for reflecting a consistent notion of identity, for reconciling different identity notions and different entity schemas.
There’s the challenge – can we integrate all of this data using topic maps ? Can we make use of the tools provided by the Amazon platform to build something even more cool – a cloud based index of the entities and relations in these data sets ? Because I believe that when we can do that we will really have 1TB (or given the expansion of the topic maps model probably 1PB 
of useful knowledge, rather than 1TB of bits and bytes that you can hack with a mash-up.
There’s the glove. Who’s going to pick it up ?
Today sees another release of TM4J. This release completes the support for the recently finalised TMAPI 1.0 APIs as well as including a number of bug fixes and documentation improvements.
Downloads are available from the SourceForge project page.
A while ago I started work on some XSLT transforms for creating a topic map of a W3C XML Schema. Its definitely not a straightforward task, just handling nested includes and class hierarchies is hard from XSLT.
Anyway, I haven’t had a chance to work on these stylesheets for a while, so I’m posting them under the Reciprocal Public License in the hope that some one else (LMG ? might pick them up and contribute more to them. There are two stylesheets:
schema2xtm.xsl works on the schema file directly. This is probably the least-finished of the two.
psv2xtm.xsl works on the PSV output generated by the XSV schema processor. To use this stylesheet, you must first validate your schema with XSV, requesting the PSV output as XML, then run the stylesheet against that output.
Today sees the release of the latest version of TM4J, the open-source topic map engine. This release is primarily a bug-fix release, but also enhances the tolog query implementation to enable dynamic association predicates to use a variable for the association type, thus allowing queries such as: $ASSOC(foo, bar) to find the type of all associations between topics foo and bar.
Source and binary distributions can be downloaded from the TM4J SourceForge page.
Here’s an idea that just struck me on the train on the way home.
Phil Gyford runs an excellent blog of the diary of Samuel Pepys. He blogs entries from the diaries one at a time 343 years to the day from the date of the diary entry (so at the time of writing we are in July 1661).
My idea is to attempt to topic map the entries in parallel with Phil’s blogging.
The diary has a lot of interesting features:
Topic maps excel at enabling references between these features to be made and connections not at first apparent by a serial reading of the entries might be revealed. As far as I know there have not been many attempts to topic map a series of events like this so it would be an interesting experiment. I wrote some thoughts on topic mapping events a few weeks ago, so it might also be an interesting way to put that to the test.
So here’s the plan:
I’m going to start blogging topic-mapped entries from Pepy’s diary. In each blog entry I’ll reference the entry, write some explanation about why I’m modelling it in my topic map in the way that I am and then include the topic map for that entry. Initially I plan to use LTM notation as it is more compact and readable than XTM.
It would be nice to make this an interactive process, so I welcome comments and proposals of alternate representations! Also, as I am no Pepys scholar, all I can do is attempt to represent what I read in the entry – I would welcome pointers in the right direction from anyone with more knowledge of this subject.
Seeing as at present there is a hiatus in the diary (there are no entries from 8th to 13th July), I’m going to attempt a bit of catch up from 1st July 1661. Watch this space…
Some thoughts on representation of historical events using topic maps.
In a large part, the thoughts here are inspired by Historial Event Markup Language.
TMTab is a plugin for the Protege ontology editor that enables the creation of topic maps (in XTM syntax).
TMTab version 0.5.0 is a straight-forward port of TMTab to the latest version (2.1 beta) of Protege.
TMTab is released as charity-ware. If you find it useful please consider making a donation to Shelter – the Help / About Plugins… menu option has more details (select TMTab from the list of plugins you have installed!).
For more information and to download the software, click here.
A new version of TM4J is now available for download from the TM4J SourceForge page. This version supports TMAPI 1.0 beta and includes a few additional bug fixes and minor API changes. Full details here.
The TMAPI team today released the first TMAPI 1.0 beta.
TMAPI is a proposed common API for topic map processing software. The API definitions have been developed in an open process and released as Public Domain software.
The download is on the TMAPI SourceForge project page.