Hi
Glad to hear we are not breaking too many of the
Groovy rules with our extensions. I agree the object model is the core to
providing a useful scripting base the difficulty is defining it to
give best support to the user community or for that matter defining the user
community.
As to Java 5 support our version of Groovy does
have some support, again requirements driven so incomplete but the intention is
to complete after release 1. I did see a Jython project, I think from the LSE,
working on 5 support but have not heard of any further work.
I have requested permission to release sections of
the work streams into the public domain and expect this to be granted, in any
event, work done on uDig Groovy etc. will be returned to their projects. If you
would like a preview of our work you may like to look at the research base:
For the data stream manager
and last but not least
The PIPES stream processing project at http://dbs.mathematik.uni-marburg.de/Home/
this is a pure java implementation of
a publish subscribe architecture stream processing engine which we implemented
as a plug-in under the uDig catalogue for use in demonstrations, proof of
concepts, where an integration bus would be inappropriate, the associated XXL
library was used together with JTS in the spatialization of the DSMS,
integration bus and Derby. We had a similar approach to D Blasby's work in
the spatialization of the DSMS and integration bus and follow the SFS and OGC
standards in all our spatial work, and would agree with your thinking on this.
All of the above projects have a GUI approach to query generation and scripting,
some of the demos will give an insight into our scientific workflows and traffic
workstreams.
For the Integration Bus:
SEDA http://www.eecs.harvard.edu/~mdw/proj/seda/ is
the basis of the architecture and the architectural patterns use can be found on
http://www.eaipatterns.com/ on
top of these we have implemented management systems for e-gif (the UK
governments interchange format) and data protection, privacy, freedom of
information acts (important in location information) and evidential (chains of
custody etc). for use in enforcement, the main integration mechanism is the
picocontainer, the systems is rules driven using Drools or iLog.
A more general approach is available at http://docs.codehaus.org/display/MULE/Home which
follows the SEDA architecture in a more relaxed way and has many more
integration mechanisms including Groovy which may be of interest.
Regards
Tony Kennedy
|