On 31-Jan-15 00:06, Stian Soiland-Reyes wrote:
Great job with the XSDs!
Ta muchly. There is a lot that I would want to do to improve it, for example making the port -> data an actual Java Map and making the lists into Lists.
I think we should keep it at Apache Taverna under incubator-taverna-language - I know Baclava will be legacy as it's only made in Taverna 2, but Ernesto has shown interest in using it to do a tool for databundle conversion as they have many Baclava files from Taverna 2. That is exactly the kind of engagement we should be encouraging under Apache.
It's in there now. [snip]
We just need to ensure that tool can't WRITE baclava files.. ;-)
There is a writer in there. However, although Baclava files generated by Taverna 2 conform to the schemas, Baclava generated from the schemas will not always be readable by Taverna 2 as that makes too many weird assumptions.
Should we also put the XSD at http://taverna.incubator.apache.org/ns/ ? I know baclava is under a dubious namespace like http://org.embl.ebi.escience/baclava/0.1alpha - but http://ns.taverna.org.uk/ is where we have published all the other schemas.
I do not mind. Though it would be a good idea to wait until Ernesto checks if his files can be parsed according to the schemas.
(.. although .escience is not a top-level domain, .science is now! http://nic.science/ )
Alan
