Yes I aggree with that Ricardo - "gap" analysis was not the best choice of words.
I didnt really have problems using RDF with the Java LSID stack as the download came with examples of how to use RDF. The biggest problem here is trying to work out how to correctly create an RDF document from the index fungorum database I am connecting to (so really more of a training issue).
The next problem was when trying to create RDF documents within Microsoft .NET development projects (this was also during setup of to the IndexFungorum LSID resolver). There is limited code available for facilitating the creation of RDF documents for the Microsoft platform (not a huge issue, as I could just use java code to create them, but this has platform dependent issues later on).
My thoughts really were to make sure that code and tools are available to help everyone create RDF documents from their datasets, ensuring the adoption of the RDF techonology is as easy as possible, and has no obvious barriers. As Ben pointed out there are quite a lot of tools, and Java code, for working with RDF, but I think it would be useful if everyone thought about how they would get their data from the database to an RDF document to make sure there are no missing components to aid this process.
Kevin,
That is an excellent point, thanks for bringing it up!!
Yes, I believe we should include RDF into the analysis. We just need
to be careful about the wording, I guess.
We clearly don't want to perform a gap analysis of all the existing
RDF software. That would be too big of a task for the GUID group.
In my opinion, we should evaluate:
1) How well the current LSID server software supports the generation of
metadata responses in RDF; and
2) What capabilities the LSID client provides for ingestion of RDF.
Besides that, we can also try to compile a list of RDF toolkits and
try to get evaluations of them, without performing a complete gap
analysis. Some people was asking for links about RDF tools on the
TDWG-TAG mailing list and on the GUID list as well. Take a look:
http://lists.tdwg.org/pipermail/tdwg-tag_lists.tdwg.org/2006-March/000063.htmlhttp://listserv.nhm.ku.edu/cgi-bin/wa.exe?A2=ind0602&L=TDWG-GUID&P=R6382http://listserv.nhm.ku.edu/cgi-bin/wa.exe?A2=ind0602&L=TDWG-GUID&P=R6104 So when you mention that it has been difficult to generate RDF, are
you talking about the Java toolkit that you used to set up Index
Fungorum LSID resolver or are you refering to the port of the client and
server stacks to .NET?
Cheers,
Ricardo
Kevin Richards wrote:
> Ricardo
>
> Do you think this analysis should include a gap analysis of RDF
> software as well , or would it be best to keep this separate?
> I think this is probably a more contentious issue - ie some doubts
> about RDF over XML. I have found setting up an LSID
> resolver/authority reasonably straight forward but trying to implement
> the services using RDF has been a steep learning curve (and limited
> software/tools to help).
> I'll add more later.
>
> Kevin
>