[Developers] Cactus RDF: URIs

Thomas Radke tradke at aei.mpg.de
Thu Sep 7 12:30:40 CDT 2006


Dylan Stark wrote:
>>Of course, thorns change with time, so even that identifier does not
>>have a static meaning.
>>
> 
> 
> That's true that the thorns change, it would just be a name for the
> concept of the thorn. Versioning information could be maintained (in
> RDF) somewhere (ideally at that address). Is there a good/clean way of
> specifying what 'version' of a thorn an exe was built with? Maybe a
> checkout date?

That is a very good question we would like to find an answer for. There 
was a request from our physicists that the Cactus metadata management 
should store such version information so that they can later do queries 
like "compare these two simulations and list the differences in the 
source code used to build the executables" (see the bottom of 
http://www.cactuscode.org/old/pipermail/users/2006-August/001955.html).
We thought about adding CVS version info of thorns' source files to the 
simulation metadata but found this to be insufficient. Local code 
modifications aren't covered, same for code which isn't managed via CVS 
in the first place (like Carpet).

I see no simple solution to this problem; the only method I can think of 
is using the source code tarballs archived by Formaline and compare 
those. In AstroGrid we are planning to implement this functionality, not 
using the metadata service alone but in combination with some external 
file management service (which keeps track of all archived Formaline 
files generated by Cactus simulations).

-- 
Cheers, Thomas.


More information about the Developers mailing list