Re: ELIXIR tools registry participation - richer metadata for our packages
Hi,
short greetings from Vietnam. :-)
On Thu, Nov 20, 2014 at 05:37:10PM +0000, Booth, Timothy G. wrote:
> This is Steffen and Tim sharing a desk at the Copenhagen Hackathon
> of the ELIXIR-DK Catalog of resources in computational biology.
> There is a general excitement of the collection of tools that
> are associated with Debian Med and its derivatives and we are
> here to
Cool. Thanks for your work on this.
> Catalog entries are meant to be provided by the maintainers of
> the software tools in the ELIXIR network. For resources (binaries)
> provided through the Linux distros, we could certainly just fall
> back to the information we already have, but we would like you
> (this list) to comment on the extension of the
> debian/upstream/metadata
> file to accommodate also structured references to semantical
> catalogs like the EDAM ontology. There is a related effort by
> Matus to annotate the DebTags. The format we think about is like
>
> Ontology: http://prefix.of.ontology.org
> feature_name: ontological_description_of_that_feature
> another_feature: id<blank>human_readable
> scope: <list of binaries> | summary
> feature_name: ...
> another_feature: ...
Sounds sensible.
> ...
> or if we want to reduce the level of detail to just the summary this could be compressed to:
I think the more explicite form has advantages - as long as we can motivate
maintainers to fill in these details. Perhaps we should be explicite to
ELIXIR people that they are very welcome to commit this kind of data
directly into Debian Med VCS.
> If the list likes this approach, then we can continue annotating a bit more and amend our task pages for it all.
I agree that we should think about ways to add this information to our tasks
pages. Any design suggestion is welcome.
> Some tools and suites (eg. EMBOSS) have existing annotations from other projects that we can inherit.
> We are not yet confident about what this effectively means e.g. for the Ultimate Debian Database. @Charles, can you
> direct us, please?
>From my point of view we are quite flexible if it comes to the question of
injecting data into UDD. Currently the only information from
debian/upstream/metadata which is taken over into UDD is the citation
information where we created a dedicated table for. I could perfectly
imagine another table which keeps this classification information.
Kind regards from Vietnam
Andreas.
--
http://fam-tille.de
Reply to: