[Dev] update RepositoryRegistry
Dieter Van Uytvanck
dieter.vanuytvanck at mpi.nl
Tue Sep 18 12:09:35 CEST 2012
On 10/9/12 13:35 , Matej Durco wrote:
> 1. Proposal to rename the page RepositoryRegistry to FCS-Endpoints
> - reflects better the actual content of the page
> + removes the ambiguity with CenterRegistry.
Hi Matej,
thanks for your suggestion - I've renamed the page (old links stays
available though with a link to the new one)
http://trac.clarin.eu/wiki/FCS-Endpoints
> 2. I am missing a description of how the Center CMDI-files
> end up in the registry.
> Are they to be uploaded into trac:
> http://trac.clarin.eu/browser/CenterRegistry/data/
> and are imported from there?
No, the data has to be entered via a webform, as explained at
http://trac.clarin.eu/wiki/CenterRegistry
We are planning to fill it further in, but currently there is only 1
user account to enter/change data, hope that will be fixed soon so that
centers themselve can enter the data... Will keep you posted.
> 3. Is it correct that the FCS-Endpoints in the Center-record now encoded as
> <WebReference>
> <Website>http://weblicht.sfs.uni-tuebingen.de/rws/sru/</Website>
> <Description>CQL</Description>
> </WebReference>
>
> Is the "CQL" the magic token?
> This looks to me a bit sloppy:
> a) `FCS` would be better than `CQL`
> b) `Type` would be better than `Description`, as Description seems to be
> otherwise used for human-readable description of the website, e.g.:
> <Website>http://corpora.informatik.uni-leipzig.de/</Website>
> <Description>Web access to the corpora of the wortschatz
> project.</Description>
>
> Looking forward to feedback
I agree this way is suboptimal. It's not necessarily the final way to
indicate the FCS endpoints though. We will see what the requirements for
this field are (eg can a center have multiple endpoints - apparently
yes) and then make a proposal for a change. But this is not extremely
urgent I think.
best,
--
Dieter Van Uytvanck
Max Planck Institute for Psycholinguistics, Nijmegen, the Netherlands
tel. +31-(0)24-3521-191 | <http://www.mpi.nl/>
More information about the Dev
mailing list