David Nickerson wrote:
> What I'd like to see is that for models which don't have any metadata 
> and/or curation information that rather than a complicated plone site 
> error users are presented with a simple message that the model doesn't 
> contain  any data of the type required for that particular view. 
> Similarly, I'd expect the same kind of page when I click the view math 
> tab for a model with no math in it. Just like the data tab typically 
> gives an "empty" page now.
> 

Seeing that as a critical issue, I have diverted my attention to create new 
pages that use my library to fetch the data.  The 'curation' tab should display 
a nice list of creator/modification data (if such data is present), and users 
of the model upload/editing facilities will be able to edit the modification 
data eventually (creator editing is already available).

However, you will still see that many of them do not actually have data 
populated, due to the various different "methods" of handling of RDF done by 
various software/code that messed up the graph (which more compliant RDF parser 
will further corrupt).  I have identified a couple models that have at least 
some data to demonstrate the new viewer:

http://www.cellml.org/models/boyett_zhang_garny_holden_2001_version01/pmr_view_history

These two will appear to have certain information mission as the RDF graph in 
the original file before it was uploaded to this repository was incomplete 
(certain nodes requiring rdf:parseType="Resource" attribute did not have them):
http://www.cellml.org/models/difrancesco_noble_1985_version01/pmr_view_history
http://www.cellml.org/models/beeler_reuter_1977_version01/pmr_view_history

That was just one of the many different ways the RDF was broken (I have 
encountered) that prevented data from being picked up.  There are many other 
files without the proper data and so nothing would even display.  They will 
need to be repaired eventually, and I have written tools for that for the 
people who are adding/managing the models in the repository.

As for the 'model metadata' tab, it only displays the citation information for 
now, but it will be expanded to include comments for other individual 
components within the CellML file, if they exist, much like the former 
functionality.

Again, it may not display because of bad RDF or mismatching cmeta:id (because 
they were renamed improperly, and they shouldn't be renamed anyway because 
those cmeta:id's could have been referenced by CellML 1.1 models, and so they 
do not get renamed anymore).

As for why I didn't do this sooner, it's because I placed my priorities on 
entering/editing the modification data (which the 'curation' tab displays) and 
move the 'model status' (really I think it should be more generalized, more 
like a comment about the file as a whole) in the tmp-documentation namespace 
into RDF (specifically into the cmeta:comment of the node about the file (i.e. 
rdf:Description node identified by rdf:about="")), as that is unused by all 
models in the repository (or elsewhere, I believe).  I digress.

I didn't touch them because they were not always working correctly and so no 
data gets displayed.  Actually, I did fix the page to trap the exception, but I 
did not see that there was a customized page that was overriding that and so 
the exception page still shows.  Took this change to raise that issue, that too 
has been fixed.

> Just to note that the first model list on the model repository page as 
> well as the first "starred" model both give the error for the model 
> metadata - these are probably the first models the casual repository 
> browser will click on. Hence they might be getting a bad first 
> impression of the model repository.
> 
> As for the tabs required, I guess we'll need to see how well the 
> navigation works on your combined pages. If it all works fine then there 
> shouldn't be any problems just having a single metadata tab. Otherwise 
> we might need to look at further categories into which to separate the 
> metadata (publication, curation, simulation, graphing, etc...). Whether 
> such separation is best accomplished via tabs or some other mechanism we 
> can leave that decision for now.

I have not combined the pages yet, but do check out the ability to collapse 
each pane (by selecting the labels in the upper left corner of each pane) and 
see if that will result in presenting all metadata in a single page as the 
better way.  Also, panes can be nested, so I can group all modification history 
into a single pane and all of them can be collapsed at one go.

Feel free to play with it.

Kind Regards,
Tommy.


_______________________________________________
cellml-discussion mailing list
[email protected]
http://www.cellml.org/mailman/listinfo/cellml-discussion

Reply via email to