On 6/12/07, Tommy Yu <[EMAIL PROTECTED]> wrote:
> Matt wrote:
> > On 6/11/07, Tommy Yu <[EMAIL PROTECTED]> wrote:
> >> Matt wrote:
> >>> I have concluded that they are now talking about the web site and not
> >>> keywords in general.
> >>>
> >>> My assumption was that the category field selections are not persisted
> >>> in the model metadata at all.
> >>>
> >> Actually they are.  Keywords are defined in the CellML metadata 
> >> specifications and are already being used in various files.  Feel free to 
> >> check the CellML files of the old repository and scroll down the to 
> >> keyword section.  An example follows.
> >
> > You are talking about keywords, I was talking about these new things
> > Peter brought up called Categories.
> >
>
> I will say it again.  Categories are not 'new things'.  It is something 
> specific to the CellML Repository I was told to implement.  The options in 
> the category are keywords themselves.  Items there are not 'new things'.

I was simply pointing out that it hadn't been said prior to that.

>
> They are just keywords hand-picked by Peter to show up in the drop box that 
> happens to be labeled as categories.  Nothing more.
>
> Before you ask further, I only did what I was told.
>
> >
> >> >From http://www.cellml.org/examples/models/beeler_reuter_model_1977.html:
> >>
> >>       <!--  Keyword(s) -->
> >>       <bqs:reference rdf:parseType="Resource">
> >>         <dc:subject rdf:parseType="Resource">
> >>           <bqs:subject_type>keyword</bqs:subject_type>
> >>           <rdf:value>
> >>             <rdf:Bag>
> >>               <rdf:li>ventricular myocyte</rdf:li>
> >>               <rdf:li>electrophysiological</rdf:li>
> >>             </rdf:Bag>
> >>           </rdf:value>
> >>         </dc:subject>
> >>       </bqs:reference>
> >>
> >> I do understand it may be different from the full CellML metadata 
> >> specification as found in 
> >> http://www.cellml.org/specifications/metadata/cellml_metadata_1.0#sec_bqs, 
> >> but all other models pretty much follow this RDF format and so I wound up 
> >> having to follow the above format to pick up the keyword metadata.
> >
> > It's different, but says the same thing, even though the graph comes
> > out different.
> >
> > Personally I think we should probably dump bqs; the rdf schema we
> > advertise is non-standard and broken.
> >
>
> I very much agree, it was not easy working with and around it, especially 
> given the pile of cruft that is already in place that I have to deal with...
>
> > Dublin core is not the easiest thing to follow, but at least it is
> > standard and used in the world, so we should at least keep that.
> >
>
> If someone come up with a proper metadata specification based only on 
> industry standard with everyone in the community agreeing with use it I would 
> be happy as it should make my life easier when dealing with CellML metadata.
>
> Also, I don't want to be dealing with five different graphs that says the 
> same thing.  I have enough headaches dealing with constructing multiple versa 
> queries to pull relevant data as it is.
>
> >>> I would have liked some indication that the 'categories' used also end
> >>> up in the model keywords attributed to the model in addition to the
> >>> keywords supplied by the author when creating or uploading the model.
> >>>
> >> That is already the case, the 'categories' *are* keywords that are chosen 
> >> by Peter as a selectable choice in the filtering drop box for the 
> >> repository listing.
> >
> > You need to make sure the keywords then are ordered collections so
> > that you can create some rule for your 'category' interpretation.
> >
>
> Not really.  I don't care what they are ordered.  If it exists it gets 
> "highlighted" as per demand.

See below.

>
> > I don't like this special attention that certain keywords gets.
> >
>
> They get no further treatment aside from the "special attention" which is to 
> limit the listing of keywords.  Peter does not want that list to be filled 
> with over 100 keywords.  A condensed listing was desired and I implemented 
> with keywords as those keywords used must exactly describe what the models we 
> have in the repository (and the old repository) are about.  Feel free to 
> check these links out.

This is the special attention that doesn't make sense. It has no data
model equivalent within the metadata specification.

>
> http://www.cellml.org/examples/repository/index.html - Under Table of 
> Contents - Basically the same list of "categories".
> http://www.cellml.org/models/pmr_search - Feel free to see/search by all 
> keywords with the 'Keywords' selection box.  It works exactly the same as the 
> category filter found on top of the main listing except the short list found 
> on the main listing is a predefined list of keywords.
>
> Again, having the "special category listing" that only includes the few 
> chosen keywords is intended to be an usability aid for the filtering of the 
> major types of models we deal with, which happens to be described via the 
> keyword metadata.
>
> If you still think category describes something new (at least in the way it 
> is implemented by the CellML repository), I don't know what else to convince 
> you that it isn't.

It does describe something new. It's "a special kind of keyword" is
about all I can see at the moment with no specification of how that is
managed. Other software will have no reference vocabulary from which
to offer these to authors. I was offering a way to help build such a
reference vocabulary.


>
> >>> I would like there to be as many keywords allowed as the
> >>> author/uploader wants (perhaps just a lines field will do for now for
> >>> this). Constraining them to a single extra keyword in addition to a
> >>> selected category makes no sense to me.
> >>>
> >> In the Edit Keyword interface, any keyword of the model that matches one 
> >> of the 'blessed' keywords will be highlighted in the category list.  All 
> >> other keywords will be in the lines field editor.  Feel free to log into 
> >> the site (I assume you have an account) and try out the editing interface. 
> >>  I do agree it is currently slightly clunky, but James has no complaints 
> >> with it and he has already added/verified keyword for half the curated 
> >> models (I think) of the repository.
> >>
> >> Again, the category field is a *subset* of keywords to limit the number of 
> >> choices in the filtering menu and not a distinct entity.
> >>
> >> Hope this clear things up.
> >>
> >> Tommy.
> >>
> >> _______________________________________________
> >> cellml-discussion mailing list
> >> [email protected]
> >> http://www.cellml.org/mailman/listinfo/cellml-discussion
> >>
> > _______________________________________________
> > cellml-discussion mailing list
> > [email protected]
> > http://www.cellml.org/mailman/listinfo/cellml-discussion
>
> _______________________________________________
> cellml-discussion mailing list
> [email protected]
> http://www.cellml.org/mailman/listinfo/cellml-discussion
>
_______________________________________________
cellml-discussion mailing list
[email protected]
http://www.cellml.org/mailman/listinfo/cellml-discussion

Reply via email to