Re: Updated Invitation: Linked Life Data @ Mon 2013-05-13 11:00 - 12:00 (HCLS)

Regrets, last minute issues :(
I'll check the docs!

best,
Andrea

Il giorno 12/mag/2013, alle ore 23:24, Health Care Life Sciences <w3.hcls@gmail.com> ha scritto:

> 
> This event has been changed.
> more details »
> Linked Life Data
> Changed: This meeting will be held using fuze: please join at http://fuze.me/19843842
> Thanks to Michel for hosting!
> Duration: ~1 hour
> (Variable) frequency: ~weekly
> Convener: M. Scott Marshall
> 
> Session Theme: Metadata for data discovery and dataset description using SPARQL
> 
> Agenda:
> * look at statistics - All
> * AOB
> 
>  Relevant docs:
>   the abstraction
>   - https://docs.google.com/drawings/d/1e6qsxPkc-qKecVTJGJePE1Nuy2sD8Puu-FsEtUtGC-o/edit?disco=AAAAAFWHcnw
>   sample implementation using chembl: 
>   - https://docs.google.com/file/d/0B4y0zfdRviKsS1l2NEttN3pfc1k/edit
> * Remaining metadata attributes in working draft - All (time allowing)
>    - Working draft:
>    - https://docs.google.com/spreadsheet/ccc?key=0Aoy0zfdRviKsdFJWTDFpblNXc3BtelhrdEpNYTdvbXc#gid=1
> 
> 
> **********************************************************************
> Notes from Mon. Apr. 29 (Thanks Alasdair!):
> 
> Michel presented the dataset description model for unversioned -> versioned -> formatted -> data
> 
> Overview: https://docs.google.com/drawings/d/1e6qsxPkc-
> qKecVTJGJePE1Nuy2sD8Puu-FsEtUtGC-o/edit?disco=AAAAAFWHcnw
> Detailed: https://docs.google.com/file/d/0B4y0zfdRviKsS1l2NEttN3pfc1k/edit?
> usp=sharing
> 
> Only an RDF data item would be able to point back to its description due to the allowances of the data model.
> 
> Discussion on whether to repeat metadata at all levels. Each description would be complete: simplifies query; redundancy in metadata. Flexibility of the URIs for each level means you have several entry points. Could have contradictions in the metadata at the different levels. Idea would be to limit the metadata in the unversioned part to minimal data that would not change over time.
> 
> Abstract data format, e.g. triplestore. One option would be to model it as a separate
> versioned/formatted dataset description. However there is an underlying formatted dataset that has been loaded into the the underlying datastore and this is what would be described. For a relational database accessible through D2R the description would be a SQL versioned dataset with an accessibility protocol of SPARQL. Service points to the versioned/formatted dataset that they expose.
> 
> Different syntaxes provide different views of the data. In RDF you can capture the relationship between the data and the metadata. This is not generally possible in other syntaxes.
> 
> Sources: points to the exact file that was used so that bugs can be tracked
> 
> MIRIAM is a catalog that described datasets. Catalog was added.
> 
> Format types would be captured with URIs.
> EDAM are amenable to extending to cover the file type that we require, Nick has been in contact with Jon Ison. http://edamontology.org/page
> 
> ToDo: Go through different scenarios, e.g. Bio2RDF, Open PHACTS, MIRIAM and see how these look in the model (not using any particular vocabularies). By hand generate a full description for a dataset.
> 
> ToDo: Revisit properties in the spreadsheet to ensure that they are all still required.
> 
> Do send questions and comments to the list!
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> 
> When	Mon 2013-05-13 11:00 – 12:00 Eastern Time
> Where	#hcls (map)
> Calendar	HCLS
> Who	
> •	
> w3.hcls@gmail.com - organizer
> •	
> public-semweb-lifesci@w3.org
> •	
> dbcatalog
> Going?   Yes - Maybe - No    more options »
> Invitation from Google Calendar
> 
> You are receiving this courtesy email at the account public-semweb-lifesci@w3.org because you are an attendee of this event.
> 
> To stop receiving future notifications for this event, decline this event. Alternatively you can sign up for a Google account at https://www.google.com/calendar/ and control your notification settings for your entire calendar.
> 
> <Allegato di posta elettronica.ics><invite.ics>