Re: schema.org upcoming topics

Hi Dan,

Just to give an ELIXIR perspective: as I see it, dataset descriptions are a fairly complicated area where there are a number of different sub-communities each with a slightly different take on the subject.

In life sciences, the W3C Health Care and Life Sciences group has proposed a community profile [1] based on DCAT, PROV, VoID but as far as I am aware is mainly applied to RDF datasets. On the other hand the NIH Big Data 2 Knowledge Biocaddie initiative, which is about dataset discoverability, has come up with a new, encompassing DATS schema [2] that maps to various other metadata exchange models both generic and domain-specific. They plan to map the DATS model to schema.org through the Bioschemas community [3] so you may see some schema.org activity in the near future.

At the ELIXIR hub we have been discussing some potential use cases for schema.org dataset descriptions that aren’t about discovery as well but these are at an early stage.

One thing I would be interested to hear about, is who is using schema.org to describe datasets and what are their use cases?

[1] https://www.w3.org/TR/hcls-dataset/ <https://www.w3.org/TR/hcls-dataset/>
[2] https://github.com/biocaddie/WG3-MetadataSpecifications <https://github.com/biocaddie/WG3-MetadataSpecifications>
[3] http://bioschemas.org

Kind regards
Andy

  
Andy Jenkinson <mailto:andy.jenkinson@elixir-europe.org>
ELIXIR Data Co-ordinator
www.elixir-europe.org <http://www.elixir-europe.org/>

ELIXIR Hub, South Building
Wellcome Genome Campus
Hinxton, Cambridge, CB10 1SD, UK
Tel: +44 (0) 1223 492618
E-Mail: andy.jenkinson@elixir-europe.org <mailto:andy.jenkinson@elixir-europe.org>  <http://www.elixir-europe.org/>                 

> On 29 Apr 2016, at 00:32, Dan Brickley <danbri@google.com> wrote:
> 
> On 28 April 2016 at 23:40, R.V.Guha <guha@guha.com> wrote:
>> What is the current state of data set descriptions?
> 
> 1. we have some very basic pieces in schema.org for talking about
> datasets, more or less inspired by DCAT and related RDF vocabularies.
> http://schema.org/Dataset and nearby.
> 2. W3C CSVW is complete. It defines its own basic vocab for talking
> about tabular data including a medium-expressive templating system to
> map tables into triples.
> 3. This release includes some bugfixing for problems we introduced in
> 2.0 (messed up a property name)
> 
> https://github.com/schemaorg/schemaorg/issues/1083 tracks "Improving
> Dataset descriptions" including notes from chatting with Natasha Noy
> (cc:'d). We've also had a few discussions e.g. with folk around
> https://www.elixir-europe.org/ on applying this stuff more deeply to
> dataset sharing in the life sciences.
> 
> I've also been toying with ideas for describing neural net model zoos
> along lines of https://github.com/BVLC/caffe/wiki/Model-Zoo but that
> might be a bit niche interest :)
> 
> Dan
>