Hi,

1. prefix for <http://www.bl.uk/schemas/bibliographic/blterms#>: I know prefixes are arbitrary but the preferred prefix for the British Library RDF Schema is blt:
2. cc:license. Could we use dct:license throughout instead? It would save adding yet another namespace. I saw that BBC RES was using that property for their Shakespeare dataset
3. Could we have a look to see if the Learning Resource Metadata Initiative (LRMI) provides all the terms we need rather than creating a BL ontology? http://dublincore.org/dcx/lrmi-terms/1.1/ . There's a Special Session on LRMI at DC-2016 next month http://dcevents.dublincore.org/IntConf/index/pages/view/specSession#lrmi so this looks like an ontology very much "alive"
4. I don't think you can use owl:sameAs for bl:pdfLink. The [filename].pf is not the resource but the electronic location of the resource. RDA has got a property called "has uniform resource locator", either http://rdaregistry.info/Elements/u/P60511 (unconstrained set) or http://rdaregistry.info/Elements/m/P30154 (constrained set: manifestation properties)
5. event: time. I think the way this property is used throughout the current proposal "breaks" the range defined in the Event ontology. My understanding is that it cannot have a range of Literal.
6. Question: what is the rationale/benefit of having granular relationships properties, e.g. bl:relatedArticle, bl:relatedWork, etc. rather than a generic property such as dct:relation?

Thanks, Corine

I agree with Corine's comments above, and have some additional ones:

1. You could use rdfs:seeAlso for bl:pdfLink.
2. It would be good if teachingSubject linked to an actual resource (e.g. DBPedia resource), rather than being a string literal, e.g. http://dbpedia.org/resource/English_literature. Though this would probably need some manual intervention to ensure that the correct topic was linked to.
3. The nativeLanguage predicate could use values from https://tools.ietf.org/html/bcp47 (as used to specify the language of an RDF string literal).
4. ageRange might be better split into "from" and "to" values, otherwise a client has to parse the string to work this out.
5. If I understand the intention of the void:dataset property correctly, I think you probably need void:inDataset (http://vocab.deri.ie/void#inDataset) instead.
6. It would be useful if the #resource itself had a foaf:primaryTopic to a DBPedia resource, e.g. for the above example http://dbpedia.org/resource/Wuthering_Heights.

Suggest reviewing the Learning Resource Metadata Initiative http://lrmi.dublincore.net/lrmi-schema/
Also there has been work integrating LRMI with schema.org - e.g. http://schema.org/learningResourceType which may be worth reviewing - Phil Barker is the person to talk to on this

Quick comment; Phil Barker (@philbarker) pointed me at the data embedded in http://www.bbc.co.uk/education/clips/z4f2fg8. This uses URIs from http://www.bbc.co.uk/education/levels/z3g4d2p for curriculum levels - might be good to use these esp. for RES

As I don't have a resource in the CMS to describe the curriculum, I can't easily create a new deferenceable endpoint at <http://www.bl.uk/curricula/UKNationalCurriculum>

So instead I've created a #curriculum node that contains the assertions described in http://museum-api.pbworks.com/w/page/112417531/RDF%20Definition%20-%20Curriculum

Does this sound reasonable?

I think this is a reasonable compromise. Would you have one of these #curriculum resources for each teaching resource, or would you have a single one referenced by all of the teaching resources? The latter would be preferable, but the former would be a reasonable work-around.

In this current release the #curriculum is included in each teaching resource as the endpoints are driven by the available content urls in the cms.

At some point today I'm going to update the wiki with what has been implemented in the current planned release (which I'll refer to as v1) and indicate what should be considered in a future release (vnext)