Re: A question on RWPM: why the 'metadata' tag?

I may have lost the thread in the conversion conversation.


I just used the JSON-LD playground

https://json-ld.org/playground/

to visualize the Moby-Dick example

https://github.com/HadrienGardeur/webpub-manifest/blob/gh-pages/examples/MobyDick/manifest.json

and the graph that comes out does contain more than the metadata.


So, it seems my concern about data loss during conversion was not a concern.


However, mapping "identifier" to the JSON-LD's "@id" (which is what's currently done in the context), does seem dangerous if the primary use of a manifest is to talk about a publication's address (and what's represented there). There certainly should be places to put additional identifiers, but the primary one should be `@id` or `id`.


But, that's probably a bug I should file elsewhere.


Cheers,

Benjamin


--

http://bigbluehat.com/

http://linkedin.com/in/benjaminyoung

________________________________
From: Ivan Herman <ivan@w3.org>
Sent: Tuesday, January 9, 2018 11:02:30 AM
To: Benjamin Young
Cc: Hadrien Gardeur; Leonard Rosenthol; W3C Publishing Working Group
Subject: Re: A question on RWPM: why the 'metadata' tag?



On 9 Jan 2018, at 17:00, Benjamin Young <byoung@bigbluehat.com<mailto:byoung@bigbluehat.com>> wrote:

The result, though, is that the RDF has none of the hypermedia value that the JSON-LD had.


I am not sure I understand this remark. Can you elaborate what you mean?

I.

Ideally, you can get the same value out of all/most expressions of the data model. If not, then something needs fixing.

--
http://bigbluehat.com/
http://linkedin.com/in/benjaminyoung
________________________________
From: Hadrien Gardeur <hadrien.gardeur@feedbooks.com<mailto:hadrien.gardeur@feedbooks.com>>
Sent: Tuesday, January 9, 2018 10:44:59 AM
To: Ivan Herman
Cc: Leonard Rosenthol; W3C Publishing Working Group
Subject: Re: A question on RWPM: why the 'metadata' tag?

My main problem is not the blank node. I know that many in the LD community are up in arms against it, but I may be less sensitive.

My problem is making owl:sameAs a central aspect of the data; what I expect that, in practice, all the linked data that is hanging on that blank node will be, essentially, useless in (Linked Data) practice, because tools will not reach that data (apart from OWL reasoners and triple stores that know how to reason with OWL terms, which is a small minority of the usage).  My claim is that the current structure produces possibly lots of Linked Data that will become totally useless… then why do it?

A possibility is that _only_ the JSON object in metadata should be considered as RDF metadata, and all the rest is ignored. I am not yet sure how to achieve that with @context  (maybe by using @context only in that object, I have to check with the JSON-LD spec).

If we don't care about the default reading order in RDF, then we could simply remove spine from the default context document.

The Turtle output would then be:

<urn:isbn:978031600000X> a schema:Book ;
    schema:author "Herman Melville" ;
    schema:dateModified "2015-09-29T17:00:00+00:00"^^xsd:dateTime ;
    schema:inLanguage "en" ;
    schema:name "Moby-Dick" .


----
Ivan Herman, W3C
Publishing@W3C Technical Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
ORCID ID: http://orcid.org/0000-0003-0782-2704