W3C home > Mailing lists > Public > public-annotation@w3.org > November 2015

Re: Advice on Referencing External Vocabularies

From: Benjamin Young <bigbluehat@hypothes.is>
Date: Wed, 4 Nov 2015 15:36:05 -0500
Message-ID: <CAE3H5FL+Cu8MVt6jwRC2w0bT4SuaNG3PqFV2nTK+th3egYvmEQ@mail.gmail.com>
To: W3C Public Annotation List <public-annotation@w3.org>
Cc: Doug Schepers <schepers@w3.org>, James M Snell <jasnell@gmail.com>, Dan Brickley <danbri@google.com>
Thanks for the input, all!

Mostly, I think this boils down to what we use on the "right hand side" of
this document:
http://www.w3.org/ns/anno.jsonld

And consequently, what folks using triple stores, SPARQL, or some other
more complete-RDF stack will have to "deal with" by way of mapping and
matching terms. As James mentioned, "There
are mechanisms to help smooth those bits over, as necessary, at other
layers in the stack."

Additionally, both `schema` and `dcterms` are provided in the "initial
context" for RDFa:
http://www.w3.org/2011/rdfa-context/rdfa-1.1
"essentially, RDFa users can use these prefixes without having the
obligation of defining the prefixes in the HTML code"

Which, imo, means they're all fair game as they're likely to also be quite
common in RDF consumption in general.

I do believe we're likely to mix in some Schema.org terms. I don't think
it's likely (or prudent) to drop any Dublin Core usage simply because of
Schema.org's popularity--especially given DC's usage within the library and
publishing sector...which, consequently, are pretty "into" annotation as it
turns out. ;)

On Wed, Nov 4, 2015 at 11:18 AM, James M Snell <jasnell@gmail.com> wrote:

> For Activity Streams 2.0, the approach we have taken thus far is to
> ensure that the majority of implementers would not need to rely on an
> external vocabularies in order to properly grok the information being
> expressed. External vocabularies can and likely should be used in
> "real" applications that use AS2, but it is not a requirement and
> implementers that do not understand the same set of external
> vocabularies can still extract valuable information out of the data.
>
> That does end up meaning that parts of the AS2 vocabulary overlap with
> common terms from other vocabularies but that's perfectly ok. There
> are mechanisms to help smooth those bits over, as necessary, at other
> layers in the stack.
>

Given that AS2 is built upon the meaning of existing AS1 JSON key
definitions, this makes good sense.

In as much as `Annotation` is a new Class (building on Open Annotation Data
Model), we'd be best served (afaik) to stick with what started there and
map our vocabulary term to commonly used terms from other
vocabularies...and call it a day. :)

The agnostic (or perhaps ecumenical ;) ) approach makes the most sense. I'd
restrain ourselves to the RDFa "initial context" list (linked
above)--because expressing annotation in HTML seems like a Good Thing.

Beyond that, though, switching what we've got toward a "single" vocabulary
doesn't seem to get us much...except more emails and/or editorial work. ;)

Thanks again for the input from the "outside", Dan & James!
Benjamin
--
Developer Advocate
http://hypothes.is/


>
>
>
> On Tue, Nov 3, 2015 at 11:12 PM, Doug Schepers <schepers@w3.org> wrote:
> > Hi, Dan–
> >
> > Thanks for the discussions at TPAC.
> >
> > (Context: Danbri is the coordinator for Schema.org, one of the
> contributors
> > to Dublin Core, founder of FOAF, and a long-time SemWeb expert,
> experienced
> > in both application development and in standards. I asked him over dinner
> > what approach we should use in referencing external vocabularies for our
> > terms.)
> >
> > If you'll recall, I asked you for advice on what vocabulary to reference,
> > and relative influence and usage of `dc-term`s vs Schema.org.
> >
> > I was surprised by your answer… If I understood correctly, you suggested
> not
> > using any one canonical external vocabulary in our spec, but rather to
> offer
> > a set of equivalent vocabulary terms that might be used, depending on the
> > project. On the one hand, this makes sense, and is a decentralized
> solution;
> > on the other, it doesn't really reduce the complexity, as I'd hoped to
> do by
> > referencing only a single external vocabulary. Could you explain the
> > rationale there, or correct my misunderstanding?
> >
> > Also, I asked about patterns of usage in `dc-term`s and Schema.org. My
> > understanding was that Schema.org had already overtaken the usage of
> Dublin
> > Core in the wider Web (though perhaps not in older libraries), and that
> it
> > would be easiest for future developers if we used Schema.org; TimBL
> > suggested during our F2F that more projects, and thus more tools,
> natively
> > understood Dublin Core today; ultimately, I guess we need to figure out
> the
> > right balance (or, maybe not, if we follow your advice on including
> multiple
> > references). I think you had a more nuanced answer on usage patterns,
> too.
> > Can you speak to that as well?
> >
> > All your explanations made sense to me at the time, but not enough for
> me to
> > convey facts and explain it to others in this WG… I appreciate your
> helping
> > us sort out some long-standing (if not particularly contentious) issues.
> >
> > Thanks–
> > –Doug
> >
> >
> >
>
>
Received on Wednesday, 4 November 2015 20:36:33 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 18:54:42 UTC