Re: Friendly JSON serialization (Was: Annotation Serializations)

Hi Doug,

On Wed, Jan 22, 2014 at 11:51 PM, Doug Schepers <schepers@w3.org> wrote:

>
> Rob's new vocabulary and mapping is great; it's pretty much just what I
> was asking for. (I won't quibble now about the few terms I'd change.)
>

Please quibble away, if you have time Doug! :)


I don't know enough about the RDF community to know if standardizing around
> these new terms for all contexts would be tolerable; I'm sympathetic to
> Waleed's viewpoint, and think 2 different sets of terms is asking for
> trouble, but since I'm not going to be the one consuming the RDF, I don't
> feel qualified to judge.
>

As Ivan said, even if we changed all of the OA defined terms (and I would
be happier if there were fewer of those, not more) then we still have all
of the others from the other vocabularies that we use directly.

I, personally, don't think it's a big deal, and is what JSON-LD context
documents were designed for -- to bridge the world views of the application
web and the semantic web.  If I was starting again from zero today, my
steps would be:

* Write out the desired structure in JSON
* See which constructions really aren't possible in JSON-LD
* Formalize the JSON-LD structure
* Look for existing terms to map the fields to
* Mint new terms where there are no existing terms.

Thankfully, I think we would end up very close to what we have today anyway
(modulo the potential names in JSON-LD being different from in the ontology)

Agree that for RDFa the problem is much greater, and Ivan will necessarily
be burnt on the stake. I suggest sending sympathy cards early. :D

Rob



>
> Regards-
> -Doug
>
>
> On 1/22/14 3:06 PM, Karen Coyle wrote:
>
>> I hope I'm not over-interpreting Ed's, Ivan's and Doug's points, but
>> _I'd_ summarize the point as this: that for a standard to be widely
>> adopted, it needs to be as easy to learn and implement as HTML and/or a
>> WordPress or Drupal module. In fact, the latter two would be excellent
>> tools if such is possible.
>>
>> I work with a number of small organizations, mostly archives, that not
>> only do not have a programmer on staff, but do not have an "actual" web
>> master. Instead, once a web site is set up (usually by a consultant) it
>> is then maintained as needed by someone on staff who has learned just
>> enough to keep things running. Yet these organizations have unique
>> collections that could be of great value if linked.
>>
>> kc
>>
>> On 1/22/14, 10:30 AM, Edward Summers wrote:
>>
>>> On Jan 22, 2014, at 1:16 PM, Robert Sanderson <azaroth42@gmail.com>
>>> wrote:
>>>
>>>> So, did you have any examples of what is useful or not useful from a
>>>> tool perspective?  Everything in the current data model is based on
>>>> use cases and requirements, perhaps not requirements for everyone,
>>>> but requirements none-the-less.
>>>>
>>>
>>> Nope, I was just writing to agree that (more?) library support for
>>> working with OA annotations is an important idea. Just saying that
>>> some data is in JSON doesn’t mean it’s useful for a particular task.
>>> It just means you can call JSON.parse on it (yay) instead of having to
>>> install some esoteric RDF toolkit to work with it (boo).
>>>
>>> //Ed
>>>
>>>
>>
>
>
>

Received on Thursday, 23 January 2014 15:51:10 UTC