Re: Standards Requirements

Hi Doug, all,

On Thu, Nov 27, 2014 at 2:29 PM, Doug Schepers <schepers@w3.org> wrote:

> Since the Web Annotation Data Model spec is not yet complete, and because
> it's not much different than the Open Annotation data model, it seems
> sensible to compromise by allowing annotation tools to only support the
> Open Annotation data model initially.
>

Great, thanks for your support :)



> As you note, the timing is the issue.
> Right now, in the Annotator codebase, there is code for Open Annotation
> data model ready for integration in Annotator 2.0; you know this, of
> course, since you helped write it. However, it will be a couple weeks
> before the Hypothesis folks complete the integration into Annotator 2.0,
> and a few weeks more before they complete rebasing the Hypothesis fork with
> the Annotator 2.0 code. So, while supporting the Open Annotation data model
> is a high priority, it will be several weeks before the version of
> Annotator that we're using at WebPlatform.org will support it.
>

So, to be clear, the system that WebPlatform.org will work from the base
Annotator 2.0 rather than the hypothes.is Annotator fork?
>From the annotator roadmap at
http://docs.annotatorjs.org/en/architecture-proposal/roadmap.html, 2.0
looks like it will be finalized around the beginning of February 2015, or
about 8 weeks spanning the winter holiday period.


> We want to publish the Web Annotation Data Model spec as a first public
> working draft on 9 December, as you know.
> Many of us would like to include the Annotator tool with the FPWD. Not
> only would it be easier for us to collect feedback, but it would illustrate
> and publicize the use of Web Annotations to readers. We will use the
> Annotator with other specs, so people can see it there, of course, but
> showcasing it on the FPWD of the Web Annotation WG's first spec has a
> certain aspirational and inspirational aspect.
>

I would be very happy to have annotation enabled at publication time.



> Here's the options, as I see them:
>
> 1) Publish the Web Annotation Data Model spec as planned on 9 December,
> with the Annotator tool as it stands today.
> 2) Publish the Web Annotation Data Model spec as planned on 9 December,
> without the Annotator tool, and only publish a version with the Annotator
> once it supports the Open Annotation data model.
> 3) Delay publication of the Web Annotation Data Model spec until the
> Annotator supports the Open Annotation data model.
>
> I personally advocate for option 1, and think option 3 is unreasonable.
> Option 2 seems like a lost opportunity to me.
>

And we can confirm that the most desirable option

0) Publish the Web Annotation Data Model spec as planned on 9 December,
with an annotation tool that supports the Open Annotation data model.

is not available?


> If there is a desire and urgency to to have interchange with other
> annotation clients or servers before Annotator supports the OA data model
> natively, perhaps we could document Annotator's current data model, and
> offer an export/import option and a simple converter between the Annotator
> data model and OA data model; I don't know how difficult this would be.
>

We have a sandboxed Annotator that does this in and out in Mirador, which
took about a day of effort including having the (paired) developers
understand the mapping between annotator and open annotation.

https://github.com/IIIF/m2/blob/annotation-menus/js/src/widgets/annotationsLayer.js

So not very difficult :)

My perspective is very similar to yours:  3 is unreasonable, 2 has the
potential to miss the opportunity to engage with readers directly for no
strong reason as we don't, today, have more applications in mind that would
require or make use of the information and are unlikely to have before the
end of January.  Thus 1 is the most reasonable, lacking option 0.

Getting a firm deadline and commitment that OA will be supported in the
webplatform.org system seems the best approach.  That exact date is
obviously the result of negotiation and discussion. My suggestion, given
the amount of work is actually very small, is end of February 2015 to allow
for the holidays and other scheduling requirements.

Best,

Rob

On 11/26/14 7:11 PM, Robert Sanderson wrote:
>
>> Dear all,
>>
>> My thoughts on requirements for standards support in the spec annotation
>> software to be used.
>>
>> * We clearly must support the model and protocol coming out of the Web
>> Annotation WG.  The timing is less obvious, as it has only just begun.
>> Engagement and implementation early will drive the WG faster and further,
>> but is reliant on the software providers' good will to track a spec that
>> is
>> certain to change.  Setting too low a bar will reduce the value as a
>> demonstrator, while setting too high a bar will make it too costly to
>> implement.  I think getting this right will be critical for the success of
>> the experiment, and towards Annotation in the W3C moving forwards.
>>
>> * Supporting the Open Annotation CG specification seems like a good
>> requirement as it will not change, has multiple implementations already,
>> and would make transition to the WG spec very easy.  This is my suggestion
>> as a way to hedge our bets on the WG spec output:  by requiring support
>> here, we do not have to keep client(s) and server(s) in step with a
>> changing document, but ensure that they are already at (say) 80% of the
>> final solution.
>>
>> * As the OACG did not specify any transport protocol, following basic REST
>> with the default JSON-LD context seems like the most interoperable
>> baseline.
>>
>> * For extended protocol features such as search, notify, and so forth, I
>> don't know of any significant prior art though would welcome references
>> towards the WAWG work!
>>
>> Best, and happy thanksgiving,
>>
>> Rob
>>
>>
>
>
>


-- 
Rob Sanderson
Technology Collaboration Facilitator
Digital Library Systems and Services
Stanford, CA 94305