Re: WA vs. OA

Robert, everyone,

it seems that we have engaged into a false direction. Per

https://www.w3.org/blog/2016/05/https-and-the-semantic-weblinked-data/

the official advice of the Data Activity is to keep to HTTP for namespaces (and contexts). This means that this invalidates:

https://www.w3.org/2016/05/17-annotation-minutes.html#resolution02 <https://www.w3.org/2016/05/17-annotation-minutes.html#resolution02>

as well as making

https://github.com/w3c/web-annotation/issues/230 <https://github.com/w3c/web-annotation/issues/230>

questionable as a consequence. In other words, I propose we stay by http and continue using 'oa' for the namespace URI.

This does not directly affect the options #1 and #2 outlined by Sarven, though, which remain valid options for that case as well, because we have to define what the /ns/oa vocabulary contains in future. As for choosing between #1 and #2 I tend to agree with you. As a possible compromise we can put explicit deprecation for non WA (and keep them) in the namespace document *now* (ie, when publishing CR), with the explicit comments that those terms *will be removed* if and when the WA document is published as Rec.

Ivan

> On 20 May 2016, at 17:30, Robert Sanderson <azaroth42@gmail.com> wrote:
> 
> 
> Hi Sarven,
> 
> To me, option 1 is the preferable one. The CG document is very clearly marked as a draft, and there has been two years for people to see that there's a new version coming. There is still at least 6 months before it will be replaced.  So it's not like this is going to be sprung on anyone out of the blue.
> 
> Also, given that properties that are not understood are ignored, the older properties will just fall away when an original system tries to create an annotation in a new server, or vice versa. The vast majority of systems will still be compatible for the majority of structures if they use RDF, and will be completely out of date for JSON-LD as every key has been renamed.
> 
> The brand name of "Open Annotation" has had thousands of people using it for 7+ years since we coined it in late 2008 (13th of November it turns out, thanks whois) :)
> 
> Rob
> 
> 
> On Fri, May 20, 2016 at 5:14 PM, Sarven Capadisli <info@csarven.ca <mailto:info@csarven.ca>> wrote:
> On 2016-05-20 10:40, Robert Sanderson wrote:
> 
> Hi Sarven,
> 
> Yes, there are backwards incompatible changes such as the removal of
> SemanticTag and oa:serializedBy (etc) replacements. The new ontology
> should thus replace the OA CG *draft* namespace.
> 
> Rob
> 
> Option 1: Update OA namespace with WA vocab
> Option 2: Update OA namespace with WA vocab and deprecate OA vocab
> Option 3: Create WA namespace with WA vocab
> Option 4: ?
> 
> Option 1 gives dead-ends to existing deployments (although it was a draft). A good practice is to mark the changes so that deprecation could be phased out, which is why option 2 may be an okay compromise. Perhaps there are no tangible differences between options 1 and 2 at the end. Option 3 is costly, but with some benefit to using https right away.
> 
> I think option 2 is doable if the definitions are not concept drifts. Handling deprecation in the vocab is simpler.
> 
> -Sarven
> http://csarven.ca/#i <http://csarven.ca/#i>
> 
> 
> 
> 
> --
> Rob Sanderson
> Semantic Architect
> The Getty Trust
> Los Angeles, CA 90049


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