Re: API design principles - HTMLXML literals

good afternoon john,

> On 2015-10-12, at 10:56, John Walker <john.walker@semaku.com> wrote:
> 
> Hi Rob,
> 
>> On October 11, 2015 at 11:28 AM Robert Sanderson <azaroth42@gmail.com> wrote:
>> 
>> 
>> The approach that we have taken in the Web Annotation Working Group [1] (and
>> elsewhere) is to have an embedded resource with value, language and format
>> properties:
>> 
>> {
>>   "@type": "EmbeddedContent",
>>   "value": "<span>This is some <b>marked up</b> content.</span>",
>>   "language": "en",
>>   "format": "text/html"
>> }
>> 
>> As RDF 1.1 does not allow both language and format to be associated with a
>> literal value, this is the best that we could do.
>> 
>> Hope that helps,
>> 
> 
> Thanks for the input.
> Very relevant as we also need to deal with multilingual content.
> Did you consider to put the lang="en" attribute in the HTML?
> If so, what was the reason to go for chosen approach?
> 
> Brings up some interesting questions about if we might look at language-based
> content negotiation.

there is an Accept-Language header, but the metadata for that would have to be per document v/s per term as it operates at a different protocol level.

> Would be nice in theory, but not sure how widely this is supported.
> Also considering the translation processes, the different languages could well
> be based on different
> versions of the primary content, how to deal with this in a clean manner?

Accept-Version analogous to the Accept-Datetime which memento introduced, but with additional parameters beyond an atomic designator to apply to version metadata?

best regards, from berlin,
---
james anderson | james@dydra.com | http://dydra.com