Re: [web-annotation] Require dct:Text class for interpretation of literal bodies

> On 27 Apr 2015, at 16:18 , Jacob Jett <jjett2@illinois.edu> wrote:
> 
> This presents a slippery slope. Commenting isn't the only role the text literal may play. What happens if we have to expand this to encompass things like explanation, description, etc. Haven't we just made motivation a completely obsolete concept within the model?
> 
> Better would be if we could do away with the string literal bodies.

This came up several time in this thread… But I still believe literal bodies are important.

It is of course true that if an annotation is created by some process/program/plugin, then having an object for the literal body is not a big deal. However, there are genuine use cases when annotation is done by humans manually. The original use case for this came from the CSV on the Web Working Group: for CSV files (but, I would expect, for many different data files published on the Web) metadata is added by the human publisher of the data as an adjunct to the raw data itself. Such metadata should also provide hooks for annotations (the CSV Metadata, which is typically in JSON, provides a specific "notes" property for that purpose) and using the Annotation Model has great advantages. But if humans have to edit/provide this, I believe the ability to add a string literal body is essential. (If we do not provide this we can bet that people will do it anyway:-)

Ivan



> Alternately we might move the motivation to the body node?
> 
> I'm starting to think that the multiple bodies portion of the model presents a significant challenge to the utility of motivation. It's ambiguous what role each body plays in such an annotation because the motivations don't neatly pair up with the bodies. This was also a problem when we tried to develop a taxonomy of annotation types.
> 
> Regards,
> 
> Jacob
> 
> 
> 
> _____________________________________________________
> Jacob Jett
> Research Assistant
> Center for Informatics Research in Science and Scholarship
> The Graduate School of Library and Information Science
> University of Illinois at Urbana-Champaign
> 501 E. Daniel Street, MC-493, Champaign, IL 61820-6211 USA
> (217) 244-2164
> jjett2@illinois.edu
> 
> 
> On Mon, Apr 27, 2015 at 8:55 AM, Rob Sanderson via GitHub <sysbot+gh@w3.org> wrote:
> Currently the usage of a literal body is ambiguous as to whether the
> text is a comment or a tag, leaving implementers to guess based on
> motivations and heuristics.  This would make it explicit that a
> literal body is *always* a comment and never a tag.  The reason why
> motivation is insufficient is that there can be multiple bodies on an
> annotation, all could be literals, with a mixture of comments and
> tags:
> 
> ```
> {
>   "@type": "oa:Annotation",
>   "motivatedBy": ["oa:commenting", "oa:tagging"],
>   "hasBody" : ["comment", "tag"],
>   "hasTarget": "http://example.org/"
> }
> ```
> 
> It's an outcome from the thread:
> https://lists.w3.org/Archives/Public/public-annotation/2015Apr/0139.html
> 
> --
> GitHub Notif of comment by azaroth42
> See
> https://github.com/w3c/web-annotation/issues/31#issuecomment-96661471
> 
>