Re: lang tags on literals

Right, this is one of the reasons why we didn't have literal bodies in the
CG draft -- it causes exactly this confusion by having two different ways
to represent the same requirement of an embedded literal body.  Also, the
confusion as to when to use language and when to use @language (and value /
@value)

Thus the model tries to limit the damage by severely restricting the
situations and ways in which the literal body can be used, including no
datatypes, no language tags, no other properties, no role, and so forth.

Especially given that RDF literals cannot have both language and format, we
should promote a single solution that meets the requirements.

R


On Mon, Sep 14, 2015 at 1:59 AM, Ivan Herman <ivan@w3.org> wrote:

> The problem is that this requirement (even if it is a SHOULD) would create
> problems on the JSON-LD representation, which does not let you use a simple
> syntax for the language (as turtle does). Ie,
>
> "body" : "simple textual body"
>
> would become
>
> "body" : {
>         "text" : "simple textual body",
>         "@language" : "en"
> }
>
> which is quite convoluted. And I am not sure how I would explain to a
> JSON-LD user why adding that "@language" term if "dc:language" is also used.
>
> Ivan
>
> > On 14 Sep 2015, at 10:20 , Vladimir Alexiev <
> vladimir.alexiev@ontotext.com> wrote:
> >
> > http://www.w3.org/TR/annotation-model/#simple-textual-body and
> > http://www.w3.org/TR/annotation-model/#embedded-textual-body
> > don't avail themselves of the opportunity to use lang tags on the text
> > literal, and instead use dc:language.
> >
> > Using dc:language is a good practice, especially for resources that are
> not
> > a simple text string.
> > But I think that for strings, the language should be repeated as a lang
> tag
> > for consistency & completeness.
> > (This approach is adopted by Getty LOD, e.g. see
> > http://vocab.getty.edu/doc/#Language_Dual_URLs)
> >
> > So for http://www.w3.org/TR/annotation-model/#embedded-textual-body I
> > propose:
> > - after the phrase "additional properties such as dc:format and
> dc:language
> > should be given if known"
> > - add this sentence: "If the language is known, it should also be
> provided
> > as a language tag on the string literal"
> > - and add lang tag in the example:
> >   rdf:value "content"@en ;
> >
> > For http://www.w3.org/TR/annotation-model/#simple-textual-body, I would
> > allow it to be used with a lang tag,
> > which will extend this economical representation also for this case, but
> I
> > don't feel strongly about it.
> >
> > Cheers!
> >
> >
>
>
> ----
> 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
>
>
>
>
>


-- 
Rob Sanderson
Information Standards Advocate
Digital Library Systems and Services
Stanford, CA 94305