Support for rdf:dirLangString and format versioning #54
Conversation
|
@gkellogg why the rdf:dirLangString section is non-normative? |
Those sections are all explanatory. Normative behavior is defined in the events and processing rules. |
gkellogg
force-pushed
the
dirLangString
branch
from
945d0aa to
e0b0a44
Compare
| class="termref"><span | ||
| class="arrow">·</span>literal-direction<span | ||
| class="arrow">·</span></a> is the empty string, | ||
| the value is the concatenation of """ (1 double quote), |
There was a problem hiding this comment.
""" (1 double quote) gets rendered as """ (1 double quote). There are a number of possible ways to better present this (which should be applied to all similar instances). Probably the simplest would be <code>"</code> (one double-quote).
There was a problem hiding this comment.
I'm not happy with this general way of emitting triples (see #55), and more could be done on character presentation, similar to how it's done in Turtle. This just adapts the existing pattern; I think another pass is needed separately to clean up these things. Just changing one of them because it happened to be touched by this PR doesn't really address the problem.
| @@ -1140,7 +1199,10 @@ <h2>RDF/XML Internet Media Type, File Extension, and Macintosh File Type</h2> | |||
| systems be given a file type of <code>"rdf "</code> | |||
| (all lowercase, with a space character as the fourth letter).</p> | |||
|
|
|||
| <p>The `application/rdf+xml` media type has been registered at IANA as [[RFC3870]].</p> | |||
| <p>Features introduced in RDF 1.2 require a version announcement in the | |||
There was a problem hiding this comment.
That's an independent decision?
I think "required" is a burden because to stream the whole data must be seen. Therefore, I think the version should be optional.
The majority of RDF 1.2 data will be legal RDF 1.1 .
There was a problem hiding this comment.
I agree with @afs. This should be part of the more general conversation on how we handle versioning in concrete syntaxes.
There was a problem hiding this comment.
This is just the PR which introduces versioning.
There was a problem hiding this comment.
The wording says that using 1.2 features requires a version announcement, not that any use of RDF/XML requires a version announcement.
spec/index.html
Outdated
| @@ -1867,9 +1980,12 @@ <h3>Plain Literal Event</h3> | |||
|
|
|||
| <p class="note">RDF/XML plain literals are in RDF 1.2 treated as | |||
There was a problem hiding this comment.
"plain literals" went away with RDF 1.1.
We can define "plain literal" for the purposes of RDF/XML as "no given rdf:datatype".
It does not have to be a full <dfn>.
Digression 😀
It may be useful to define "string literal" in RDF Concepts with alt "RDF string literal".
SPARQL has one because the grouping of xsd:string, rdf:langString and now rdf:dirLangString comes up repeated.
There was a problem hiding this comment.
I updated the language, although it may not flow quite as well. Suggestions are welcome.
| @@ -1140,7 +1199,10 @@ <h2>RDF/XML Internet Media Type, File Extension, and Macintosh File Type</h2> | |||
| systems be given a file type of <code>"rdf "</code> | |||
| (all lowercase, with a space character as the fourth letter).</p> | |||
|
|
|||
| <p>The `application/rdf+xml` media type has been registered at IANA as [[RFC3870]].</p> | |||
| <p>Features introduced in RDF 1.2 require a version announcement in the | |||
There was a problem hiding this comment.
I agree with @afs. This should be part of the more general conversation on how we handle versioning in concrete syntaxes.
| <code>xsd:string</code> (when no language tag is present), | ||
| as a literal with datatype <code>rdf:langString</code> (when | ||
| a language tag is present and a direction is not present), | ||
| or as a literal with datatype <code>rdf:dirLangString</code> | ||
| (when both a language tag and a direction are present). |
There was a problem hiding this comment.
That does not tell us what to do if we have a direction but no language tag...
There was a problem hiding this comment.
The result would be illegal according to Concepts, but the triple generation wording in this spec is archaic and should be taken up separately.
pchampin
changed the title
|
I edited the title of this issue to reflect the actual content: |
gkellogg
changed the title
Co-authored-by: Ted Thibodeau Jr <tthibodeau@openlinksw.com>
Co-authored-by: Ted Thibodeau Jr <tthibodeau@openlinksw.com>
Co-authored-by: Andy Seaborne <andy@apache.org>
…k, and re-number examples to be consistent with the added example for dirLangString.
Co-authored-by: Ted Thibodeau Jr <tthibodeau@openlinksw.com>
gkellogg
force-pushed
the
dirLangString
branch
from
89de473 to
82738b6
Compare
Still a Work in Progress, more work needed on the event definitions.
Fixes #50.
Preview | Diff