W3C home > Mailing lists > Public > public-cwm-talk@w3.org > January to March 2005

Re: N3 media-type in SPARQL protocol

From: Graham Klyne <GK-lists@ninebynine.org>
Date: Tue, 18 Jan 2005 11:26:50 +0000
Message-Id: <>
To: Dan Brickley <danbri@w3.org>
Cc: public-cwm-talk@w3.org

Hmmm... adding to unregistered MIME types in-the-wild?  If this were a URI 
scheme, I think some folks here would complain.  I suggest submitting a 
MIME-type registration if going this way.  Also, since N3 isn't intended to 
be a widespread interchange format (is it?), wouldn't application/vnd.w3.n3 
be better?

By the way, from my recollection of the discussions, the adoption of 
<foo>+xml for MIME types was a one-off for XML, recognizing XML's 
preeminent role as a meta-format for defining new content types.  I don't 
think n3 can make any such claim, so the suggestive use of rdf+n3 would be 
misleading in that respect.  I suggest sticking to n3.


At 15:32 16/01/05 -0500, Dan Brickley wrote:

>Hi. Nice to see
>out and about :)
>Just one brief and boring comment now while I remember:
>"text/n3" should probably be "application/rdf+n3"
>or "application/n3". Googling for these, it seems both
>are in various codebases. Perhaps having a conneg-happy
>protocol spec out there will motivate some concensus
>(and a media type registration) in the N3 scene? CC:'ing timbl...
>(text/n3 is worse due to UTF-8 vs ASCII issues). Ah,
>Tim says "text/rdf+n3" is also in use. Hmm. I'm going to hit
>"send" on this mail anyways, to flag the issue!
>ps. typos in 2.2:
>"Accept-Chareset" -> "Accept-Charset"
>  "prefix foaf:  <http://xmlns.com/foaf/0.1/%gt; ."
>         ... s/%gt/&gt/ I think
>also suggest "Host: my.example"
>         -> "Host: my.example.org"
>although http://www.rfc-editor.org/rfc/rfc2606.txt
>allows .example as a fictional TLD

Graham Klyne
For email:
Received on Tuesday, 18 January 2005 11:31:29 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 20:01:04 UTC