W3C home > Mailing lists > Public > www-rdf-interest@w3.org > September 2001

Notation3 woes

From: Bjoern Hoehrmann <derhoermi@gmx.net>
Date: Tue, 18 Sep 2001 17:32:15 +0200
To: www-rdf-interest@w3.org
Message-ID: <h8neqtg0tirt61r4ns07ma01gssnd512vd@4ax.com>
Hi,

   I like non-XML syntax since XML is almost ever hard to read, thus I
like Notation3, but I'm not happy with available materials on Notation3.
The authoritative source seems to be [N3], so my comments will be based
on this document. 

First I like to encourage the RDF WG to take Notation3 in the W3C
standards track ending up with a Technical Report on Notation3; this
should be either a Note or a recommendation. In general, I prefer
recommendations, but a Note may fit better here. To have a normative
specification for Notation3 would ease dependencies on Notation3 for
other W3C Technical Reports, e.g. the recently published RDF Test Cases
Working draft could reference this document in favour to define itself a
grammar for N-Triples.

I will list here what things I like to have addressed. These are general
Notation3 issues, they should be addressed even if Notation3 won't
become a W3C specification.

Design Rational:

  I'm currently no RDF expert, otherwise this might be rather obvious,
but I didn't find out, whether Notation3 and RDF can be converted into
each other one by one without loss of information. Is this the case and
if not, what limitations has either of both?

Character Encoding:

  There is no means provided by Notation3 do define the character
encoding of the document. This is a serious issue, since Notation3 is by
no means limited to US-ASCII, at least [N3] doesn't say so and I
certainly may want to define something like

  :Björn :name "Björn"

The N-Triples as used by the Test Cases WD require me to use \(u|U)
escapes, but

  :Bj\u00F6rn :name "Bj\u00f6rn"

is certainly not very easy to write. I propose to add an @encoding rule,
that has to appear as the very first token in a Notation3 document, that
specifies a IANA character set name, e.g.

  @encoding iso-8859-1.

URI encoding:

  It's likely that Notation3 documents use Characters not literally
allowed in URIs but they may have to be mapped to URIs. Section 8 of
[CHARMOD] defines what specifications must define for escaping non-ASCII
characters in URIs, this should be stated for Notation3 (and RDF!)
aswell.

MIME Type:

   Notation3 documents should be identifiable with an own MIME type,
this would require the registration of such a MIME type. I remember
Aaron Swartz has written up an Internet Draft for a general purpose RDF
MIME type, but It hasn't been submitted to IETF yet. Maybe this document
can be extended to cover also text/notation3 (or whatever MIME type is
appropriate). Text/plain isn't appropriate, see section 3 and 4.1 of RFC
2046. Aaron: Why didn't you submit your document yet?

Syntax and Grammar:

   The BNF in [N3] is a mess. I can't make any clue out of the comments
used there and certain tokens lack of definitions. Some of them should
be specified by reference to other syntax definitions, e.g. URIs and
QNames. A formal grammar should use the EBNF notation used in XML 1.0.
It seems that there are certain things still work in progress (that's
what I read out of those questions marks and the @@ markers in the
grammar), this makes it very hard to write a parser for Notation3,
doesn't it?

[N3]      - http://www.w3.org/DesignIssues/Notation3
[CHARMOD] - http://www.w3.org/TR/charmod/

regards,
-- 
Björn Höhrmann { mailto:bjoern@hoehrmann.de } http://www.bjoernsworld.de
am Badedeich 7 } Telefon: +49(0)4667/981028 { http://bjoern.hoehrmann.de
25899 Dagebüll { PGP Pub. KeyID: 0xA4357E78 } http://www.learn.to/quote/
Received on Tuesday, 18 September 2001 11:33:23 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:51 GMT