BLD comments

Reading the frozen editors draft of 18th May 2008 I noticed a couple of 
mostly editorial points.

o Process question: can we have a Last Call spec parts of which are 
defined by non-Last Call documents? Specifically the definitions of 
Symbols Spaces (e.g. in 2.1) and Schemas for Externally Defined Terms 
(e.g. in 2.3) are in documents not currently included in the Last Call set.

** Section 2.4.

o The definition of a RIF rule does not permit an externally defined 
term as the consequent but does not exclude equality terms involving 
externals. Is that deliberate and OK?  Thus rules like:

   Forall ( External(fn:foo("1"^^xsd:int)) = "3"^^xsd:int) :- And() )

are legal?

** Section 2.5

o Suggest deleting sentence starting "Tool developers ...". The text 
makes it very clear that the EBNF is not a concrete syntax whereas the 
reference to "Tool developers" suggests otherwise.

** Section 3.4

s/In this case, we same that/In this case, we say that/

** Section 4

o Must define which version of XML RIF assumes. This affects which 
subset of unicode strings are legal.

** Section 4.1

s/XML serialization of the presentation syntax/XML serialization of RIF-BLD/

[I still don't understand the value of discussing the serialization of 
the conditional language separate from that of the rule language but 
that's old ground.]

o The reuse of the "upper"/"lower" role names for Member as well as 
Subclass is a little confusing. How about "instance"/"class"?

o In the explanation of the "type" attribute the example should not use 
CURIE notation. Suggest writing xsd out in full otherwise use an assumed 
entity declaration.

** Examples 4 - 6

o The reference to "compact URI prefixes" here is not appropriate and 
can mislead the reader into thinking they can use CURIE notation in the 
XML. Use entity declarations.

[If the above point is rejected for some reason then need to at least 
include rif and xsd in the list of compact URI prefixes.]

o At least one of the XML examples (and maybe all) should be a complete 
legal XML document, include *all* of the XML header lines and no pseudo 
macro notation.

** 7.1

o Suggest that attributes should be qualified so we use "rif:type" 
rather than plain "type". [I'm not an XML person so am prepared to be 
overruled on this.]

o In the definition of Const suggest the (xsd) type restriction for the 
(rif) type attribute should be "xsd:anyURI" rather than "xs:string".

Dave
-- 
Hewlett-Packard Limited
Registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England

Received on Sunday, 25 May 2008 16:08:35 UTC