W3C home > Mailing lists > Public > www-rdf-comments@w3.org > July to September 2003

Re: status of comments pfps-02 pfps-03 pfps-04 pfps-05 pfps-06

From: pat hayes <phayes@ihmc.us>
Date: Thu, 25 Sep 2003 13:43:24 -0500
Message-Id: <p06001f12bb98c1f65e0a@[]>
To: "Peter F. Patel-Schneider" <pfps@research.bell-labs.com>
Cc: www-rdf-comments@w3.org, Brian_McBride <bwm@hplb.hpl.hp.com>

>From: Brian McBride <bwm@hplb.hpl.hp.com>
>Subject: status of comments pfps-02 pfps-03 pfps-04 pfps-05 pfps-06
>Date: Thu, 25 Sep 2003 14:19:52 +0100
>>  Peter,
>>  I've been reviewing the results of RDFCore's last call process and note
>>  that my status list is showing that you have not accepted some of the
>>  WG's responses on comments relating to the semantics document:
>>  http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-02
>>  http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-03
>>  http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-04
>>  http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-05
>>  http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-06
>>  Pat tells me that some progress has been made on at least some of these.
>>  Accordingly, could you please state whether any of these issues are
>>  acceptably resolved in the current RDFCore published WD:
>>  http://www.w3.org/TR/2003/WD-rdf-mt-20030905/
>>  or in subsequent discussion.
>>  Brian
>-> http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-02
>This comment has not been adequately addressed.  I believe that there
>has never been a translation from RDF to LBase that even resulted in
>syntactically correct LBase.  The most-recent communications on this topic
>that I have received from the RDF Core Working Group have been
>The translation from RDF to LBase still results in incorrect LBase syntax.
>For example, the RDF untyped literal "\" is translated into '\', which will
>result in syntax problems in LBase.

That is correct, and reflects a slip-up made in 
the last-minute editing. The backslash character 
should be self-escaping. I will correct this in 
the Lbase document by requiring this, changing 
the example in section 2.1.2 to read:
..so that '\'A\\\'' denotes the string 'A\'
and adding the requirement:
...with any internal occurrences of ''' or '\' prefixed by '\'
to the Lbase translation (also in the translation in the semantics appendix.)

>  Translation of RDF URI references that
>contain non-ASCII characters are also problematic, I believe.

This level of lexicographic detail isn't really 
relevant to the goals of the Lbase translation. I 
have therefore modified the text of the appendix 
by inserting the following comment:

"It is possible for the translation given here to 
produce syntactically illegal or ambiguous Lbase, 
eg if a URI reference begins with the symbol '('. 
An exact translation for machine use would need 
to perform some more detailed lexical analysis 
and use a more sophisticated encoding of URI 
reference syntax in order to avoid these issues 
in all possible cases. "

and modified the Lbase document (section 2.1.2, end of first para) thus:

"A name may be any string of unicode characters 
not starting with the characters ')','(', '\', 
'?','<' or ''' , and containing no whitespace 
characters, or any URI reference. In order to 
avoid syntactic ambiguity in all possible cases, 
for example when a URI begins with one of the 
reserved characters, it may be necessary to use 
more sophisticated renderings of URI names as 
character strings: we will ignore such 
complexities in this document."

Given this cop-out, I have also removed the 
hastily-added <-> syntax for encoding strings 
which was added to the recent version.  The 
consequences of this have not been thought out in 
enough detail.

>There may be
>other problems here - I have not checked thoroughly.
>I also note that the LBase syntax itself is ambiguous, which I have already
>pointed out.  I have not received a satisfactory response to this

I believe that this refers to a bug in an earlier 
version (which did not use escaping of quotation 
marks in strings), pointed out in your message


If so, this has now been corrected, cf. above. I 
see however, on looking at the archive,  that 
this was not formally acknowledged in an email 
response, an omission for which I apologize. 
Please take this message to be the official 
notification of the change.

>My messages of 31 August 2003 and 5 September 2003
>are relevant to this issue.  The latter has not received a response.
>-> http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-03
>This comment has not been satisfactorily resolved.  There is still no
>indication of the use the translation from RDF to LBase serves in the RDF

I think we have already agreed to disagree about 
this matter. Your comment is noted.

>-> http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-04
>This comment has not been satisfactorily resolved.  There has been no
>official communication on this issue since my message of 14 August 2003

The original comment refers to language tags in 
XML documents. Such language tags were removed 
from XML literals some time ago, so this comment 
is now moot. The communication of 14 August seems 
to be devoid of content other than as an 
expression of a subjective state; I do not see 
how there could be an appropriate official 
response to it.

>-> http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-05
>This comment has not been satisfactorily resolved.  There is still no
>syntactic characterization of entailment in 
>RDFS.  My message of 14 August 2003
>details this and other concerns I have with the RDFS entailment rules.
>Some of the minor issues in this message have been somewhat resolved in
>later RDF Core Working Group Working Drafts, but this does not resolve the
>major concern I have in this area.

I do not accept this comment, in the form that it 
has been understood in the most recent message 
traffic. The original comment, referred to as 
pfps-05, which was clearly different in content, 
was accepted and has, I think, been adequately 

It is incorrect to assert that there is no 
syntactic characterization of RDFS entailment. 
The text gives one explicitly, in a style which 
has been unchanged except in details since the 
very first draft of the document and which has 
been used as the basis for working 
implementations. I have explained and amplified 
this point in several email messages, for example:


>-> http://www.w3.org/2001/sw/RDFCore/20030123-issues/#pfps-06
>This comment has not been satisfactorily addressed. 
>I have not received any official communication from the RDF Core Working
>Group after my message of 23 July 2003
>on this topic.  The current versions of the RDF Core Working Group Working
>Drafts may have adequately resolved this issue, but there has been no
>indication that I can find to this effect.

The original comment has been addressed, I 
believe adequately: in particular, malformed 
typed literals are excluded from LV.

The above-cited message of 23 July referred to a 
phrase in the wording of the change log, and did 
not address the actual text of the revised 
document.  In response, after pointing this out, 
I revised the wording of the change log so as not 
to give a misleading impression: this change was 
noted in
and is reflected in the text of the document cited by Brian.

The relevant parts of the actual document text 
have not been changed for some time, and await 
review and comment.


IHMC	(850)434 8903 or (650)494 3973   home
40 South Alcaniz St.	(850)202 4416   office
Pensacola			(850)202 4440   fax
FL 32501			(850)291 0667    cell
phayes@ihmc.us       http://www.ihmc.us/users/phayes
Received on Thursday, 25 September 2003 14:43:33 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:15:21 UTC