thought experiment: Oracle will not object to *every* substantive change to NTriple

If I understood correctly what Zhe was saying in the telecon today, it 
was that the position he was being asked to represent for Oracle was 
that any substantive change to NTriple is unacceptable and would raise a 
formal objection.

The purpose of this message is to demonstrate that
a) this is not actually the case
b) to suggest to Zhe that the role of representing a W3C member in a WG 
is a two-way communication process, and not a one-way process - and that 
this is most critical in areas where there is greatest business concern


So, as a thought-experiment, which the chairs may choose to actually 
enact, but I am not particularly advocating, here is a proposed 
substantive change to NTriple.


PROPOSAL: That the specification for ASCII N-Triples in the RDF Test 
Cases document be modified by deleting the following text:
[[
*NOTE*: N-Triples is an RDF syntax for expressing RDF test cases and 
defining the correspondence between RDF/XML and the RDF abstract syntax. 
RDF/XML [RDF-SYNTAX 
<http://www.w3.org/TR/rdf-testcases/#ref-rdf-syntax>] is the recommended 
syntax for applications to exchange RDF information.
]]


Rationale: some W3C members, most notable Oracle, and also TopQuadrant, 
already use N-Triples to exchange RDF information, this change reflects 
that the WG will have adequately addressed weaknesses in the 2004 spec 
that motivated this text limiting the scope of N-Triples.

Given Zhe's statement today, if I have understood correctly, Oracle 
would formally object to this change - since it is a substantive change 
to N-Triple. Since this is incoherent, I believe I have demonstrated 
that Oracle will not object to each and every substantive change to NTriple.

====

So having established that Zhe misstated the reality of the Oracle 
position, I would like to suggest that the role of representative in 
such cases is:
a) to understand the actual business issues of the member that one is 
representing
b) to understand the technical content of proposed changes to actual or 
de facto standards or older ways of operating
c) to assess the impact of such changes on the business concerns
d) to articulate that material impact to both the WG and one's 
colleagues [the two way communication]
e) if the impact is bad then that is normally a compelling argument to 
the WG
f) in the case where the impact is bad, but the WG does not find it 
compelling, then one may need to reluctantly proposed to your immediate 
colleagues and AC rep that a formal objection is in order


In particular, I think I have heard several WG participants say that 
they wish to ensure that older data that conforms with the spirit of the 
2004 Ntriples spec will still conform with the updated spec, so that 
older data can be read by newer software. (I think I have heard some 
anxiety that that document is actually broken - in the sense that the 
grammar technically does not work). A position that 'no change is 
acceptable' (i.e. in addition newer data can still be read by older 
software) needs to be accompanied by identifying which software we are 
talking about, and why such a change will cause actual business 
problems, and an assessment of how bad these problems are. I think most 
of us are use to this issue with a wide range of software products and 
realize that if, for example, we want to send a MS Word document to 
someone who only has older software, that we have to save the document 
in the older format.

====

While I was pretty insistent on the call that the note above is there to 
indicate that the older WG was of the opinion that change would be 
needed before recommending NTriples as an exchange syntax, I would not 
like that to be misinterpreted as an unwillingness to listen to 
reasonably articulated concerns about business impact of specific changes.

Jeremy

Received on Wednesday, 7 March 2012 19:05:20 UTC