Re: Suggested charter changes

Hi Jeremy, thanks for your attention on this. I expect to synthesize a
list of proposals along with their proponents and detractors this
weekend. In the mean time, a proposed change below:

* Jeremy J Carroll <jjc@syapse.com> [2014-08-07 06:49-0700]
> In-line +1, -1s … to proposal immediately above
> On Aug 6, 2014, at 1:16 PM, David Booth <david@dbooth.org> wrote:
> > 
> > SUGGESTED SUBSTANTIVE CHANGES:
> > 
> > 1. In Section 4 (Deliverables), change:
> > [[
> >  5. Test Suite and/or Validator: to help ensure interoperability
> >  and correct implementation. The group will chose the form of
> >  this deliverable, such as a git repository.
> > ]]
> > to:
> > [[
> >  5. Test Suite, to help ensure interoperability and correct
> >  implementation. The group will chose the location of this
> >  deliverable, such as a git repository.
> > 
> >   6. (OPTIONAL) Reference Validator: a reference implementation
> >  that may be used for non-production purposes to test validation
> >  rules against test data.

6. (OPTIONAL) Validator: a potentially partial implementation that
demonstrates the technology by testing the conformance of validation
rules as well as verifying instance data against those rules.
 
> > ]]

• removed "reference" because real "reference implementations" can
  have normative bugs.
• added "potentially partial" because the language may have protocol,
  e.g. downloading value sets à la Resource Shapes, that we don't want
  to run on a public endpoint.

• differentiated verifying the expression of the rules from adherence
  of data to those rules.


> > REASON: A test suite is essential; a reference validator would be nice but is not essential.  Also, the previous word was completely unclear able what was meant by "Validator".
> > 
> 
> +1
> 
> > 2. Change "The WG MAY produce a Recommendation for graph normalization." to something like: "OPTIONAL - A Recommendation for normalization/canonicalization of RDF graphs and RDF datasets that are serialized in N-Triples and N-Quads."
> > 
> > REASON: Canonicalization needs to be relative to a serialization in order to be most useful.  Otherwise "canonicalized" RDF may be serialized in multiple ways, and still could not be usefully compared for regression testing or other purposes.
> 
> -1: I think I need to produce my own proposal here; in my view David's wording is slightly worse

proposed in <http://lists.w3.org/Archives/Public/public-rdf-shapes/2014Aug/0105>
pfps wants normalization struck entirely.


> > 3. In Section 4 (Deliverables), change "OPTIONAL - Compact, human-readable syntax" to "Compact, human-readable syntax", i.e., make it required.
> > 
> > REASON: I think a compact, readable syntax is important.  Writing validation rules in RDF would be simpler than writing them in SPARQL, but still much more tedious than having a concise, compact syntax for expressing them.
> > 
> > 
> 
> +1: I believe Karen has made the case for this deliverable

There seemed to be critical mass behind this at the workshop, but I
expect TopQuadrant and potentially Clark and Parsia will argue against
this deliverable. If there is support, I suspect we'll have to
exlicitly state whether turtle is considered "human-readable".


> (all the editorial comments seemed good to me)
> 
> 

-- 
-ericP

office: +1.617.599.3509
mobile: +33.6.80.80.35.59

(eric@w3.org)
Feel free to forward this message to any list for any purpose other than
email address distribution.

There are subtle nuances encoded in font variation and clever layout
which can only be seen by printing this message on high-clay paper.

Received on Thursday, 7 August 2014 19:49:24 UTC