Re: RDF tests errata

> On Jul 20, 2015, at 2:38 PM, Andy Seaborne <andy@apache.org> wrote:
> 
> On 19/07/15 19:19, Eric Prud'hommeaux wrote:
>> 
>> On Jul 19, 2015 12:53 PM, "Andy Seaborne" <andy@apache.org
>> <mailto:andy@apache.org>> wrote:
>> >
>> > On 19/07/15 16:18, Eric Prud'hommeaux wrote:
> ...
>> > Eric,
>> >
>> > There are two things:
>> >
>> > 1/ The tests as approved during the WG process
>> >    These were the state implementations reported on.
>> >
>> > Will there still be these frozen files (on w3.org? <http://w3.org?>)?
>> 
>> Yeah, but I think it's fine to stick them someplace obscure linked from
>> the live tests and the snapshot of the implementation report.

I think the most relevant information to developers now is what is the present state of the implementation report, not what was used historically to move to REC. While that needs to be preserved (and should be what’s linked from the RECs), it should be clear where the up-to-date implementation report is. Linking it from the test suite is good, and perhaps something like an Editor’s Draft link off of the version used for REC would be useful.

Some form of automation process, where developers could upload their EARL reports and have the consolidated implementation report pop-out automatically would be good; we can probably re-purpose Travis-CI for this purpose.

>> > 2/ A live evolving set of tests
>> >
>> > Is the governance going to be the same as document comments?
>> 
>> I suspect that the right level of governance is a community group. What
>> say you?
> 
> Mainly I was asking about the details of
> [[
> Per consensus in an INK Domain call, W3C would like to move tests and implementation reports to github
> ]]
> 
> A community group feels the right thing to do; IPR etc. It answers the practical matter of who own the github repo (W3C).
> 
> A question left is about decisions of that group.  What if any, standing do the changes have? Are they errata for the formal test suite? or a bunch of changes made by a self-selected group of volunteers?

Obviously, a Community Group can’t affect the status of released specifications directly, but it seems that the members of a Community Group will self-select to operate in the best interests of both the standard and the tests. As long as we make it clear that the test suite, and are solely the result of that CG, I think we’re clear. This may result in proposals for a PER; there is no process I’m aware of for making “official” updates to the test suite of the WG, once the WG is gone. This now becomes the test suite of the CG, which needs to be clarified. Of course, the work of the CG could potentially roll into that of a future WG.

Gregg

> (For SPARQL, W3C still officially controls and mediates additions to the SPARQL errata for example)

I’d like to see a similar process for SPARQL tests, which are in greater need of curation, IMO. Maybe a single RDF Test Platform CG could take this on for all existing RDF spects/test-suites.

Gregg

>> 
>> >         Andy
>> >
>> >
>> >
>> >
>> 
> 
>

Received on Monday, 20 July 2015 22:07:21 UTC