Re: JSON-LD skipping CR?

On 03/29/2013 10:38 AM, Ivan Herman wrote:
> On Mar 29, 2013, at 13:12 , Markus Lanthaler <markus.lanthaler@gmx.net> wrote:
>
>> On Friday, March 29, 2013 4:46 AM, Sandro Hawke wrote:
>>
>>> Have we talked about the possibility of skipping CR
>>> for JSON-LD?
>>>
>>> It has a lot of implementations and a test suite; if
>>> the implementations have passed the test suite by the end
>>> of LC, then CR isn't necessary.
>> We talked about this once but some people in the RDF WG (sorry can't
>> remember who it was), said that implementations from people working on the
>> spec do not count. I don't really buy that argument (it's true for most
>> specs IMHO) but we should definitely discuss that *now* to avoid bad
>> surprises later.
>>
> Hm. I do not buy that argument in general either.
>
> In RDFa Gregg and I were both working on the spec, and we were both working on our implementations. The two implementations are radically different (mine is in Python based on RDFLib, Gregg's is based on Ruby); I think I have not even seen, let alone run his implementation. During the discussions we both realized on numerous occasions that we did take different design choices internally, too. But both our implementations complete the full test suite; I regard that as a satisfactory outcome of the CR phase.
>
> I think the issue is mutual independence of these implementations. Not whether the implementer was part or not of the design.
>
> (In the case of RDFa we had other implementations, too, including from people who were not in the group. But that is besides the point.)

It was actually me who made this point, and I meant to be clear that 
this was an idea, not any kind of rule.    W3C process certainly doesn't 
require implementations to be from outside, but we're also on the edge 
of W3C process with doing most of this work in CG.   To me, the fact the 
the work happened without staff participation kind of raises the bar for 
the other quality tests.    In particular, I'm trying to avoid the 
situation where the 5 people in the room know that ofcourse you're 
supposed to do it like this, when the spec never says that.     It's 
great to have some people implement just from reading the spec, to help 
avoid that problem.  We can't always have that, but we at least usually 
have a staff contact in the room trying to help avoid this problem.

       -- Sandro


> Ivan
>
>
>> Cheers,
>> Markus
>>
>>
>> --
>> Markus Lanthaler
>> @markuslanthaler
>>
>>
>
> ----
> Ivan Herman, W3C Semantic Web Activity Lead
> Home: http://www.w3.org/People/Ivan/
> mobile: +31-641044153
> FOAF: http://www.ivan-herman.net/foaf.rdf
>
>
>
>
>
>

Received on Friday, 29 March 2013 15:06:44 UTC