Re: on validation reports

On 9/12/2016 3:59, Peter F. Patel-Schneider wrote:
> If I was creating an issue related to this observation I would create one like
> this:
> The SHACL specification is much too loose in its description of validation
> reports.  Conforming implementations can produce widely varying validation
> reports for the same validation.  This variance imposes a much too difficult
> burden on applications that process validation reports.

We had intentionally designed this for flexibility, even included an 
"abstract" superclass for all validation results, encouraging people to 
create different kinds. This may include subclasses of 
sh:ValidationResult. So I would consider the current design a feature.

On the process question, we could indeed now make this a formal ISSUE, 
present it to the WG and spend meeting cycles on this. From your time in 
the WG you do remember how precious that weekly meeting time is, and 
that on many topics very little progress is made by emails alone. I am 
sure you are also aware that the WG will be ending in the middle of next 
year, and we need to go through various phases to fulfill our formal 
reviewing obligations. It is now unclear to me how we can structure the 
process so that we can realistically achieve the charter, but I guess 
we'll know more next week. Until then my apologies if I am not yet 
turning your comment into a formal issue. Please hold the line.

Thanks,
Holger

Received on Friday, 9 December 2016 00:41:18 UTC