Re: SHACL complexity

Hi all,

I follow the reasoning of the WG, makes sense.  That said, an informal
Group Note or something even less official would definitely be a valuable
resource. Something along the lines of implementation guidelines and
best-practices. Not high priority at the moment though

Best,

Miel

Op ma 17 mrt 2025 om 22:46 schreef Nicholas Car <nick@kurrawong.ai>:

> Hi Valdimir,
>
> The discussion in the WG meeting yesterday seemed to indicate that many
> people were quite happy with the idea of defining profiles of SHACL - like
> profiles of OWL, with different complexity expectations or, more
> importantly, with different expectations about the possibility of
> computational issues like infinite recursion.
>
> There was, however, pushback on the idea of working out and providing
> computational effort per profile or per feature.
>
> If the standard were to investigate computational effort, this might bias
> certain forms of implementation that strove for speed via engineering
> solutions such as parallelisation, which isn't really something the
> conceptual standard should look in to.
>
> Cheers, Nick
>
> On Wednesday, 12 March 2025 at 17:15, Vladimir Alexiev <
> vladimir.alexiev@graphwise.ai> wrote:
>
> It would be nice for each new feature (or more realistically a bundle of
> features, i.e. Profile) to have some idea about its implementation and
> execution complexity.
>
> I'd like to take 20-30 min from the next Monday meeting to discuss these
> issues:
> https://github.com/w3c/data-shapes/issues/321
> https://github.com/w3c/data-shapes/issues/242
>
> Is this realistic, and would people like to comment in these issues
> beforehand?
>
> Cheers!
>
>
>