Re: can subqueries be executed first in SPARQL?

good morning;

> On 2016-06-17, at 09:52, Axel Polleres <droxel@gmail.com> wrote:
> 
> Hi all,
> 
> interesting discussion...
> 
> First, let me say that I find the claim that something is "broken" here exaggerated.

well, while perhaps not “broken” in the sense of “inherently or evidently not functioning”, both hernández &co and patel-schneider have provided examples sufficient to support the claim that the definition constitutes a severe discontinuity.

if instead of

> The definition of EXISTS is broken so bad in SPARQL that it should be replaced with a completely different mechanism.

the claim had been, 

> The definition of EXISTS breaks with several central principles of the SPARQL evaluation model, such as the bottom-up nature of SPARQL query evaluation and rules for variable scope, in such basic ways and with so little motivation and so little explanation, that it leads to non-interoperable implementations and should be replaced.

would that be acceptable?
while the second is perhaps less polemic, it and its consequences are just as severe.

best regards, from berlin,
---
james anderson | james@dydra.com | http://dydra.com