Re: reification test case

[...]

>>and even
>>
>>   <http://example.org/eg#s> <http://example.org/eg#p> _:s1 .
>>   _:s1 <http://example.org/eg#q> <http://example.org/eg#v> .
>>
>>is *not* entailing
>>
>>   _:s2 <http://example.org/eg#q> <http://example.org/eg#v> .
>
>That should be entailed, again by anyone's standards. It is always OK
>to say an existential statement again using a different bound
>variable, even inside the same graph.
>
>>
>>but
>>
>>   _:s1 <http://example.org/eg#s> <http://example.org/eg#p> .
>>   _:s1 <http://example.org/eg#q> <http://example.org/eg#v> .
>
>>is entailing
>>
>>   _:s2 <http://example.org/eg#q> <http://example.org/eg#v> .
>
>?? Why is this different from the previous case??

OK
it's good to go back to such examples

all we said in the beginning was that
  [any big map
   _:s1 <property> <value>]

does entail
  [_:s2 <property> <value>]

****************************************

but I must confess that we *don't* have
that always in Euler...

it is a long story, and I'm not sure
I will be able to explain it,
but it is important to have
a clear decision on that.

I assume that, while asserting
  aaa bbb _:ccc .
  _:ccc ddd eee .
we for sure assert the triple
  aaa bbb [ ddd eee ] .
but *not* the "contents" of the node [ ddd eee ]
(this is N3's [ ] and ako "tree")

this is btw a case of unasserted triples
(just like N3's { } set of triples)

this is just to say how things are in
our implementation, and it would be
quite straightforward to change it
and have extra statements like
  [ ddd eee ] .
for all branches of all trees
(so efficiency would go down)

I'm feeling better now that I have
confessed my sins against MT
and I beg for some help

--
Jos

Received on Wednesday, 6 February 2002 19:44:54 UTC