[Pellet-users] bug report - Pellet 1.5.1.direct reasoning unsound while OK with DIG

[Raphael Coulonvaux]

  Hi,

I am using Protegé 3.4 build 125 and I am experiencing a problem when using
Pellet 1.5.1. direct reasoning.
The problem does not occur while using Pellet 1.5.1 through DIG but this is
not a usable workaround because of the DIG too strong limitations (e.g.
hasValue, cardinality, ... not supported).
The problem is the wrong computation of inferred types with a class being
equivalent to the union of two other classes.

You will find my (very small) test ontology in attachment but here is the
detailed description of the problem :

I have four classes :
- Class_1,
- Class_2,
- Class_3 and
- Class_4.

I have only one instance in my test ontology : individual_1 being asserted
as instance of Class_1.

Class_2 is equivalent to Class_1 and not surprisingly, individual_1 is
inferred as being an instance of Class_2, Pellet 1.5.1. being used directly
or through DIG - no problemo so far.

Class_4 is equivalent to the UNION of Class_1 and Class_3 and therefore,
individual_1 should be inferred as being an instance of Class_4.
Unfortunately (hélas, c'est là qu'est l'os ;-), this INFERENCE OCCURS ONLY
WHEN asking Pellet 1.5.1. to compute inferred type THROUGH DIG : the same
computation of inferred types using Pellet 1.5.1. in DIRECT mode does NOT
infer individual_1 has being an instance of Class_4 !!!

Raphaël Coulonvaux
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.owldl.com/pipermail/pellet-users/attachments/20080213/f3713030/attachment.htm 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: test_union.owl
Type: application/octet-stream
Size: 1104 bytes
Desc: not available
Url : http://lists.owldl.com/pipermail/pellet-users/attachments/20080213/f3713030/attachment.obj