Re: test case results expressed in ResultSet terms

EricP wrote:
>> I was also trying to run your "nested optional arcs" test case
>> and had to make some corrections in our implementation so that
>> in fact only those 4 results are returned (also using a design
>> that allows result interpretation as it refers to the query).
>> So querying the facts given in
>> with filter
>> results in
> Yeah, the trick (for me, at least) was to make sure I NULL'd
> all the otherwise unbound variables in any optional part that
> was not satisfied. I think this is pretty easy to define so
> I'm not to worried about this as a requirement from that
> perspective. However, it is worth noting that introducing
> optionals does mean we have an ordered set of constraints.

when we have found all solutions coming through the filter
we put them in a new engine and take out the unique triples
using the unification method, but indeed, you are right, we
have to build something (plan to do that later..) to do an
automatic clause and premis reordering to make it order

> I'd like to experiment with Euler(sharp?). I guessed a CVSROOT:
>   cvs login
> but didn't guess a module name. I tried Euler, euler, EulerSharp,
> and eulersharp.

Oops.. will have to write some more documentation..
(ther's now some at

You could checkout module 2004/01swap
$ cvs -z3 co 
and then run
$ java -cp 2004/01swap/classes euler.EulerRunner --nope --think \ \
for the java version, or
$ 2004/01swap/bin/Release/Euler --nope --think \ \
for the .NET version

(yesterday I only checked in the sources, and today I've
checked in also the classes and bin, so it could take a
while before can get them)

Jos De Roo, AGFA

Received on Sunday, 27 June 2004 07:05:31 UTC