X #= 1.1*1.1 Evaluability

[jacobfriedman]

I'm trying to find an ISO system (free / non-sicstus) that can handle the problem of X #= 1.1*1.1.

The BNR system is useful in that it provides an answer, whereby GNU does not:

• If trying to use FD, which isn't ideal for this case:uncaught exception: error(type_error(fd_evaluable,1.1000000000000001),(#=)/2)
• 1.21 =:= 1.1*1.1. (yields 'no').

What should be done about this?

Thank you - and I hear v2 may be coming out soon ;)

[didoudiaz]

What you need is a solver over reals (GNU Prolog provides a solver over finite integers). You can look at the clpr library (provided in SWI Prolog)

[jacobfriedman]

Thank you! I'm looking at BNR- very interesting. Take care.