2015-05-27 14:15 GMT+02:00 Sean P. DeNigris <sean@clipperadams.com>:

Nicolas Cellier wrote
> we want to have this property a=b & (a=c) ==> (b=c).

If IUC it sounds more inconsistent because
�� (5463/20) = (10926/40) "true"

What is less consistent is that these fractions are all different:

fractions := (101 to: 200) collect:�� [:n | 1<<n+1/(1<<n)].

[fractions asSet size = 100] assert.

But they are all "equal" to the same float:

[fractions allSatisfy: [:f | f asFloat = 1.0]] assert.

[(fractions collect: #asFloat as: Set) size = 1] assert.

Don't you see a problem now?

-----
Cheers,
Sean
--
View this message in context: http://forum.world.st/Inconsistent-number-comparison-tp4828795p4828890.html

Sent from the Pharo Smalltalk Developers mailing list archive at Nabble.com.