On 2015-12-02 00:23, Volker Braun wrote:
IMHO undecidability has exception semantics so really there is only one
correct solution, namely raise. Otherwise you get
sage: foo() == bar() # haha both Unknown
True
But we can override __eq__ (or __richcmp__), so it could be
sage: foo() == bar()
IMHO undecidability has exception semantics so really there is only one
correct solution, namely raise. Otherwise you get
sage: foo() == bar() # haha both Unknown
True
On Tuesday, December 1, 2015 at 11:20:53 PM UTC+1, vdelecroix wrote:
>
> Unknown is not broken. It just does not behave like
An error is perhaps the most appropriate but not very user friendly.
None is better but a bit disappointing
sage: my_equation.has_solution() # haha answer is None
sage:
And moreover has the same "boolean" behavior as Unknown
sage: None or False
False
sage: False or None
sage: not None
False
Its quite terrible, in Python you can either raise an exception or maybe
return None if you can't decide on a boolean return value. The only
redeeming quality of Unknown is that it documents that it doesn't work,
horray. IMHO we should deprecate it.
On Wednesday, November 25, 2015 at 5:56:02
