> I suppose it would be more helpful for it to signal an error.
Well, not if that's a well-understood convention, as Rich indicates.
It seems as good a convention as any, to be honest, and it makes
sense.
>> (%i1) assume(x<-2);
>> (%i4) limit(floor(x),x,0);
>> (%o4) limit floor(x)
>> x -> 0
>
> The limit here means that -1 < x < 1. But your assumption was x<-2. Faced
> with this contradiction in user input, Maxima returns a noun form, meaning
> that it can't calculate a result. So it is in some sense a "correct"
> result.
But then why, under the same assumption, do I get:
(%i3) limit(x^2,x,0);
(%o3) 0
(%i5) limit(sin(x)/x,x,0);
(%o5) 1
Should they not also remain unevaluated? That's why I'm confused.
On the upside, chasing down these things is making me much more
comfortable with Maxima syntax! And that is good, as a beginner.