[Axiom-developer] Automated calculation of AXIOM bugs

[Vladimir Bondarenko]

Hi Martin, hi *,

Much thanks for your quick comment.

I expected that it is a bug but being right this moment a tyro wrt
AXIOM, in contrast with MMA, Maple and many other systems, at the
very beginning of the learning curve I need some help on side of
an AXIOM connoisseur.

If AXIOM developers would expect to get the maximal efficiency
from my diagnostic effort, it would be great also to hear, during
some period of time, a kind of explanation.

I mean, what is your decision procedure, that is your internal
work, your train of thought, your criterion/criteria when you
make the decision that this IS a bug?

Say, "it is a bug BECAUSE there is inconsistency in observable
behavior between the case A and the case B while the input data
are similar (or are of identical nature) for A and B" or kind
of it?


I am asking you because I am trying to build a discrimination
net for AXIOM bug analysis, that is I should learn how to make
the decisions correctly.

Currently, AXIOM seems to be very buggy, at least much more than
I had expected from reading J/S book. But having had a solid bug
list, hopefully, it would be easier to get insight about the nature
and the interplay and the evolution of the bugs, and prioritize the
bug manifestations in a way that fixing them would take the minimal
efforts.

Much thanks again.


Best,

Vladimir


MR> Of course it's a bug. Please file it!

MR> Martin

MR> Vladimir Bondarenko writes:
MR>  > Hello all,
MR>  > 
MR>  > Any comments on the following stuff?
MR>  > 
MR>  > 
MR>  > .......................................................
MR>  > -- behavior type # 1
MR>  > .......................................................
MR>  > 
MR>  > -> integrate(%i^log(z), z)
MR>  > 
MR>  >         log(%i)log(z)
MR>  >     z %e
MR>  >     -----------------
MR>  >        log(%i) + 1
MR>  > 
MR>  > -> integrate(f^log(z), z)
MR>  > 
MR>  >         log(f)log(z)
MR>  >     z %e
MR>  >     ----------------
MR>  >        log(f) + 1
MR>  > 
MR>  > -> integrate(sin^log(z), z)
MR>  > 
MR>  >         log(sin)log(z)
MR>  >     z %e
MR>  >     ------------------
MR>  >        log(sin) + 1
MR>  > 
MR>  > -> integrate(tan^log(z), z)
MR>  > 
MR>  >         log(tan)log(z)
MR>  >     z %e
MR>  >     ------------------
MR>  >        log(tan) + 1
MR>  > 
MR>  > -> integrate(sqrt^log(z), z)
MR>  > 
MR>  >         log(sqrt)log(z)
MR>  >     z %e
MR>  >     -------------------
MR>  >        log(sqrt) + 1
MR>  > 
MR>  > -> integrate(atan^log(z), z)
MR>  > 
MR>  >         log(atan)log(z)
MR>  >     z %e
MR>  >     -------------------
MR>  >        log(atan) + 1
MR>  > 
MR>  > -> integrate(asinh^log(z), z)
MR>  > 
MR>  >         log(asinh)log(z)
MR>  >     z %e
MR>  >     --------------------
MR>  >        log(asinh) + 1
MR>  > .......................................................
MR>  > 
MR>  > 
MR>  > BUT
MR>  > 
MR>  > 
MR>  > .......................................................
MR>  > -- behavior type # 2
MR>  > .......................................................
MR>  > 
MR>  > -> integrate(exp^log(z), z)
MR>  > 
MR>  >    >> System error:
MR>  >    Cannot take first of an empty list
MR>  > 
MR>  > -> integrate(log^log(z), z)
MR>  > 
MR>  >    >> System error:
MR>  >    Cannot take first of an empty list
MR>  > 
MR>  > .......................................................
MR>  > 
MR>  > 
MR>  > Should this be considered as a bug or a feature?
MR>  >    
MR>  > 
MR>  > Best,
MR>  > 
MR>  > Vladimir
MR>  > 
MR>  > 
MR>  > 
MR>  > _______________________________________________
MR>  > Axiom-developer mailing list
MR>  > address@hidden
MR>  > http://lists.nongnu.org/mailman/listinfo/axiom-developer