Re: Quality assurance

jacob navia wrote:

JB> What is interesting for me, is that this bugs are being
JB> reported now, not before.

It would be a challenge to test Mathemathica 6 and Maple 11
prior they are released.

JN> But it is startling that bug reporting was never done here

It is really difficult to explain why you think so.

Being a CAS customer, we report about defects in Maple and
Mathematica over many years, since 2002 on, both to the
manufacturers and publicly. Just search for these groups

comp.soft-sys.math.mathematica
ucsc.software.maple
comp.soft-sys.math.maple
Maple User Group

For example,

Jul 1 2002, 11:20 pm, Yet another bug in LaplaceTransform

http://groups.google.com/group/comp.soft-sys.math.mathematica/msg/5451f00adea57c7c?dmode=source&hl=en

-----Original Message-----
From: Vladimir Bondarenko [mailto:v...@xxxxxxxxxxxxxxx]
Subject: Yet another bug in LaplaceTransform

Hello.

Alas, Mathematica 4.2 does sadly remind me Mathematica 3.0.
Something like a small breakthrough plus ocean of bugs
(as compared with the previous release).

In particular,

$Version

"4.2 for Microsoft Windows (February 28, 2002)"

LaplaceTransform[Exp[-z^3]/z^10, z, p]

(Null*p^9)/(39366*Sqrt[3]*Pi) + (9/80)*p*Gamma[1/3]*\
HypergeometricPFQ[{-(8/3)}, {2/3, 4/3}, -(p^3/27)] +\
(3/56)*p^2*Gamma[-(1/3)]*HypergeometricPFQ[{-(7/3)},\
{4/3, 5/3}, -(p^3/27)]

Question 1. What is the origin of Null?

Question 2. What is the easiest workaround for this bug,
that is how to calculate the correct answer?

Thank you for your help in advance!

Best,

Vladimir Bondarenko


jacob navia wrote:
We have seen a lot of bug reports in this newsgroup in the last weeks.
What is interesting for me, is that this bugs are being reported now,
not before.

I am a compiler writer (http://www.cs.virginia.edu/~lcc-win32) and the
bugs problem is for me a big issue. I am used that my users will
immediately scream and yell at me at the slightest suspicion that my
software has a bug. It is considered normal that the developers go quite
fast to blame a compiler when they see something that can't be explained
right away.

Obviously, I would not compare the complexity of my small compiler
system with something like mathematica, either in price (my compiler is
distributed at no charge) or in sheer complexity.

But it is startling that bug reporting was never done here, even if many
users surely found bugs and problems. Why?

Is it that there is an implicit "law of silence" what CAS are concerned?

My question then:

Why there were no bug reports before?

jacob

.