Professor Stein, you are completely missing the point.
Any serious attempt to check source code would include in a thorough check
of what variables are used, their types, what they are used for, how they
are used, what (if any) values they are initialized to. That would be true
whether you were looking at your own code or that written by someone else.
I can not possibly see how an author can check his/her code properly unless
they have done that. Doing that, one would soon realize these problems.
Neither can I understand how mathematical software than needs to be compiled
by an end-user can be taken seriously unless there is a test suite. Lcalc
lacks one.
When someone can't be bothered to check their code, then why should I
believe they have taken due care to implement the mathematics properly? They
may be the best mathematician in the world, but I personally would not trust
their ability to write high quality software.
Of course, any non-trivial program is likely to have bugs, but the frequency
of bugs could be reduced by careful testing - something that has clearly not
been done in lcalc.
I can see some of the source code from Mathematica in the form of .m files.
The care that has been taken in writing them is clearly more than the care
which was taken in writing lcalc.
In principle with open-source code, it is possible for a user to check the
code themselves. In practice, that is not likely to be the case, for all
except the smallest of sections, especially when the code is poorly
commented.
I do not fully trust any program, but my confidence in Sage is decreasing as
I look at code like lcalc and see some of the comments on public mailing
lists.
It's a shame, as I like using Sage, but after looking at some of the source
code, in particular that of lcalc, my enthusiasm has decreased markedly.
lcalc just happened to be the part that interested me, which is why I have
looked at that in greater depth.
Anne.
On 8 June 2010 20:30, William Stein wst...@gmail.com wrote:
On Tue, Jun 8, 2010 at 12:21 PM, Anne Driver annedrive...@gmail.com
wrote:
When building open-source software, it is not unusual to see a few
compiler
warnings - usually deprecated code. There's quite a lot of them in Sage.
But looking at lcalc, the code seems particularly poor, with lots of
unused
variables deprecated code. It does not exactly inspire confidence in
Sage.
I don't know how it possible to check code core correctness under such
circumstances.
The person who wrote Lcalc is easily the world leader in algorithms
for numerical
computation of zeros of L-series of all types, and has done vastly
more along these
lines than anybody else. I would greatly prefer code with unused variables
from a person that understands the relevant mathematics, than the
other way around.
tmp_x is unused. I thought that would be defined in a C or C++ file, not
a
header file. Lfind_zeros.h It's actually normal practice to have code in
C
and C++ files, not header files.
Functions are declared to return values, but do not do so. 'missing_data'
is
unused. Is that since it's not needed, or because the author forget to
use
it, so the code is unstable?
Maybe you should ask the author...
There are reports of deprecated or antiquated header which may be
removed
without further notice at a future date.
Unfortunately, lcalc was the part of Sage I intended using - now I think
I
will look at other software. Hopefully some high quality open-source
software. Parts of the code are quite well commented, but other parts
have
few if any comments. This has rather put me off.
Good luck.
-- William
Anne
--
To post to this group, send email to sage-support@googlegroups.com
To unsubscribe from this group, send email to
sage-support+unsubscr...@googlegroups.comsage-support%2bunsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sage-support
URL: http://www.sagemath.org
--
William Stein
Professor of Mathematics
University of Washington
http://wstein.org
--
To post to this group, send email to sage-support@googlegroups.com
To unsubscribe from this group, send email to
sage-support+unsubscr...@googlegroups.comsage-support%2bunsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sage-support
URL: http://www.sagemath.org
--
To post to this group, send email to sage-support@googlegroups.com
To unsubscribe from this group, send email to
sage-support+unsubscr...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/sage-support
URL: http://www.sagemath.org
