I'm sure you are aware of the Sage open-source mathematics software

http://www.sagemath.org/

which has a mission of creating a viable free open source alternative
to Magma, Maple, Mathematica and MATLAB.

Obviously Sage has a test suite where results from Sage are compared
to a set of known results. For example, one test for the factorial()
function is:

sage: factorial(10)
3628800

As you are no doubt aware, all non-trivial software contains bugs. It
would be very useful to compare the result from Sage to that of other
software which is developed independently.

One way, which could be used in some circumstances, is to compare the
Sage result to that obtained from Wolfram Alpha. For example

http://www.wolframalpha.com/input/?i=10!

shows 10 factorial is 3628800, so there is a very high probability
that WolframAlpha and Sage are both correct.

It would sometimes be useful to add a comment to the Sage test suite
that the result has been compared to that obtained by WolframAlpha. So
we could write something like:

---------------------------------------------------------------------------------------
WolframAlpha gives the same result as Sage - see:
http://www.wolframalpha.com/input/?i=10!

sage: factorial(10)
3628800
-------------------------------------------------------------------------------------

Sage has tens of thousands of tests and that number is increasing all
the time. Only a fairly small fractions of those tests could be
computed with WolframAlpha, and even in cases where they could, we
might not chose to do so.

Looking at the terms of use of WolframAlpha,

http://www.wolframalpha.com/termsofuse.html

I personally can't see anything that would suggest that comparing
results with Wolfram Alpha, and documenting this  would breach the
terms of use. But when I suggested we could verify a result in
WolframAlpha

http://www.wolframalpha.com/input/?i=N[Integrate[+Sin[x]%2Fx^2%2C{x%2C1%2CPi%2F2}]%2C50]

one Sage developer questioned whether this would be within the terms
of use. See:his comments at:

http://groups.google.com/group/sage-devel/msg/1f8af294fbf40ccc?hl=en&;

One section in particular of your terms of use says::

"You are not allowed to use Wolfram|Alpha to create something that is
likely or intended to be reused as a data source for further
processing, or that in some other way serves as a replacement or
alternative to using Wolfram|Alpha itself. This applies whether what
you create is in electronic or print form."

Sage, has a web based interface that allows one to perform advanced
mathematical calculations. Clearly there are some calculations that
could be performed in WolframAlpha, but which could also be performed
in Sage. If you try Sage  - you can get a free account at

http://t2nb.math.washington.edu:8080/

you will soon realise that Sage is quite different to WolframAlpha.
Sage is certainly not intended to be a replacement for WolframAlpha -
in fact, Sage existed several years before WolframAlpha.

Sage has its own language, which is based on Python. Sage can only
process input using that syntax. It does not attempt to process
questions the way WolframAlpha does.

To save any further discussions on the Sage developers list about
whether the use of WolframAlpha in the way I explained would be
permissible, could you please clarify the matter.

Obviously using WolframAlpha to compare results with Sage would be of
benefit to the Sage project. But it would also benefit Wolfram
Research too. In the event that comparisons with WolframAlpha showed
different results, and we concluded WolframAlpha had a bug, we would
out of politeness let you know. In fact, only recently I made your
technical support team aware of a documentation error in PrimePi[] and
PrimeQ[], which I understand will be fixed. This documentation error
was discovered when some comparisons were made between Sage and
Mathematica.

Dr. David Kirkby (a developer of the Sage mathematics software).

-- 
To post to this group, send an email to sage-devel@googlegroups.com
To unsubscribe from this group, send an email to 
sage-devel+unsubscr...@googlegroups.com
For more options, visit this group at http://groups.google.com/group/sage-devel
URL: http://www.sagemath.org

Reply via email to