1. Forum
  2. Usenet
  3. SCI.MATH.SYMBOLIC

  • FYI, Sympy dsolve added to independent CAS differential equations test

    From Nasser M. Abbasi@21:1/5 to All on Fri Mar 14 12:15:03 2025
    I do not know if any one still reads this newsgroup.

    But in case, there is a new update to the Computer
    Algebra Independent Differential Equations Tests

    <https://12000.org/my_notes/CAS_ode_tests/index.htm>

    Added Sympy 1.13.3 in addition to Mathematica V 14.2
    and Maple 2024.2.

    The tests now uses almost 20,000 differential equations.

    This is percentage solved for each CAS
    ======================================
    Maple: 96.219%
    Mathematica: 95.592%
    Sympy: 70.477%

    Full reports are online (PDF is over 33,000 pages)
    with plain text commands used for each CAS and for
    each ode.

    Sympy seems to handle standard ode's well, but for
    more advanced ones (non-linear and so on) it still
    lacks behind the commerical CAS ones.

    Lots of tables are included showing different types
    of results obtained.

    --Nasser

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Jeff Barnett@21:1/5 to Nasser M. Abbasi on Sat Mar 15 01:31:14 2025
    On 3/14/2025 11:15 AM, Nasser M. Abbasi wrote:
    I do not know if any one still reads this newsgroup.

    I do! I am no longer active with symbolic math. At one time I had a
    (paid for) copy of Macsyma and later the freer one of a similar name. I
    was hired at the old System Development Corporation around the mid 1960s
    to work on the development of Lisp 2 along with a gang from Ed Fredkin's
    III. A fellow at SDC, Clark Weissman, at about that time, wrote a Lisp
    Primer for what, today, we would call a Lisp 1.5. It had a special
    problem with exercises in each chapter that incorporated the language
    forms used in that chapter. The special problem was developing a program
    that could do symbolic math including differentiation, some
    simplification, factoring, etc.

    I was quite enamored how easy that was to do in Lisp. I had a job
    working on the Apollo program a few years before that. A control
    engineer talked to me one day about how hard it was to combine and
    manipulate S-plane rational fractions and do analyses on the results. I designed and wrote an implementation of an interpreter for a "machine"
    whose memory could hold polynomials, rational fractions, and constants.
    The order codes were add, subtract, multiply, divide, etc for these
    primitive representations, graph them, and derive various features of stability, vibration modes etc. That code changed the way the engineers
    thought about their work and the processes they used.

    As a coincidence, Tony Hearn, during the Lisp 2 years, was developing
    his Reduce system on the same computers that we used for Lisp 2. He used
    an existing Lisp there for his research. Ever since those intros to
    symbolic math in my youth, I've been fascinated and tried to keep in
    touch with progress in the field.

    I particularly value the kind of results that you develop and publish.
    Symbolic math in some ways resembles the halt problems. Say someone says
    they have a halt decider that's really good. You might come along and
    grab a few thousand programs that are between a 100 and 1500 lines of
    code (in various stages of development) run them through a decider and
    report percents of right, false positive, and false negatives. And that
    would tell us a lot more about the field than any of the claim-ware advertisements.

    Keep up the good work and I'll continue to peek in and see how everyone
    is doing.
    --
    Jeff Barnett

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From [email protected]@21:1/5 to Nasser M. Abbasi on Sat Mar 22 14:08:07 2025
    "Nasser M. Abbasi" schrieb:

    I do not know if any one still reads this newsgroup.

    Thanks for the message.

    Half a year has passed since the preceding post (by Mild Shock) arrived
    on Eternal September. Accordingly, I am checking <sci.math.symbolic>
    only occasionally these days.


    But in case, there is a new update to the Computer
    Algebra Independent Differential Equations Tests

    <https://12000.org/my_notes/CAS_ode_tests/index.htm>

    Added Sympy 1.13.3 in addition to Mathematica V 14.2
    and Maple 2024.2.

    The tests now uses almost 20,000 differential equations.

    This is percentage solved for each CAS
    ======================================
    Maple: 96.219%
    Mathematica: 95.592%
    Sympy: 70.477%

    So Maple manages to stay ahead still!

    Martin.


    Full reports are online (PDF is over 33,000 pages)
    with plain text commands used for each CAS and for
    each ode.

    Sympy seems to handle standard ode's well, but for
    more advanced ones (non-linear and so on) it still
    lacks behind the commerical CAS ones.

    Lots of tables are included showing different types
    of results obtained.

    --Nasser

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Nasser M. Abbasi@21:1/5 to Nasser M. Abbasi on Fri Apr 4 22:55:21 2025
    Fyi;

    Updated Computer Algebra Independent Differential Equations Tests
    for the new released Maple 2025 version.

    <https://12000.org/my_notes/CAS_ode_tests/reports/maple_2025_mma_14_2/index.htm>

    Very little changed in result from Maple 2024.2, 8 more ode's
    was solved in Maple 2025 vs. from Maple 2024.2. Added only
    few more ode's to the collection. No time to add more.

    I am now using Maple on Linux and not windows.

    Here is the result
    ------------------------------
    Maple 2025 (Linux): 96.246%
    Mathematica V 14.2 (windows 10): 95.558%
    Sympy 1.13.3 (linux): 70.458%

    Might be able to make new version of the Integration tests
    before end of year. Waiting for a new version of Maxima
    and for Mathematica v 14.3 to make it worth doing as
    it takes about 2 months to run the integration tests.

    --Nasser

    On 3/14/2025 12:15 PM, Nasser M. Abbasi wrote:
    I do not know if any one still reads this newsgroup.

    But in case, there is a new update to the Computer
    Algebra Independent Differential Equations Tests

    <https://12000.org/my_notes/CAS_ode_tests/index.htm>

    Added Sympy 1.13.3 in addition to Mathematica V 14.2
    and Maple 2024.2.

    The tests now uses almost 20,000 differential equations.

    This is percentage solved for each CAS
    ======================================
    Maple: 96.219%
    Mathematica: 95.592%
    Sympy: 70.477%

    Full reports are online (PDF is over 33,000 pages)
    with plain text commands used for each CAS and for
    each ode.

    Sympy seems to handle standard ode's well, but for
    more advanced ones (non-linear and so on) it still
    lacks behind the commerical CAS ones.

    Lots of tables are included showing different types
    of results obtained.

    --Nasser





    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)