SandMath/Matrix and others: Module Updates Available

[Ángel Martin]

The "Xmass Edition" of the SandMath, SandMatrix, 41Z, Library#4 and PowerCL Extreme modules is available at TOS. With Updated manuals in case you need a refresher ;-) Enjoy!

[HP-Collection]

You're unbelievable.... How many hours is your day? It must be at least 27 =
Thanks for your Xmas present and happy Chrismas to you from Switzerland.

[Michael Fehlhammer]

Yes, indeed, unbelievable, thank you, Angel!

Angel, are you in search for new challenges for the next year?
What about a BASIC interpreter for the 41CL - speed and memory size should be sufficient! Just the display width is poor, I know ...
:-)

My wish ( dream :-) ) is a result from the following observation:
FOCAL language is totally overchallenged with matrix expressions that are very common in numerical mathematics: terms like K(i,j) = M(i,k) * N(k,l) + P(l,j) f.e. are easily implemeted in BASIC, easy to read and edit and quite fast. Try this with the CCD Module matrix functions - possible, of course, but somewhat ridiculous.
So even if the implementation of a complete BASIC environment might not be an option, a BASIC subset for definition of matrix and vektor calculations would be very helpful, and feasible perhaps.
Such a definition could be done in form of an alpha string which is interpreted by the function that follows immediately ( in a FOCAL program).
I know that there are special functions for row- and column-sums, determinants, cross-products ... but that doesn't really help much. Real-world numerical algorithms ( like differential equation solvers, simplex method etc. ) always need individual combinations of matrix elements that cannot be reduced to standard functions like the above mentioned ones!

[Michael Fehlhammer]

Just forgot to say:

Merry Christmas to all MoHPC members!

[Ángel Martin]

(12-24-2014 05:03 PM)Michael Fehlhammer Wrote:  So even if the implementation of a complete BASIC environment might not be an option, a BASIC subset for definition of matrix and vektor calculations would be very helpful, and feasible perhaps. Such a definition could be done in form of an alpha string which is interpreted by the function that follows immediately ( in a FOCAL program).

I know that there are special functions for row- and column-sums, determinants, cross-products ... but that doesn't really help much. Real-world numerical algorithms ( like differential equation solvers, simplex method etc. ) always need individual combinations of matrix elements that cannot be reduced to standard functions like the above mentioned ones!

Very true Michael, but I'm afraid you overestimate my abilities. Such an implementation would no doubt require a variable management facility completely lacking in the platform, and writing it from the scratch is a very tall order.

Perhaps somebody else can get it started and we'll join in ;-)