+- HP Forums (https://www.hpmuseum.org/forum)
+-- Forum: HP Calculators (and very old HP Computers) (/forum-3.html)
+--- Forum: General Forum (/forum-4.html)
+--- Thread: Quasi Literals (/thread-12846.html)
Quasi Literals - ttw - 04-22-2019 05:30 AM
I found useful trick for the HP50g and perhaps others. I was doing some matrix work and I wanted to work on some general vectors with entries: a,b,c,.... which often causes problems if one of these letters is in the folder chain above. Also some things just don't work on literals.
So (if I don't square an object) I can uses Sqrt(2), Sqrt(3), etc in exact mode as literals. Thus I can get the coefficients of a product of matrices as a transformation as the Sqrt(2) and Sqrt(3) (and other primes) "exist" in another realm from the integers.
I think I could use very large primes Sqrt(1007) or the like if squares are needed but the problem is that differences of squares can be small numbers and lead to confusion of numbers from the matrices and from the "literal" stuff.
[[3 5][5 8]]*[Sqrt(2), Sqrt(3)} = [3Sqrt(2)+5Sqrt(3) , 5Sqrt(2)+8Sqrt(3)] and the square roots act as symbols.
I'm trying some stuff for multiplications. The problem is that addition of products may cause a symbol to become numeric. I need a couple of independent "free groups" on two or more symbols that can correspond to addition and multiplication.