Still needed for the HP 41CL

[Sylvain Cote]

(08-05-2016 01:42 PM)Ángel Martin Wrote:  The "Next Program" feature can be extracted into an independent routine called by all others, instead of residing everywhere.

Yep, that is what I was looking at the optimization phase.

(08-05-2016 01:42 PM)Ángel Martin Wrote:  However I've looked at the manual and those 222 data files may be a problem: assuming each of those requires 10 registers, we're talking of a humongous database of ~15k !! More than three ROMS just to put all that data in...

I did not looked at that part yet, but you are right this is an issue, actually it is worst than that.
Using the raw version of the W6X who contains a parser and 4 data loader as a reference, the math gives the following:
512 bytes (RAW) / 4 (data loaders) = 128 bytes / 7 (bytes/register) = 18.29 -> rounded to 19 registers
19 registers (133 bytes, per data loader) x 222 data sets = 4218 registers or 29526 bytes.
Adding to the already used 8KB for program code we are looking at something like a 40KB ROM which does not fit into the 32KB directly available.
Even with all the tricks that can be applied, 4 x bank switching, a bank switcher function in MCODE, etc, we are still looking at a 2 slots solution for the 41CL or an MLDL (Clonix/NoV/MLDL2K/etc).

(08-05-2016 01:42 PM)Ángel Martin Wrote:  The alternative is MCODE tables, but such is a huge effort than I'm not really contemplating doing... not even in the space-time continuum I live in :-)

Common, the GREAT Angel not tackling a challenge! hi! hi! hi! just teasing here

Taking a step back, we could offer an hybrid solution with the code in ROM and the data in an external IL storage (tape,disk,pilbox+soft)

Sylvain