|Re: Getting HP to market an HP42sx - comment to Christof's comment...|
Message #9 Posted by Bill Wiese on 27 Aug 2003, 1:24 p.m.,
in response to message #7 by christof (NoVA US)
[Christof wrote: ]
i'm a bit skeptical about their actual ability*,
let alone willingness, to provide something of
the hardware and software capability of the 42S.
look at the obvious and silly bug with program
entry limits in the 12CP, the case durability of
Hence my desire to add on to my 42S further :)
I agree about the 42S :) Dream calc.
HP/Kinpo failed on the 12CP because they had a small time & cost budget and decided to do a rewrite from scratch. In situations like this it would've been faster & more reliable to write a 'Nut CPU emulation layer - poss w/a few 'hotspot' areas optimized - that runs on the 6502. Even if written in assembly code, this wouldn't take too long; not that hard to emulate a 'Nut in C or assembly macros.
Instead of a "code interpreter" the emulation could also be done as a machine translation (i.e, at "compile time") via macros/pseudoops & subroutine calls, taking into account ROM space use.
Since I/O setups would be different, you'd jump in & out of emulation for KB, display I/O.
Then the 12C code could could run in 100% conformity to a regular 12C. Once this is done, minor changes (# of program steps, etc) could be done incrementally.
Building a new-old calculator is not difficult these days software-wise if you can change/replace/reflash a ROM with your choice of firmware. The new ARM-based 49 would be WAY cool for this. (Linux on a 49 anyone? :) The hard part for small runs on consumerish widgets like calculators is finding/manufacturing keytops w/appropriate legends and keyboard overlays w/legends. This takes at least (guesstimating) $25K-$40K for even small runs because setup costs are the big thing.
San Jose, CA