Survey posted 2005-04-19 18:01 by Jon.

Contribute ideas to surveys by sending a mail to survey@ticalc.org.

TI should probably make a special calculator CPU since it would be better suited for a calculator since it would be designed for a calculator. It would probably be better to make a NEW slow and efficient CPU rather than just using really old slow ones.

But WAY more expensive. THe cheapest the first calc with a custom processor could be is 500 USD. Nobody'll buy it, it'll be awhile before anybody can program in other than BASIC for it.

It would be an absolutely pointless waste of time, money, and R&D power.

LIES! The TI-80 had a proprietary processor, and it wasn't near that expensive.

Uh... The TI-80 sucked though...

120d//!

Uh...I see that you haven't tried designing a processor anytime recently...

While you're at it, why not just stick in a couple low-power FPGAs, one with a hardwired OS, and just let users configure their own programs in the other FPGA? It would be a lot faster than software, and I think there are programs that create VHDL code from C. :-p

I'd definately say ARM. RISC instructions are so nice to work with.

Yeah, they need to grab the ARM. :) Maybe do what HP does and make it back-compatible with the 89.

How 'bout Pentium? er Ok, I know that would get too hot. But I have a old mac that has a 20Mhz Moderola (no, its not spelld right) processor and it seems fast...faster than ti's calc programs...

Actually, an underclocked PI would probably work just fine (albeit they are somewhat large).

How much underclocked? The first ones ran at I think 20MHz

The size is the main problem with using an underclocked early Pentium, or 386, or 286, even if it doesn't use much power or get hot - there would be serious problems just fitting it in the case! And imagine stuffing northbridge and southbridge chips in there too to interface with the processor...or TI would have to design their own custom ASIC for connecting the memory, linkport, and LCD with the FSB.

i think they should definately get a new cpu. a new programminig language could give newbs (like myself) a head start on the veterans so we could see some variety in the programmers

n00b alert!

Naw, the "veterans" will learn the language quickly and won't wait for the n00bs to catch up. It's the way it's always been and always will be.

There's much more to programming than just learning a language.

Yep, the more languages you know, the faster you can pick up another one. It just works that way.

StrongARM, baby!

That would take up too much power I suppose, but the HP-49G+ has an ARM running at 75Mhz.

The HP49g+ is a nice calc, but its resolution, OS, and
adaptability are not so great,(nothing against HP!!) atleast in my "research".

XScale. It has a bunch of stuff on the chip, everything from a USB host to a controller for a VGA screen, and is already used in portable systems (the Axim PDA, for example).

68k of course! It's nice to program in assembly, and it's easy to retarget TIGCC for another 68k-based calculator.

I'd like to see a 68030 (or higher) chip in the next calcs, and then you'd just have to change the floating-point stuff in TIGCC.

Floating point arithmetic would be a lot faster on the calculator with a built-in coprocessor.

Actually, the _68040_ and higher have floating-point support built-in. My bad. The 68020 and up have memory management units, though, which would be really nice to write a real OS for the calculator!