SD: Re: Let me propose something then...
[Prev][Next][Index][Thread]
SD: Re: Let me propose something then...
Doing this would require that one could access external files which is hard
on some calculators, but not impossible. Instead of making it a shell, it
could easily be implemented as a program for one of the existing shells.
Another idea would be to make it possible to includes small functions in
asm, which would make it mroe useful and faster.
Dines
----- Original Message -----
From: "Robin Kay" <kaypictures@clara.net>
To: <shell-developers@lists.ticalc.org>
Sent: 1. januar 2000 23:09
Subject: SD: Let me propose something then...
>
> How about creating a new tokenised language (TI-BASIC is tokenised)
> for TI calculators. First of all scrap expression evaluation becuase
> that is too slow, replace is with assembler like arithmatic. Allow fixed
> point maths for increased speed. Then replace the Lbl statement with
> absolute jumps like with jp and jr. Create a comprehensive platfrom
> independant API for accessing functions. Then replace slow interpreted
> code with Just-In-Time compiling. The majority of the work of the JIT
> would simply involve replacing language tokens with machine-dependant
> macros or ROM calls. The end result would of course be slower than
> writing programs in pure assembly, but still much faster (and more
> functional) than BASIC and still machine independant (between Z80 and
> 68K). Or course this require programmers to port the underlying shell to
> all the different calculators first, but after that (in theory) any
> program written in the new language would work without modification of
> any TI-Calc. Such a system --could-- revolutionise programming TIs as we
> know it, much in the same way that Java is doing on the PC platform.
> Think about it, and then reply with your position...
>
> --Robin Kay--
>
>
References: