I think TI should only make a TI with a FAST Basic language. I mean so fast as ASM. That would be cool. Then we can make perfekt games on our TI. Thats really my BIGGEST desire.

-Fabs

Well, besides the fact that this is physically impossible...

The reason BASIC is slower than ASM is that the BASIC program is *compiled* oncalc as the program is run. ASM programs are precompiled on a computer (most of the time) and therefore are much faster. If BASIC programs were somehow precompiled, the calc would crash every time an error was encountered.

Can't they just include an error handler in the assembled program?

i really dont think so, have you heard of xpak?? it was written in the c compiler GTC that seems to be a soon release, if tigcc were able to make this compiler inside the calc why TI cannot improve their prgm writter to make it a compiler??

So are you saying that it is or is not possible?

Note that GTC is NOT a release of the TIGCC team, it is just compiled WITH TIGCC because that compiler can't even compile itself. And I am not sure Paul Froissart will ever release it, he keeps delaying the beta release all the time.

PS: And it is not a port of TIGCC either. The acronym is misleadingly close to "GCC", but it hasn't anything to do with GCC. GCC needs way too much memory to run on a calculator.

You're definitely right about the "physically impossible part". Also, it's just not worth making BASIC any faster-learn C++, assembly, or TIPower-they're not so hard.

Not as fast as ASM, but faster... If the calculator compiled it BEFORE you run, not during, it would go much faster, but the startup would take forever... my suggestion, allow both... run the program as is, and then a little option in the PRGM menu, "CMPL", to CoMPiLe the program... and the compilled program gets saved as a normal program, and to run it you use Asm(. If this is an 83+ Family... It would take more, but be cool.

Actually, it sort of does do that on the 89, but still runs slow. It doesn't actually COMPILE the program, but the first time it runs it checks the whole thing for errors making it take forever, then after that it starts up immediately as long as the source hasn't been edited.

It's called tokenizing. It doesn't do that on the TI-82/83 series because the programs already use BASIC tokens.

No, I think he's talking about a system where when a program was run for the first time, each token would be replaced with a bit of assembly code, then the whole program would be run as an assembly program.

The reason that Basic is slower is that it's an interpreted language. In a way, a Basic program isn't even code, it's a piece of data you give to the interpreter, which then performs actions based on it.
And there is no reason why it would crash on every error; if the interpreter can detect and catch errors, why would it be impossible for a "compiled Basic" program to do the same?

You're right. Every language feature could be duplicated in a compiled version, much the same way I can compile QBasic programs, and the compiled versions still catch errors (runtime errors only, of course).

Why can't the BASIC interpreter be user-accessable? It seems to me that that would allow for on-calc assembly programming.

The BASIC interpreter is the same as the command-line interpreter.

Exactly....by using ":" you can run BASIC programs from the homescreen, and you can perform homescreen commands and operations in BASIC programs.

It's just like in bash, for example. You use the exact same syntax and commands both on the command line and in scripts.

How would it allow for on-calc assembly programming?

What we all need is an on-computer TI-BASIC editor, which compiles into an ASM program. To prevent unexpected ram clears, the programmer will be notified of the syntax errors to fix.

So this built-in USB port...What does this do for the fate of the graph link, and graph link software? I assume they'd still be compatible, TI wouldn't be that dumb to abandon all graph link consumers and software. Atleast I don't think so..

Update of TI Connect I'd imagine...

The easiest thing is simply to act like you're a SilverLink. I can't see any reason for them to do anything else.

Well, the 86 is one of the few calcs that can only use the graphlink and not USB, and they've already abandoned the 86 users (just about)....

That's becuase 86's suck... On a different note, the calc will still have the 3/32in port (standard I/O port on calc).

Just because you don't know how to use the 86, or you are uncomfortable with them, doesn't mean they suck. I'd like to see the 83 find inflection points on a graph. (That's something the 86 can do and the 83 CAN'T.)

Actually it can...with symbolic (is that the calculus program? I may be mixing up names...), you can take derivative of a function, and then find the zeros of the new functions which obtains the inflection points :) I love being right...

If you look at TI's calc comparison chart, the TI-86 is closer to the TI-89 in functionality than it is to the TI-83 Plus Silver Edition in everything other than Flash Memory.

The 86 also can find the roots of polynomials, solve systems of simultaneous equations, etc. without needing an app that can only be used by people who buy a cable, and takes up a sizable portion of the memory. You can also type program names and other tokens directly, instead of having to insert them from a menu....I could go on and on....

No, I can access an 86 in TI-Connect on my Mac. I can also access the 82, 85, ans 81. Yes, i can link to an 81. Since it is only to the ViewScreen version, I think they do it through that.

The legacy I/O port will still be there, so I think you will probably be able to use the TI-89 GraphLink software just fine. Maybe the directory listing will have some glitches, as it does when using the TI-92+ GraphLink with a Voyage 200.

As for the TI-84+, the TI-83+ GraphLink may or may not work. The processor speed might cause problems. (TI recommends against using the TI-83+ GraphLink with a TI-83+ SE for that reason.)

As for TI-Connect, the autodetection makes things trickier (as it won't just let you connect a calculator and claim it is another one), but I'd expect it to work with new USB-enabled calculators on an old GraphLink cable. Let's see.