Hey everyone
I really want a ti-89 to play all these cool games if anyone is selling any or knows where to get one please tell me. Also is anyone interested in building an ASM-C compiler or interested in building a CAS for the 86?

for a 89 used or new try ebay.com

The z80 used in TI calcs does not allow a program to view or modify any stack values other than the top one, and thus it is not conducive to C programming.

Not only have people been interested in coding a symbolic CAS for the 86, but at least one person was able to code a large portion of one. However, he decided against continuing or distributing the work he had completed so far on the grounds of liability (if people fail a math test due to a bug in his program) and morality (anyone who really wants a symbolic CAS will buy a better calc. A CAS for the 86 would likely only be used to cheat on tests).

If you're serious about buying a new 89 (as opposed to a used one off ebay), Dimension-TI (http://www.calc.org) is cheap.

What about ADD HL,SP? Say you want SP+4, it is simple:

ld hl,4
add hl,sp
; now HL points to the value, and you can
; do whatever you want with it

Since the stack grows downwards, you would use positive values to get function parameters and such, and use negative values to allocate variables. However, this is with a "traditional", non-optimizing compiler design. A code generator for Z80 would need to use register parameter passing, and for auto variables (if you haven't heard of this, you don't know the C language very well :) use fixed memory locations.

Which brings up a good question: why doesn't TI-GCC allow the use of register parameter passing? The only time you would need stack based calling is when interfacing with the ROM routines. And this wouldn't be necessary if the compiler that TI used had been decent.

I think that the author of the 86 CAS discontinued it because he is lazy :) Not that I'm not lazy, but if he really wanted to, he could finish it. And it would be better than the 89's software in many ways.

There already are three C compilers for the TI-86: Small C, TCC, and TISCO. They all do allow parameter passing by the stack, using either ADD HL,SP as described, or getting the stack pointer in IX and accessing values by offset. Of course either one of these methods is slow.

Of these three compilers, TISCO undoubtedly generates the best code, even though even its output isn't highly optimized. The slowness of using the stack can be easily overcome by the programmer by putting everything in global variables. Doing this, the output of the TISCO compiler is at least reasonable.
All three compilers also don't implement the C language completely, providing only subsets of it.

Also, I have to wonder how you could place automatic variables at fixed addresses. That would make them static variables, not automatic, since they're not specific to the invocation of the function. Of course, this could be done with some sort of elaborate scheme in which the variables are temporarily held in a fixed location, and backed up to a stack when needed, but I have to wonder whether this would improve efficiency at all.

I am willing to start a project for people who are involved and want to create a CAS for the ti-86 in one of the previously described C compilers or in ASM or maybe Basic if anyone is interested. email me if you want to imbark onto this project. If we can get it good enough maybe we can elliminate the whole need for TI-89's (the math reasons of course, obviously not for the games except for that cool 32 grayscale thingy.

Ajai Sharma
aka Lenny

Bob Lagagi's 5.332343/32 sense

Speaking of the CAS, are their any programs for the 89 that allow you to see all the steps in solving an algebra problem? Any plans to make one?
On another note, I havent been able to get SMA .36 working. I launch the game, and everything works fine until it tries to locate the sprite file. Then it just freezes and I have to pull the batteries out. I have a HW 2 89 with ams 2.05, UniOs, and HW 2 patch. Anyone have a suggestion to make it work? I suppose my libraries COULD be out of date, but they worked with SMA .34

i don't think it works with hw2. i have a hw2 89 with ams 2.03 with hw2patch and can't get it working either.

Thats odd. I got .34 working with my HW2 89, AMS 2.05. I can't see why they'd make the next version so it wouldn't work with HW2. I think I'll just try updating the libraries.

well, I updated the libraries, and it works fine now. if you want them (the current libraries) go to the url above, its time to team's website. Besides downloading the libraries from their site, I got their version of SMA as well. Howcome none of the other calc sites (besides ticalc.org) have posted a version of SMA to their own server? I went to Dimension Ti, and they dont have it posted because its supposedly a beta.

Well it is a beta, but the sites can post it on their own server.
I don't know why the other sites don't do it.

By the way, you have crashed because of the new file 'monster' which is different, and I was too stupid to change the signature of the file.
Sorry !

You couldn't always place automatic variables at a fixed location, and yes, it would make them similiar to static variables, but in many cases an automatic variable can be made a static variable while still preserving the program semantics. If the compiler can make a few assumptions, such as no function aliasing, then the compiler can track which functions can possibly be called from each other and from themselves.

Name one way it would be better? Or are you just an 86 owner with a serious case of calc envy?

Using it on the ACT because they do not know 86's have a CAS (maybe) do they? How about the ease of useness for advanced mathematics students. These are just a few ways. And not to forget the ability to by a calc cheaper and still have a CAS

Bob "Lenny" Lagagi places his instructive 5.434343/32 sense into anothers hand

i have to say this:
I think eric is a much better poster than nick. granted he's new and all, but I do think that given the time that nick had, ticalc.org will have another era of news postings.


also - I'm looking for advanced 83+ developers, so give me an e-mail at
steve@suitestevie.net
kinda 'top secret' like programming for it (z80)

Steve
ps: I freak and hate this 'handle'! will somebody react to my e-mails and change it for me?

Nick rox. DOT.
Eric is almost as good.

My posts weren't good at the beginning. Being a good news editor takes time. Eric has time. Keep an open mind to things.

--BlueCalx

PS: Frank, I'd like that PHP counter code if you may ;)

You were good all the way man. I loved to read your dosage of creamy goodness. And man if you have time to read these posts then maybe you will have time for some occasional news descriptions because man my creamy goodness level is low.

A tear jerks from Bob "Lenny" Lagagi's eye as he contributes is 5.233343/32 sense

He was not good all the way. It took him a week or two. Then he got the hang of it and was great. :)

I think I'm a little rusty myself--or else Nick's just kinda changed the way we're all used to news. Anyhow, it won't take long for us to get settled again and for everyone to get used to our styles.

Trust me--it's happened twice since I've worked here. Oh, and browsing news comments isn't nearly as painful as writing news items! (right, Nick?) ;)