Post your ideas for new miscellaneous programs here, or build on ideas posted by other visitors.

Well, I've put the link project on the backburner for now. I recently discovered what I've been working on discovering on and off for the last year (give or take a few months). I've finally figured out how to make an actual OS for the TI89.

Now I know that one certain French guy and the LiteOS team both knew how to do this, but there were a few things that they left out of the available source code, so even if they did compile correctly (which I never got either to do), they still wouldn't be usable on the calc.

I'm now developing an actual OS, so I thought I would throw the topic out to the public for their review.

Firstoff, I need your opions:
Should it be a gui or DOS-like (describe)?
Should it have out-of-the-box support for grayscale, or should I leave more space for data?
Should it support UTF8 unicode or ASCII?
Do we even need a Certificate Memory (64kb)?
Which routines should it come with?
Anything else?

Also a note to anyone that has made any kind of decent routine library for the 89/92+ with TIGCC: please send me a copy of the source code for the libraries and I will include it in the OS and give you credit. But please make sure that the routines don't use any TIOS calls.

All posted opinions are welcome!

<Vx>

Are you still working on your MegaMan game?

Anyway, this sounds interesting. Patrick Pelissier (TimeToTeam) is also working on an OS (PedroM) at the moment, in case you didn't know.
http://www.yaronet.com/ sujets.php?f=82

I'd prefer an OS with a GUI if possible. I don't see a need for grayscale to be builtin natively, and it shouldn't be on all the time (that would drain the batteries).

Are you planning to be able to run TIGCC-compiled AMS-compatible programs? Will the OS include a CAS?

PedroM is almost to a public beta, and can run some/most TIGCC programs, but it doesn't include a CAS, from what I hear. It is open-source though, so maybe you could contribute/combine/something.

--Malcolm

I didn't know about that, no. I tried to find the info on it from your quoted site, but being as I don't know any French and online translators are horrible, I still don't know anything about it.

I did set MegaMan to the side for the moment, too (I just have TOO many projects! ;) ). It's in decent shape, though.

I do plan on making it AMS compatible, but a CAS is a low priority for me right now. I may add one toward the end of the development. Another thing that I hope to do is create better (faster) versions of the TI-BASIC routines and have support for executing any TI-BASIC program (from different calcs) without modification to the files. So basically I'll be including what I've learned from my linking project so you can send/recieve to non-68k calcs.

I'm interested in what our fellow Frenchman can come up with, though. If you know the language, please tell me what's going on.

<Vx>

Heh, sorry, I don't know French either. Here's one post in English that kinda talks about it:
http://pub26.ezboard.com/ ftichessteamhqfrm5.showMessage ?topicID=2160.topic

More should be known once a public beta is released, especially because it is open-source.

I don't have a TI-89, but a GUI would be GREAT-about routines, it should have a better API than the TI-86, with many different options for things like string input so that programmers won't always have to be writing their own.

hey, are you making this in basic or c\asm? (im guessing C\asm), but will this actually replace the TI-OS, (you say "actual os") it would be pretty cool, but i probably would like to leave my 89 factory default. anyway, i made a PC Front-end called G-os, based slightly after the commodore OS called GEOS. it still has some work to do, but it has its own interpreted visual programming language, and command line... its pretty cool. anyway i started making this for my 89 along time ago.. (with out vis language) and it worx pretty well (has a registry system, graphical file explorer with different view modes, like "icons" and "details".) because i suck at C, i am using BASIC with FLIB. the problem that i am having i cant solve, and i was wondereing if you could help: i need a small function (in c \ asm) that returns the file extension of a 89 file as a string ie getext("main\picture") and it returns "PIC". i saw viperos, and it uses something called SLIB. i cant find SLIB on ticalc or anywhere else for that matter, but it does use SLIB to get ext's. i just cant use SLIB, beaucse its too big to have with flib as well. (and i dont know any of the slib functions..) anyway i would appreaciate any help in getting a program to return ext's, thanx.

I sent you an email with the code on it a while ago. Maybe you didn't get it. If that's so, here it is again:

#define USE_TI89
#define OPTIMIZE_ROM_CALLS
#define SAVE_SCREEN
#define MIN_AMS 200
#define RETURN_VALUE

#include <tigcclib.h>

// Main Function
void _main(void)
{
ESI argptr = top_estack;
char *file;
file = (char*)GetStrnArg (argptr);

char *str;
str = (char*)SmapTypeStrings( GetDataType(HToESI((SymFindPtr (SYMSTR(file),0))->handle)) );

while (GetArgType (top_estack) != END_TAG)
top_estack = next_expression_index (top_estack);
top_estack--;

push_zstr((char*)str);
}

This will give you the function you want if you compile it with TIGCC. The first (and only) argument should be the filename. It returns a string which is the file extension.

Note that if it is not a builtin extension, it will come out as "OTH". I can fix this but it's not something I do very often so I would have to look up a bunch of stuff to get a custom extension.

I guess if you really need the custom extensions I could make it do that, too. Anyway, enjoy!

I'm working on a team project called Morvlon 3.00, which is an assembly-augmented BASIC OS for TI that is essentially supposed to be an improvement on Windows. It does use some Exec and ASM routines (defining windows etc) but is basically in BASIC :-). It communicates with PCs via its PC counterpart in VB.NET, so it needs some other utilities and should be a lot faster. Thus, I was wondering if, since you have an OS concept and no interface/operations and we have an interface/operations setup and way too little access to the microprocessor, we could combine our projects.

Additionally, since I notice there are a lot of books on PC OSs out there and none (obviously) on TI, I think that "How to write a TI OS" would make a seriously good technical book-lots of people want to know this.

Quite an interesting idea you propose. And depending on how your shell is structured, it could be quite good.

However, it can't happen right now. Yes, I am able to create an OS that runs, but right now I'm trying my hardest to figure out the write/read protection that affects a huge chunk of the flash ROM. My project is still pretty useless until I can write to the calc's memory.

If any of your team wants to help, I would be greatful and would seriously consider your idea after this major roadblock has been moved. Do any of your team know how to hack that well?

<Vx>

See the 89 program updates. We program in VB.NET and C/ASM, so yes, we can do programs. We just can't fit them under the var size restrictions, so they have to be on Memory Sticks and their performance is less than optimal due to the loading-executing-deleting sequence that has to occur for *each line* of code. And, no, we don't hack-we write programs in ASM and BASIC which is supported by AMS. Link to and from a PC is doable without hacking-it's been done since at least DOS 3 in BASIC, and is still available on WinXP. With x86 ASM, you can do basically anything the hardware can do, and, new since that last post, we've found out that one of the people from another project (I and my team are only a subset of a relatively large contracting group-we take other people's ideas and put them into ours, meaning less work for us, and I work in custom programming) is an OK programmer in x86 assembly, so if we need to use it, we can, but VB.NET has application-to-application interfaces, so as far as we know, we can just operate through TI-Connect.

What is the size of this chunk?

There may be utils out there that are designed specifically for this-as this, for all I know, could be the FLASH App space, with the devices made by TI and the OSs made by TI, and it could be entirely dependent on one obscure register somewhere that only TI and maybe the writer of Xpand know about.

Um-I think I have an idea (borrowed from our C programmer). The archive is supposed to be default write-protected, the reason being that it is the archive and not RAM. He says to just disassemble the archive() function or to use the Sierra C command for it which I don't recall. I think that the command can be found in the source for the mode backup programs in the TI-89 misc program archive.

I actually have no problem with manipulating the archive FLASH. The difficulty lies in changing the OS boot sector and OS reserved FLASH. This is the chunk where an OS upgrade is put, and is only capable of being unlocked by 1 of 2 methods; unlocking the FLASH from the actual area in rom that is being protected, and hardware tinkering. Now I am a great programmer, and a pretty good hacker, but I don't know squat about hardware and am horrible at soldering.

That's the problem.

<Vx>

Wouldn't unlocking the OS boot sector be done by the new OS? Any device's OS would have to do that. Disassemble a freeware OS that isn't copyrighted and email the initalization code to me. I can ask an x86 programmer that I know to tell me what it means, at which point I'll put it in 68k ASM or C and email it back.

Yes, I know what you mean. I have already isolated the code from which the block seems to be locked/unlocked. Like I said, though, this code can't work without being executed from specific regions of ROM, those regions being ones protected in the first place.

Actually I think I just about had it yesterday, but...sadly, my link port is now broken beyond my repair. I can't test anything more until I can get my hands on another calc.

That being said, anyone want to buy a perfectly operable (besides the link port) TI89?

<Vx>

Cut a SilverLink cable (or any link cable) and connect the wires to the appropriate traces or leads.

TI89 like a Photo-Camera, it's possible but is an articolate project there are many aspect:

***Hardware section:

Use of in/out mini-jack like a video-in using mini-video-camera (isn't possibile acquire video stream cause low memory) than click to make a photo. Must to be done the good interface to prevent damages to calc and camera.

***Software section:

Interface with low-level interrupt for redirect the stream to memory.
Make a program to display the stream on LCD display
Make a program to get a frame frm stream and put it in a file (not pic file!!!) eventually jpg [JPG??!!!?!?!?]
Make a program to convert the frame data to a portable PC jpg (it's possible!!!)
Make a program to high compress these files (i've made!)
Make a program to save the frame in a PIC file
Make a program for PC(win\linux) to get the pics

Ok that's all......i'm crazy??? Nooo....think at GameBoy....

byebye

Why not delete each frame as you use it? Then you only need ~800 bytes for video (you'd store it on a USB memory stick or CD).

TI89 like a Photo-Camera, it's possible but is an articolate project there are many aspect:

***Hardware section:

Use of in/out mini-jack like a video-in using mini-video-camera (isn't possibile acquire video stream cause low memory) than click to make a photo. Must to be done the good interface to prevent damages to calc and camera.

***Software section:

Interface with low-level interrupt for redirect the stream to memory.
Make a program to display the stream on LCD display
Make a program to get a frame frm stream and put it in a file (not pic file!!!) eventually jpg [JPG??!!!?!?!?]
Make a program to convert the frame data to a portable PC jpg (it's possible!!!)
Make a program to high compress these files (i've made!)
Make a program to save the frame in a PIC file
Make a program for PC(win\linux) to get the pics

Ok that's all......i'm crazy??? Nooo....think at GameBoy....

If you are intereted in helping me, contact me

byebye

huh? Are you talking about making a camera for the TI-89? If you are, it's definitely possible. You could use the MCM20027, a Motorola CCD array, which would just require a bit of interfacing through a microcontroller, but otherwise be perfect.

Yes, and it would have to be able to write to some standard format like JPEG, TIFF , or ti PIC, or have a RAW format (which would require a decoder). Have fun.

The TI picture format isn't complex at all-it's just an extremely simple bitmap (well, with 3-4 layers for grayscale). As for JPEG and other compressed formats, there are hardwired encoder and decoder chips-not simple, but easier than writing the code for the algorithms by yourself.