In reading over the discussion generated by this news item, I am impressed with the zeal there is to make an open source OS. I do not fault this but I have a caution. The programming of an OS is not a small thing. The specific hardware of a chip is reflected in its assembler to an extent, but an efficient and safe OS requires a deeper understanding of how the hardware is put together.

My caution is to make sure that the people heading the programming portion of this are people who have a good understanding of the physical hardware. This way they are able to write an OS which is usable to programmers, be they gammers or not. This will also allow the OS to be optimized for speed. A good assembler OS written with hardware optimization may be at least 50% faster, and much more stable than one without. Example from Atari, it requires a pentium 90Mhz to emulate Atari, due to the "tricks" that the programmers used that were based upon the hardware.

I do not know the level of understanding of the hardware in ticalc's community but I think that this should be considered.

(One option would be to pursue the creation of the OS as a flash-app. I don't know if flash-apps run unhooked from the TI-OS or not)

Philip Sugimoto
aspiring Computer Engineer

Another thing. In class we designed and built an 8-bit CPU from programmable logic. We wrote a program to run on it and had to translate it into hex-code by hand. One thing I noticed was that it would not be hard (though maybe time-consuming) to have written the assembler and such.

Thus another sugestion is to have the OS people release a standardized assembler based on what the OS can do. This could be added to and changed as people improve on it. This way we could start with a standardized language and try to keep it so. (Any changes being agreed on by "most" programmer. ie always backward compatible).

My $0.02

Philip Sugimoto

I think the new OS should be somewhat similar to the current one as far as the interface goes. Sure, a GUI would be nice and pretty, but you have to make it easy to swithch from one to the other(mentally, I mean), so people who are used to the TI-OS will not have a hard time getting used to the new one. Also, it has to be flexible, so it should have some GUI and command line features together, but nothing exessive like Windows or something.

More like an 86

I HATE the frilly menus that waste a huge amount of screen space.

Heck, we should be able to hack the rom to give us our screen space back.

It would also be nice to play with the split screen feature... make it waste less screen

HACK AMS 2.01 OR 2.02!
They denied that it's thiers, so they couldn't sue us for decompiling it.
And besides, would they actually sue teenagers (some in foriegn countries)?
LOL, they probably WOULD try to, but I doubt they would get anywhere with a case.
And we would probably get the support of everyone in the country, and go on Dateline, CNN, ZDTV, and other things.
lol, I live in SF where ZDTV is done, but the cable company doesn't have that channel. (my french teacher is married to a ZDTV guy)

Someone explain to me. If TI can modify the FLASH why can't we? Surely they have to leave something open so that they can modify things? Do they use some fancy checksum or something?

Just wondering.

Ciaran

They use a digitaly signed RSA encription (512 bit i believe). So cracking it is not really worth presuing, the best method would be a physical hardware crack (kind of like a playstation mod chip) or to use the current hacks that maxmem and hw2patch use

i think its about damn time someone tried to do something like this! :)

Wow Nick... broke out the red text for this one again!

Ok, this would totally kick ass. TI seems to make new limitations every time, especially on asm progs. Well, not anymore!

- Jeff

I saw all these comments about ideas for the new OS, making it a gaming machine, making it emulate the hp49, linux, CLI and other dirty words were flying around 8) Why don't we make this OS customizeable, I mean say have a directory with "dll" files (libraries, whatever) and make these updateable.

Say you want an RPN math interface, load the RPN module. Want a Linux or CLI interface? just load the module. Same goes for the libraries we currently use. make them a group of libraries (all libraries will be in assembly of course-->fast) and make these updateable, so when a new version comes out, just overwrite the previous one. This kind of OS will let us customize it as we want, and all that will be needed is the kernel that will be the OS itself. If you want a GUI, just load the dll for it, btw the ti89 (which i've got) would definately need a GUI, because a CLI would be pretty slow...typing is horrible 8, though prosit worked beautifully with the mouse thing, shame it doesn't work on AMS2)

This model of the OS will also make it possible to port it to HW3 and so on calcs... as long as they are asm compatible, just update the kernel.

Another thought that crossed through my mind... this kind of thing cannont go head to head with TI, firstly, i belive that a checksum is needed to update ths flash rom, and also that non writable area where the system loads if it fails. TI designed this calc so why not get their help for stuff like the checksum. Besides I read on that hpcalc.org site that the latest OS was developed with the best hp48 programmers, so why don't we do it as well? this would make the OS much more stable and there would be no legal problems with TI, and they would not blockade our attempts to make asm progs.

just my $0.02 on the OS
btw, I am totally for this new OS thing, and I think the Ti89/92 Hardware can more then easily take upon the task...the hardware is there, we just need the software (just like the sounlib that let's us have sound and the clock prog that gave us a clock in the taskbar)

Summoner

>Besides I read on that hpcalc.org site that the latest OS was developed with the best hp48 programmers, so why don't we do it as well? this would make the OS much more stable and there would be no legal problems with TI, and they would not blockade our attempts to make asm progs.

HP != TI. There's a reason why the HP is often considered to be slightly better than TI's.

I think a new OS should be made to resemble the Mac OS. I mean, it is the most user friendly environment in the world. Oh well, that's my $0.02 in this matter.

do you happen to have a macintosh?

He has a whole web site on Mac emulation.

Look at it at the "web page."

BTW, I am not a Mac fanatic. I don't own one. I don't actually use a Mac emulator. But I did try one once, and that's how I found this site.

Not anymore. The repo man showed up at my door and told me he was either going to take my G4 or my HP tower, and I knew I could emulate the Mac OS faster on my HP, so, away went the G4. I plan to buy a new one next summer though.

I'm all for the idea of a new Rom. There are several immediate questions that I have about the idea, though.

1.) How are you going to program many of the stats features found only on TI calculators? Mathematicians come up with algorims for off-the-wall functions such as sinusoidal regressions and later sell their ideas to either TI or HP. The companies
then close the source and add the formula to the calculator. I don't know when you'll ever need to use log regressions and the like, but for some people, these features are important.

2.) How will you ever get past TI's limitations?

3.) Are there any matheticians/math majors that are willing to work on the project?

4.) Will the new Rom be easy enough to program for that existing programmers will want to work on it?

5.) If it's a free Rom, will that mean it can be freely distributed for use on VTI?

Finally someone with an insight! :-) I have my own concerns about this project (read more way above). It seems that too many projects get started but never finished. My suggestion: Make it small at first, then add things to it.

About the mathematicians: Yes, of course they are needed. I made my own EOS once in C++ in Computer Science class, but that is only enough to graph functions, provided that all the functions like sin, cos, tan, sqrt, and so on, are all implemented in the OS (not an easy task, either, for so many).

One thing, though: Let's make a Pretty Print expression editor for the TIs, and make it easy to port to the new OS. That would just be so cool, and a lot easier to use than RPN.

Why don't we reverse engineer AMS 2.01 or 2.02?
TI claims they didn't write it, so they shouldn't have a problem with us decompiling it.
hehehe
Or someone else could do it in another country without all of the US's stupid laws :^)
If it is going to be done, then it needs to be organized.
How about setting up a website such as www.89os.com
It might be easier just to help TI write AMS's
Or we could boycott them by shuting down all of our websites so that people would get pissed that they couldn't get any games or other appz.
Does anyone know for sure that it uses a 512 bit key?
That doesn't seem right considering that it checks it so fast (unless it is continually checking while it is receiving) and because it is sold outside the US.
Don't know about the rest of you, but I just want that damn SDK so that I can know about the system better, and so that I can sell my appz ;^]
Imagine how much money Patrick Davidson would have if he got $1 for every copy of Pheonix!
Just my $.76

BTW My calc crashed during the 2.04 upgrade, but after 100 tries I finally got 1.00 on it and then upgraded again. After all that, I found that my French language APP was still there! I was amazed that it didn't get deleted by AMS 1.00, but AMS 1.00 didn't recognize it, so it must be insome special part of memory or something.

This is exactly why a lot of people hate the SDK: having to pay for programs. That's against everything the TI-Community stands for.

- Jeff

Selling quality programs for nearly nothing is what I'm for.
I would be glad to pay for programs that I now are good.
That is why I suggested that TI allow us to write free Flash APPS for free, and also allow us to write "4 sale" ones for free, but so that we would have to sell a few copies before we receive any money.
This way, no matter how old you are, you could get a small amount of money to upkeep your computer or offer your program for free if you are rich and program for %100 fun.
I don't think it would be very bad if we sold our programs for $1 because we would all be making money and could buy each others programs.
As for open source, we could leave our programs open source because who when you offer a product dirt cheap, most people pay.

Cavan

Oops
That last sentence is: ...because when you offer...

Pheonix is fun and all, but I would never pay for a TI game. Also, with the small sizes of calculator programs, imagine how much pirating would occur. I can see it now: "WaReZ for all of your needs - Windows NT, The Sims, Quake III, and of course, solitaire for the 83! Please support our site by clicking on some of the links on your way out!!!"

You can't be serious!
You wouldn't pay a dollar for a program that you know is good?
Another though of mine is that if we do sell games then TI would need a LOT more on the site than they have now!
It would need a minute long screen shot and reviews...

Cavan

Paying for programs, or apps, whatever, is a new idea for TI people. I don't hate it, but I've got a strange feeling about it. If you were to release your app for $1 or so like you said, I would get it, but it would be different than others. Mostly, I would want it to be bug free. This means REALLY bug free. Phoenix is not bug free. It may be awesomely made, but it isn't always perfect. If I paid for the program, I would expect the level of bug free where I couldn't crash it if I tried. If I broke my calculator in half, it would still be there on the disconnected LCD, functioning perfectly :). I would just want it to be really well made, and then sure, I would pay for it, knowing it was a very good program.

If it's not, you just released a program that people won't want to pay for but must, and you released it to a bunch of people who know what they're doing. Little vampires sitting there drooling with their disassemblers and hex editors ready to release the $0 version. Hehe :)

- Jeff

Of course it would have to be bug free.
TI wouldn't put it on their site if it wasn't 99.9% sure that it was bug free.
And quit talking about Pheonix that way! The only bugs I've seen in it aren't because of itself, but rather due to TI's disclosure of system info which causes the kernal to freeze.
I know the SDK is a new idea, but at least we'll learn more about the system :^)

Cavan

BTW read my other comment about open source. Even if you release your source so that anyone can have a free version, they wouldn't be able to make it into a flash APP, and most people would pay the $1 anyway.

I would pay for Street fighter if it was updated and totally bug free. It rules!!!

I wouldn't. No offense to the author, he is a very skilled programmer I'm sure, but the game isn't really THAT good. I would probably pay for Phoenix; its better than most PC Galaxian clones.

- Jeff