Two comments -- no, three -- all closely related:

First, I was looking through the VTI source very recently. Not to discredit Rusty for his fine end product, but the source is a _mess_! To someone not experienced with the program's bowels, it appears to be a loose assortment of loosely related programs churned together into something that, suprisingly, works. Okay maybe that was kind of a rant. But I was trying to use some parts of the source (specifically Z80 emulation and the TI-8(1,2,3,3+,5,6) architecture code) and I can't figure out where to start! I'm sure Rusty can bang off the exact source files I'd need, but the point is that it should be no big deal to figure it out. Comments? You know, the stuff that begins with /* and ends with */, or starts with // and goes to the end of the line? Huh?

Okay, sorry. It's not _that_ bad. But I'll let you know that I wouldn't write code like that.

Second comment: This is very closely related to the first. I'm working on a TI-83 & co. emulator for the TI-89 & co. I'll be based on VTI for the architecture-specific stuff, but it'll be quite slow, since the processor must be emulated. To solve this problem, I'll slowly rewrite the emulation core to dynamically rewrite at least some of the Z80 instructions into M68k instructions, which will execute much faster. I say slowly because this is a lot of work for someone inexperienced with assembly. But see Bochs for how it can be done for the x86 (and they say 68k ASM is easier than x86 ASM). If anyone else thinks that this is a cool idea, contact me at kcarnold@yahoo.com and we can work together.

Third: port uClinux to TI-89 / 92+ (most likely the latter first). uClinux == Linux for the M68xxx without MMUs. Exactly what I / we need.

Kenneth

(Nick and the rest of ticalc.org: a preview function would be nice. I want to see how my comment looks before I post it! Anyway, here goes nothing...)

Yea, the source is a mess, I'll definitely admit to that. The project began as a small tool to help me debug and expand PlusShell, so I wasn't really concerned with how "nice" the source was. It started as a simple 89/92+ only emulator trown together in DJGPP, and evolved as I got more ideas. In fact, the 82-86 support started as a joke with one of the other ACZ members, and I decided that night to start on it just for the heck of it.

VTI may have been downloaded a lot, but ZShell is a far greater achievement. I still don't have any idea how those guys could have possibly figured out how to write ZShell in the first place, I know I couldn't have done it. They are the ones that deserve all the fame, without their work we would still be stuck with slow TI-BASIC. Their achievement is truly amazing.

Well I do believe Virtual TI was a great achievement as well. When I was programming for the TI-85 and TI-92, it wasn't viable to use an emulator to program. I can't count the number of times I had to remove the batteries and reload... :(

I'm really glad I added that news update without reading the comments. :) When I said that VTI was a marvelous feat, I meant it.

My TI-85 was stolen from a friend who borrowed it for school (and he only had it for two periods). So I waited until summer before I got a job and saved up $116 to buy a TI-86. I only ever used ZShell until Usgard came out (I tried--but didn't like--the others), and even today with my strong ties to Usgard it is comforting to have VTI on which to run the old TI-85 programs.

Especially since there was never a way (that I know of) to view Usgard 4-color grayscale pictures on the TI-86 (I had a couple pictures of friends; they got a kick out of it).

I like your comment about Rusty's code. The two of us have discussed that topic quite often. When he can bang out source ten times faster than me, and it works, does it really matter if it is clean or organized? As long as it gets the job done and works, it shouldn't matter. I never code like that, simply because I'm not that good, but it works for him. At the same time, I have seen code that he has written to be shared with others, and it turns out quite nicely :)

Don't you think maybe there is a reason it's a "mess" as you call it. It's style, as in all programming. My friend writes the sloppiest programs I've ever seen, but they work surprizingly well, better then mine likely, and my programs don't get much more organized. He has the exact same viewpoint about my programs. The author sees what he needs. Ever noticed how hard it is understand a massive program you didn't write? Say you wrote a massive program though, and needed to make a change to it. You would know EXACTLY, where to change it and how to change it. You create you're own program maps in your head so it makes sense while you build it. A nonauthor of a program doesn't have it and must study the program quite some time to build even a fairly accurate program map, and then even longer to discover WHY the author chose to do it this way.

Yeah, I know -- I've looked at a lot of other programs also. But VTI just stands out as being quite disorganized. Quite possibly this is because there is only one maintainer. Programs that are distributed among groups of several people tend to be much more organized.

I write very neat code, and I clean up what doesn't look neat with indent(1). I haven't had anyone else really read over my code, but I comment a lot (not excessively) and thus program slowly. And yes, you do get quite a feel for the structure of a program that you wrote yourself, something someone who did not write it finds very difficult to achieve.

But still, could have been neater, for benefit of myself and others.

About the emulator idea - drop it. You see how slow VTI is, so no matter what you do, a 68k is too slow.

And BTW, I think x86 asm MUCH easier. I mean move.d, move.l, GAGH! It just sux. A simple mov is the way to be!

Uhh, I think that VTI is a great idea. It wasn't meant for people to use their calculator on their computer and not on the calculator (you even need a copy of your ROM). I think it is more for debugging applications. I haven't done too much with assembly as of now but I'm sure that it would get irritating to reset your calculator over and over again, because of bugs.

On a side note, did anyone else notice DoorsOS v0.96 crashing more than DoorsOS v0.95, or was it just me? I even tried redownloading it and not using any other programs, and it stilled crashed my calculator (try using it for a little while then graphing (x^3*y-y^3*x)/390 in function mode..it should normally produce an error that it is the wrong mode but sometimes it crashes in Doors).

I'm pretty sure he was talking about emulating the Z80 calcs on the 68k calcs, not VTI--of course VTI's a great idea!

note: the best way to avoid crashes doing math like 3 sqrt(6) is to just uninstall doors before using math. Teos seems to have less of a problem about crashing with math, but thats only me I believe

Use TeOS :)

Cd_Slayer

No, VTI is great - couldn't get my proggies bugfree without it! I meant emulating Z80 calc on 68k calcs, which is a Bad IdeaŽ IMO, due to the limitations of the CPU.

SLOW?!? I have a P133, and these are the speeds i have achived with VTI (as % of normal speed):

82,83,85,86-upwards of 300 or even 400% at times
83+-ok, so it didnt even work.
89-120%
92/92+-not quite 100%

If its slow, you either have an even crappier computer than I do, or you have no memory left.

400%? How? I've never been able to get VTI to run at even 200% of normal speed. Why does mine run at normal speed when I deselect "restrict to normal speed"? My computer has the same processor as yours.

Everybody, quit praising ZShell so much. If ZShell hadn't come along as the first assembly shell ever, i'm sure a different one would have.

Zshell is just an old shell that allows people to run assembly programs on the ti 85 graphing calculator, but VTI is a program that can be used by any platform programmer for any reason they want. Without VTI, it would be harder for programmers to come up with games for you people using ZShell (unless they had CalcEm or another emulator...).

What i'm saying is that Zshell is an old program that had its time already. Let another great program, like VTI, come into the spotlight. VTI is going to keep helping programmers of all kinds for future generations until another program comes and passes it. Let's all praise VTI now.

Long Live VTI!

Yes, another one would have. The question is how late. If ZShell hadn't been there, we might really be stuck with BASIC. Of course, you never know, but just think of how its makers figured out how to program it. It's just purely amazing. (delicious? - no, wait, that doesn't sound right!)

You cannot guarantee that another shell would have come out. Besides, if another one had, then we should praise it as much.

You say that ZShell programs are coming out because of VTI. That's ridiculous, no calculator would have assembly unless it had become popular on the TI-85. It was the only boat in town for a long time. TI obviously saw the excitement and interest it was bringing to calculators and added it as a feature.

ZShell not only was responsible for the calculators that have built-in support, but also for the TI-82. Without the TI-85 as a model, the TI-82 would probably not been hacked for much longer. The mechanism of executing Z80 instructions on the TI-82 was derived from the same method discovered first on the TI-85 using ZShell.

I don't know if you're actually serious or not. If you are I really question your mental health.

First, you don't know that a replacement would have come out. Even if it did it would probably be several years behind ZShell.

Second, you can't possibly expect VTI to lead for that long. If it does it going to need to make advancments nonstop to its program to keep it up to date. Do you know what an NES is? When it came out, I can almost garantee that the creators said exactly what you just did. Now its as obsolete as a slide-rule. If VTI does continue advancements as it is doing, it will do fine. It will NOT though rule unchallenged. There will be times where it is not number 1.

>First, you don't know that a replacement would have come out. Even if it did it would probably be several years behind ZShell.

and how do you know there could not have been a shell before ZShell? either scenario, we'd be praising it

>Second, you can't possibly expect VTI to lead for that long. If it does it going to need to make advancments nonstop to its program to keep it up to date.

Of course not, and that's not what he said. It's the next natural step, as someone blow has said.

And everybody, please, try not to be so insulting.

Cd_Slayer

I didn't intent to be so insulting. Sorry if I said anything to hurt anyone's fellings. I don't know what you have a problem with on the first part. I know we'd be praising it, I didn't mean to suggest that we wouldn't praise any suppliment for ZShell. When I said "you don't know that a replacement would have come out. Even if it did it would probably be several years behind ZShell.", I think I made it clear with "would have come out" that I was talking about when ZShell came out as if it were the first, nothing before it. If one had come out earlier, fine, that's not what I was talking about. Yeah we would praise it.
As for the second thing you mentioned, he DID say that he hoped VTI served many future generations, and I was telling him that was unlikely. Again if I hurt anybody's fellings, I as very sorry. I didn't mean all that garbage that he was crazy or whatever, I was just very frustrated with some of the things people have been saying.

Sorry, you weren't the one nagging anyone, I just wanted to put it all in one post... Anyway, I uderstand your frustration because some of the opinions are so one sided...

About VTI... I do believe an emulator is always a big push, mainly because it helps debugging tremendously, which usually takes up more time than building the prorgam, or at least for those of us without a high gift for programming. Anyway, probalby not your comment I should've replied to.

---
Cd_Slayer