Simply astounding. Well done!

I'm gonna be taking Java next year in highschool, so I'll be able to use this maybe since I'm too lazy to learn C. I don't like java though because it's slow and inefficent like old people making love.

I have heard a lot of people who are scared of learning C. It really isnt very hard. I learned it very quickly last fall, and all the programming experience i had was a tiny bit of java and some TI-BASIC.

Frankly I like writing in C better anyway, I use Java for school and work, but I don't like object-oriented prorgamming, really.

I also don't care much for OOP. Not very many problem types can be best expressed using OO.

Besides, designing programs to be data-centric is probably the best way to write good programs and eliminates much of the advantages of OOP anyway (see URL for "The Art of Unix Programming", which explains data-centric design).

I do like some non-OOP related features of C++ that are not in C, though more recent C standards and compilers support them.

I don't particularly like OOP either, but I can see where it would have some advantages. I still think there isn't anything you can program with OOP that you can't program without OOP.

You can't, Assembly isn't OO at all, so everything ends up being not OO, the only thing that makes it OO is the way you program it in. Most OOP langauges also alow you to program in !OOP mode (sorta, kinda)

I don't particularly like OOP either, but I can see where it would have some advantages. I still think there isn't anything you can program with OOP that you can't program without it.

That's true because you can convert OO code to nonOO code automatically. But using OO has advantages when your programs become really big by making it more structured. But for calculators with relatively small programs using OO doesn't give much of an advantage.

The main advantage over imperative language is that OO supports a higher level of abstraction through mechanism such as polymorphism.
Any way, Moka offers complete interoperability with C (C statements can be included in native { } blocks). Moka being a ‘translator’, these statements are simply exported to the final TIGCC project. Moka is a port of the Java programming language, not a port of the Java Virtual Machine. This means that Moka, although many effort has been done to preserve an API similar to the Java API, is adapted to the TI-68K developing realities such as optimization of size and speed.

Anything like & pass by reference in functions? That would be cool. Of course, that's not native Java, that's C++, but I wish it was everywhere.

That can be done using 2 way in Moka : the 'right' thing to do is to define classes and use objects (which are passed by reference), the 'bad' thing is to use C interoperability and proceed just as in C (that's the good thing in C++ I think :D).
Note that OO language, through polymorphism, has a more ‘friendly’ way to pass function by reference than the pointers to functions in C.

I forgot to mention that special classes exist in Java (Boolean, Integer, Long, Double, …) that wrap primitive values so they can be used to pass by reference those primitive type.

What you didn't mention is that the "right" thing is horribly inefficient compared to the "wrong" thing!

Using & to pass variables/objects by reference is purely syntactic sugar. It's identical in every other way (AFAIK) to passing a pointer to the variable to a function.

OO and imperativeness are two independent concepts. Java is an imperative OO language...

It's true, even SmalkTalk, in which everything is an object, even structures (loop, conditional, name it), is also an imperative language. I misused the term. I wanted to talk about 'traditional procedural imperative languages'.

Here's my question: Will there be new Moka directories, or is there any way to tell/any performance difference?

This probably would only be necessary if a different oncalc compiler was used (i.e. MCL programs wouldn't work in the same way as BASIC ones, and definitely would require a different directory).

Moka generate nostub programs. Once compiled, they are similar to any other asm prog.

Hi Mr. Quésoft, i have a question.
when i try to run some of the examples, it says 'ASAP or Exec string too long'. i'm running them on VTI, with AMS 2.08. its usually the ones that use the mouse, but some of those do work.
does anyone know what s wrong?

I think the AMS version 2.08 prevents programs larger than 8K (or is it 16K) to run. As the graphic environment induce a big overhead, it is logical that the windowed application are the ones that don't work.

I beleived that the EXECUTE_IN_GHOST_SPACE directive (which is enabled by default in Moka programs) prevents this, but maybe I was wrong. You may need a launcher to run larger programs :(

I'll work this out and post it back ...

I did find an utility called hw2patch developped by Julien Muchembled on ticalc.org and it should work. Whoever, i tried it with AMS 2.09 and it is incompatible (the only version supported are earlier than AMS 2.05). I hope someone know a fix that work for the newer AMS version ...

hmmm, is there a way i can get AMS 2.05? or does TI keep the old ones in a inpenetrable fortress on an unknown island in the pacific ocean...

Master Chief _is_ real. He's owned by TI and they keep him on their unpenatrible fortress island to guard the old AMS versions.

yeah, haha. ti is making their own secret government and military.

oh, i got the files to work by using tict explorer

HW2Patch is NOT the solution to your problem.

EXECUTE_IN_GHOST_SPACE is NOT a solution to your problem. What you need to do is to generate a launcher. The easiest way is to just ExePack the program. Use the -pack switch. See http://tigcc.ticalc.org/ doc/comopts.html#SEC3 .

Hey this is a bit off topic, but anyway

I got a broken TI-83+. It actually works perfectly (I can send and receive files to the calculator from the computer or another calc), except for the screen is bust. Nothing shows up on the screen. I opened it up, and the thin plastic tape-like wires going to the screen was cracked (it was folded inside the case, like usual). Anybody know how I could fix this?? (ie, where I can buy another set of wires?)

It works ! Great ! It is a very useful thing to know... Smaller programs, which runs :D

I think the next release of Moka will offer support for ExePack. But for now, developpers can modify the options of TIGCC before compiling of the default TIGCC option in config file tigcc\conf\tigcc_project.default of the mdk directory, adding the lines
Pack=1
Packed Variable=test
at the end of the file.

Thank you for posting.

NB : copy paste error, 'test' can be replace by any variable name that fits your needs :D

BTW, I had always wondered how to exepack programs. Thanks.