MLC: Multi-Platform Language for Calcs

[CrimsonCasio]

well, good news, it seems that there may be a 83 version soon after all, evidently burntfuse is succsessfully converting the code himself

also, it appears we may finally also have an HP programmer interested in MLC, so no HP calcs might have MLC compatibility.

[tr1p1ea]

Ahh ofcourse, i knew there was a thread here! (how silly of me).

Anyway, this seems to be progressing rather quickly ... i will check your forum for updates.

[CrimsonCasio]

lol, i took your advise and registered... i wasnt orriginally intending to post quite this much

you might find this intersting:
http://dysfunction.earthforge.com/mlc/s ... t-pics.php

[tr1p1ea]

Wow, thats nice! ... but i thought the screen was restricted to 96x64 for Compatibility?

[CrimsonCasio]

Thanks

all thoughs games there are in 96x64, or the extra parts of the screen are not required for gameplay, some of them adjust to fit the screen, bounce for instance has a single variable that you can addjust to move the minimum X value for everything so that you can center it on the screen (this will be automated once the model checking code goes in)

[kv83]

Well, as far as I can see, is the 83(+) version not even compiled once yet... you still got some problems with the grayscale and the shell, am I right? Maybe Duck or someone can help you or unkown_one, since it's rather a tough job to achieve such a thing
Also I wonder how this will work on the ti-83+, cause your documentation says:

|| The run speed on some calcs has also
|| been artificially dropped to match the lowest speed
|| we've done this to make sure all games run equally
|| well on all calcs.

This means that, as far as I can predit that, ti-83+ should be one of those minimum calcs. Conclusion of that is that you have either to speed down all other calcs, or step away from the above statement.

It's nice though to see something going on there, and I am kinda intrested in this, and will of course try it at least once it's released.

[tr1p1ea]

I thought the_unknown_one was porting it to the 83/+?

[the_unknown_one@bro's com]

@kv83: No its not compiled yet, burntfuse already made a greyscale routine but i assume its not good enough, as he is planning to use Kingma's. For the shell, it is not made yet, and i'll take a try at making it, with help of madskillz and burntfuse.

@tr1p1ea: I was, but i had to go on vacation and my comp broke (well harddisk is fucked up) so i couldnt, burntfuse started porting it, but i'm back from vacation now and i'll have my comp back tomorrow, so i'll try making the shell then

[CrimsonCasio]

kv, your probably right about that, and we'll play around with it some... im hoping that we'll just be able to squeeze every last bit out of the calc so that that dosnt have to happen (like burnt is doing on the 86), but in all honesty i think we would centralize that statement around the AFX and 86 models and drop the speed slightly on the Voyage and others if needed. the 83 would get the short end of the stick on that but im sure you can understand why we would choose to do that.

[Duck]

Is there any progress on the grayscale part of the ti83(+) implementation? This week I have plenty of tests and assignments for my study, next week I'll be on vacation but the week after I'll have time to code some again. If you guys cant figure out the grayscale I could do that part.

What are the processor requirements for MLC? How well does it run on the ti86?

[the_unknown_one]

It runs very well on 86 Processor requirements arent high, we'll maybe make it for 82/83/85 too :p For the TI-83+ version, burntfuse starting porting 86 version again, cause he has updated it alot already since he made first 'pre-port', so i'm waiting for that...

[Duck]

Ever considered combining the ti86 and ti83+ version into one source file? You'll always have the latest version for both.

Also: at how much % of minimal speed does it run on the ti86? I'm interested because adding grayscale adds huge overhead.

[the_unknown_one]

dont ask me man, i dont have a clue, i just know it runs well :p

[Duck]

I've just looked at the code of the ti86 version and played some games. It does quite well.

But what stunned me is that games are in the form of source code... Wow, even the simple game Phalanges takes 7k of code (!). Are there plans on making some sort of source-to-bytecode compiler or something? Its nice to be able to edit programs on your calc, but end-users dont need that feature.

[CrimsonCasio]

duck, we'd be happy to accept any help your willing to give

in fact your right about the code size, Ive suggested that we include a way to store them in a compressed format (bytecode), odds are that this will be implemented on the 83 and hopefully the 86 (it probably isnt needed for the other models, the hp version can run MLC games up to a gig in size and MLCafx can do something like 64k). in addition, we're working on code reuseability by enabling pointer variables, that way large and complex functions only have to be written once, instead of duplicated for each set of data they have to manipuate (this is why many MLC games are so big, currently you simply cannot do this), also by experimenting with our current set of demo games we've uncovered a few tricks that we decided to build directly into the language and thus increase performance (sprite velocity being a good example).

as for the source code, im looking into setting up a CVS server on sourceforge for the C versions of MLC, another could be set up for the 86 and 83 if needed.

also, id just like to say once more that burnt is amazing, he keeps on pulling space out of thin air every time he runs low... its quite awe inspiring by the time hes finished i wouldnt be supprised if MLC86 was the most optimised 86 program in existance...
(if you see any more ways to optimise it im sure he'd be gratefull to know, he needs every last byte)

thanks again for your intrest and support