Lunch was extraordinarily hectic, so I will be posting code after school. I'm on one of my school's pathetic Macs right now (which, incidentally, doesn't have TI-Connect - imagine that!) so I can't transfer code right now. (I suppose I could just type it out, but I have better things to do with my time. )

Never use Y variable. Some TI-OS glitches causes it to be reset to 0 from time to time. You'll run into problems later.

I think ztrumpet is right - the Y variable automagically gets set to zero whenever you display the graph screen. However, I don't display the graph screen while I'm moving except if I go into the menu (during which I temporarily save the value of Y somewhere else). I've already run into problems with that in which I moved up a screen whenever I exited the menu. It was fixt rather fast.

ztrumpet, it made sense. I need to test that. If it's true (and not that I'm doubting you, but I like to test these things. There should be documentation for that somewhere, or else I can't read), then it would be an interesting speed-for-memory tradeoff (and a good excuse not to use xlib).

Okay, thanks for xcopy syntax.

@ztrumpet: Not according to TI-BD, it isn’t: http://tibasicdev.wikidot.com/timings#toc3

The If:X structure is only considerably (20-fold) slower if one uses a For( loop, of which there are none in my movement engine. Otherwise, the speed is comparable (and even slightly faster).

And aahhh... xcopy uses the X variable. This could be slightly problematic, as my movement is stored (temporarily) in the variables X and Y before these variables are copied to L4(8 ) and L4(9). (Justification for this is the enormous toll saved on memory – X and Y take up one byte each, while L4(n) takes up an enormous five bytes. The default list variables take up TWO bytes each, which is extremely irritating.)

Does xcopy have deletion capabilities for the programs as well, or will I have to use another ASM library, since I don’t think I can manage otherwise? This is quite the interesting collection of ASM code I’m assembling (pun completely intended)

I just tried it without xlib, and it didn't seem to run any faster. When I counted pixels there was a tiny difference, but not really enough to matter.

I would love some help with this devilish piece of code, builderboy I'll post it tomorrow (when I hijack my friend's link cable) but I'll just give some background information now (he brings his laptop to school and we only have lunch to do this, so I won't have time to type it out then):

• In [A] is the "world" map for each portion of the game, and the element stored in a place in [A] gives you what map loads. For example, in the first zone, [A] is a 2x4 matrix with [[4,5][3,6][2,7][1,8]]. You start in room 1, and which room you progress to after that depends on which side you exit from. (i.e. if you exit to the up direction your y-coordinate in the matrix increases 1).
• I'm not going to include the menu or teacher mode programs - only the map file, the event file, and some other necessary stuff. So don't press 2nd unless you want an error.
• All the checks for stuff in L6 are completely necessary. (They're your status ailments.)
• Whether you can move or not is determined directly from the string, seeing if sub(Str1,16(Y-1)+X,1 returns a valid string that you can move on. This is probably the slowest part, but I couldn't think of anything else unless you want me to store everything to a 128-element list which is extremely memory-inefficient.
• You need xlib enabled. (Duh. )

Thanks for all the help and support so far, guys - I'm going to take a break (and do my homework) tonight! See you all tomorrow at noon

E - wait, do you guys just want the code or the actual files themselves to see how fast it runs? Code would be harder to do (my friend doesn't have a program editor on his computer).

Yeah, seq( is fun.

I forget many times the second argument and skip it. Or put without wanting the last optional argument for increase.

Whoops! Much headdesking happened there?

Unfortunately, I found a problem - a very major problem - that does not require the use of the seq( command.

While writing my movement engine, I somehow neglected to check for wall collisions. Basically, if a wall collision occurred, you would just get your x and y coordinates moved backwards and nothing else would happen. This would not be a problem, except when I implemented my code to call a fight, to take damage from poison, etc. you would still get poison damage and your blindness counter would still tick if you tried walking into a wall as well as regular movement. There's also the chance that you can get a fight after opening your menu.

Obviously, this sucks. I need to re-do my entire movement system, and being extremely busy right now (teachers like giving stuff before break ) it's probably going to take me a few days.

On the bright side, I drew out all the equipment (56 in total) and finished the inventory screen. It also uses some fancy list manipulations to prevent lots of If statements (I believe the code is H+sum(Ans{100,1000,99999,0,0,0,99999-->H. Basically Ans is a list containing which item you used, and the list it's multiplied by is the amount each item heals).

More to come later! After I get this mountain of homework out of the way, that is.

Sorry, guys - no progress today except for laying out inventory screen (items are visible but you can't use any of them yet). I was super busy - had a concert this morning, made about $300, took a friend (and my page turner) to lunch/dinner, and had to do homework. I hate weekends

Anyways, the functionality of the binomcdf( command can be improved by the cumSum command. I forget the exact syntax, but I believe it gives a list of {1,2,3,...N}.

DJ, just think of seq( like a For loop - you enter a seq(function,variable,start,end) and it iterates with the variable the exact same way it would with a For loop.
:For(X,1,5
:X^2 -> L1(X
:End
:L1would do the same thing as
:seq(X^2,X,1,5
I like seq( since it doesn't actually change the variable. And it's easier to write, and it's probably faster. Also, with seq(, you have your output directly in Ans, whereas with the For( loop you need to either store the elements to some list or use the augment( command.

I'm just not going to try to explain the binomial commands, though. That's a whole semester of Stats right there.


E: aargh, right - superscript tags don't work in code boxes. Silly me.

Wow, nice. My opinion of TI went up a little.

Yeah, I learned about the binomial model in Stats last year and am pretty familiar with the mathematics of binomcdf( and binompdf( - they're probability density functions. But my problem with that was that you couldn't use them to manipulate list elements without a For( loop, and seq( is faster and easier to program than a For( loop, so I used seq( instead.

Loliloli! -Wait what?

I'm looking forwards to playing this... and finding my link cable...

Epiphany!

I managed to make it non-buggy, relatively fast, and awesome, while using only 1/3 the code I used before. (seq( and augment( commands abound, as well as some complicated boolean thingies.) Second time's the charm.