Talk:Petit Computer 3/@comment-14600979-20140920032351/@comment-15296152-20140920053807

Sure wish they'd publish them on the website, like the entire list of command + parameters.

Now, I have a problem with the latest development. Initially, I was happy for local variables in user-made functions, but now it feels like a curse because, from what I understand, interaction is completely cut off, and the only way to transmit array-like data going in and out is to make conduits out of built-in capabilities, like background layers (using BGGET and BGPUT). The JUMP program does this in its LOADSUB function. It loads data from an array it creates preceding SDLOAD, and then proceeds to use a temporary array to load data to get sent off to layers via BGLOAD. Outside the function, data is read via BGGET, and loaded into the MAP array.

I find this to be a big inconvenience, especially since they stated in that article that people could make libraries to be used by the public to offer features that the base program does not have built-in. I'm looking at my various routines in my MM2 program, all done with labels to act like functions and work because it's all global, and yet I can't even convert those into user-made functions because of the limitation of everything in the function only being local. I can't make an SPR_INIT function to initialize all the information needed for sprite processing because when it returns, all the data is cleared, as local variables do. and making a conduit like mentioned before means I'm restricting the use of that capability for my own functions (and for others if I were to make a library out of my stuff).

Sorry for the rant, but with so many nice things happening, it's disheartening that one piece pulled from the Jenga tower makes everything crumble.