Didn't check whether or not I said it here, but I may have been jumping the gun when I said that "Replay" would be a standard game option. The way I see it, if it were to work out that way, the game would have some insane system requirements.

If I had it my way, 32 or 48KB of FlashROM would be for my game. ALL YOUR RAM would be needed to keep track of replays, high scores, and the temp file. Currently, the calculated number sits at a cozy 22520 bytes of RAM needed to accommodate all of what I want to keep track of.

Cutting the "Replay" option would remove only 9000 bytes from this calculation. I'll need to recalculate the memory requirements as I find out what I need to do to keep track of the scoring file.

mhmm
I wonder if there is a way to rename APPVars? Such program could be useful if for example I wanted to keep multiple replays in my calc archive to watch later

I remember putting Bubble Bobble's appvar into a group a long time ago so I could essentially have a loadable save state. (Last boss battle was epically hard.)

The replay files are huge. The game will swap in and out of FlashROM numbered replay files. Know that a single stage will have up to 9000 bytes for its replay. After each stage, the game asks if you want to save the replay of this stage. If you do, the game sets itself up to stash away the replay file and create another to continue recording. While the replay files are numbered, embedded in it is "who" the player is and at which stage the replay file is for, which is what the game will be reading as you go back to view the replays.

Now... will it get annoying to the player if they're asked whether or not you want the replay to be saved? Keep in mind that the replay ability stops the moment you have to use a "continue" in a run, as was in the original game...

Had forgotten to reset the environment while I was spending time developing. The new script engine shows some of its power in the screenshot linked to below this paragraph. Had to turn collision detection for yourself off to demonstrate this.

http://www.mediafire.com/imageview.php?quickkey=kz5z2itz0zy&thumb=4

Really amazing work Iambian!

On the topic related to the image, I managed to get the arctan function (the stuff responsible for aiming) to be scary-precise, well within the 127 degrees of freedom allowed by my system. The only problem is the accuracy: It seems to be generally off to such that it seems to be aiming at the top-left corner of your character sprite. Given some time, I'll probably have a fix, and a screenshot ready to show it off. Until then. Later.

EDIT: Also noticed that the incorrect angle seems to pop up at semirandom intervals (mostly along the axis) at certain extremes. I'll have to explore what the problem might be, but I believe it might relate to incorrect values being added from the table in the approximation stage of the algorithm.

dark this is so epic, awesome work as always. I wonder if you should post videos of it on youtube, like Calc84maniac did with Mario and the emu?

Convert this to a GB cartage...NOW! Amazing bullet effect!