That is cool! I like where you are going with this :) On Wed, Feb 7, 2018 at 7:42 AM, Ralph Versteegen <[email protected]> wrote:
> > On 7 February 2018 at 14:23, Ralph Versteegen <[email protected]> wrote: > >> I actually already started rewriting cursor movement yesterday, but >> didn't finish. I haven't really nailed it down yet, but can just make a >> first pass at it and then we can tune it. >> My ideas a fairly similar to yours, except that instead of an iterative >> search for a slice that matches a certain criterion, I just assign each a >> score (metric) and pick the one with the best score, like it's already done. >> When using a metric, it's useful to think about the isolines: the lines >> along which different slice positions would have equal goodnesses.I'm >> > > I experimented a bit with some alternatives, and selecting planks which > are sharply off to the side but close to the previous plank was a constant > problem. Which makes the battle cursor logic seem quite sensible (although > I didn't copy it over to plankmenu and try it). I would actually be easy to > adapt it to rectangular planks; I found just using the point on the edge of > the plank facing the previous cursor position and closest to the previous > cursor works pretty well most of the time - in other words, reducing all > other planks to single points. > > I think the isolines we want are something like these: > > > > This is similar to what you said about checking for planks within a > rectangle or within a cone. > > > >> >> BTW, I rewrote how cursor movement in battles worked, a number of years >> ago (battle_target_arrows). I used something very similar for that, except >> that the sizes of enemies are completely ignored. We weren't totally happy >> with how that worked out because the cursors might skip over a target. But >> I wrote it in such a way that (I believe) it's impossible for there to be >> unreachable enemies. However, it's still possible that you can't select an >> enemy from one direction and have to loop around from behind. >> It works like this: >> -Find all targets within a 90 degree cone from the current >> -Pick the one in the cone with the smallest distance along the 'move' axis >> -If none, pick the target that's on the right side of the target with the >> smallest angle to the 'move' axis (will be more than 45 degrees) >> >> But for grid-like arrangements, using a 90 degree cone like that is >> definitely going to have bad results. >> >> However, when you take sizes into account, it may be impossible to avoid >> unreachable slices. Rather than autodetecting such placements it might be >> easier to just let people test out their menus directly with a test mode in >> the slice editor. >> >> We could maybe replace the battle cursor movement with a plankmenu if it >> works alright or better than the existing, by just ignoring the enemy sizes >> to avoid the unreacble problem. The only thing needed is to special case >> spread attack selection, which happens when you press left/right and there >> are no more enemies. >> >> Regarding diagonal movement, I could do that. Problem is that you usually >> don't press both directions at the same moment, so I doubt it will be very >> useful. (But the keys will fire together thereafter). >> In fact I bought a USB adaptor for my old PSX controllers a few days ago, >> so now I can even try out an analog joystick, and possibly allow moving at >> arbitrary angles! >> >> >> >> On 7 February 2018 at 05:25, James Paige <[email protected]> wrote: >> >>> >>> >>> On Tue, Feb 6, 2018 at 8:22 AM, James Paige <[email protected]> >>> wrote: >>> >>>> >>>> I think the only good thing about the current method is that it should >>>> be impossible to have any unreachable plank, but its other glaring flaws >>>> outweigh that by a great deal. >>>> >>>> >>> Thinking about that more, I think my logic is probably wrong on that >>> too. ;) >>> >>> The current method sucks no matter how you slice it! (pun intended) >>> >>> --- >>> James >>> >>>> >>>> >>>> On Tue, Feb 6, 2018 at 6:45 AM, <[email protected]> wrote: >>>> >>>>> teeemcee >>>>> 2018-02-06 06:45:13 -0800 (Tue, 06 Feb 2018) >>>>> 90 >>>>> Add a screen for testing plankmenu cursor movement to the spam menu >>>>> >>>>> ...it sure is broken! >>>>> --- >>>>> U wip/custom.bas >>>>> _______________________________________________ >>>>> Ohrrpgce mailing list >>>>> [email protected] >>>>> http://lists.motherhamster.org/listinfo.cgi/ohrrpgce-motherhamster.org >>>>> >>>> >>>> >>> >>> _______________________________________________ >>> Ohrrpgce mailing list >>> [email protected] >>> http://lists.motherhamster.org/listinfo.cgi/ohrrpgce-motherhamster.org >>> >>> >> > > _______________________________________________ > Ohrrpgce mailing list > [email protected] > http://lists.motherhamster.org/listinfo.cgi/ohrrpgce-motherhamster.org > >
_______________________________________________ Ohrrpgce mailing list [email protected] http://lists.motherhamster.org/listinfo.cgi/ohrrpgce-motherhamster.org
