My GM shmups - Last Chance,MEGATANK, Flying V, XYX, ZPF

[BPzeBanshee]

Also, why the hell would you have a giant D-Pad on an iPhone for the converted program? I'd just make the entire movement touch-based like Cave games with some speed limited, seems to work just fine for them.

I don't know how detailed you can get with the touch controls, that's just one method they recommended on their iOS/Android base-line spec (which has now vanished, apparently.) Something where your ship follows where you touch seems feasible too; but I don't know precisely what all is possible.

Assuming this ever gets past simply bullshitting about it, I'm imagining the game itself will be drastically different... if for no other reasons than having a teeny screen to work with, the player likely having a shorter attention span and potentially a finger or stylus getting in the way of the action.

-very short semi-random levels - same timing; but different enemies depending on certain factors. Some might also randomize difficulty for higher scoring potential.
-psuedo RPG elements: clear a certain number of enemies to reach the next 'grade', which level selection and your weapon's power level may depend on. A limited shop system (if only to swap out ships and restock bombs) is possible too.

Whether or not that makes for a terrible shooter (portable or otherwise), I don't know. I'll obviously need to see how well most of this works before I can worry about all the prickly details of having it run on something other device... again, assuming that's even feasible.

Running on another device seems possible if this page is anything to go by. Hopefully within the next year or so we may have a fully featured engine that works multiplatform. I'm looking forward to seeing a GMOSSE or Xeno Fighters for Mac as well as XYX.

Don't worry about coming up with menus or any of that bullshit until that at least comes out and gives us the opportunity to port across the board.

[Rozyrg]

Yeah, I know I should probably just focus on what I'm doing now. 2+ games at one time is enough for anybody I think. Still fun to think about at least.

I *was* wondering about what that "GM Studio" thing was, though. >_>


Death's Head LUNA changes - you can now damage her wings, and destroying both starts a new attack phase.

Difficulty settings are back, too. There are now 6, just like in XYX, although I haven't finished tweaking them. In Medal mode, you get multiple times an enemy's value depending on how high the setting is (up to a maximum of 4X.) I did this because of how imbalanced the scoring is at the moment; but it's only a temporary solution. Basically, if you collect medals and stay alive, your score can be something like 10 times what it would be if you reached the end of the level empty handed.

I'm thinking of having ones you've collected cash-in at half value when you die; but maybe that would decrease the overriding importance of staying alive, though.

[S20-TBL]

I'm thinking of having ones you've collected cash-in at half value when you die; but maybe that would decrease the overriding importance of staying alive, though.

I have a couple of ideas:

1) Instead of halving the medal values when you die, why don't you halve the total amount of medals received for the stage instead? It's a diminishing returns thing, but at least you wouldn't have to recalculate medal values, and the player will still feel the "oomph" of losing so much hard-earned gold for that level due to the visual impact of actually losing half of your medals, instead of retaining the full number you had before messing up.

2) Second, you could try doubling the total point value for the medals if you manage to make it through a stage without a single miss, as a "Perfect" bonus. I think Jamestown does a variation of this if you complete a level without losing a life. This way it won't bloat the player's score as much, and a good player's tally will have a significantly higher difference over a lesser one.

Awesome moth wings BTW. Flying V always gave me the impression that you used PS vectors for the assets.

EDIT: Cleared up idea 2 a bit. The idea is to multiply the converted point equivalents of the medals by 2 if you successfully no-miss a stage.

[SenpaiSamaKun]

1 seems to potentially make it a, dare I say it, Imperishable Night situation where the roof of potential scores far exceed practical attempts.

For 2, I would just do an large score to do multiplier for every life not used in a level myself.
That way while there's going to be big gaps between skill levels of players in score, one life screw ups won't be so punishing for non-perfect runs.
The only problem with that is that it would need different scores for life amount. 5 at start being 20000x, 4 being 30000x, and so on.

As for my own ideas, just triple toning the bullets instead of a black outline would look nice with the redrawing.
I don't know how you would keep the same visibility without major anti-aliasing though.
I should keep up my interests up in your other games too.

[Rozyrg]

Thanks for the scoring suggestions, guys.

The black outlines were kind of a last ditch effort at maintaining visibility... I tried all kinds of stuff; but there'd always be some spot where they'd get mixed in with everything else. Fortunately, they're drawn separate from the bullet's sprite.

edit :
- all player and enemy shots have been redrawn!
- fixed the medal re-count glitch (I think...)
- Also, after rewriting and debugging the Dual Shot's targeting code many, many times, I FINALLY fixed it's behavior problems. It turns out GM seems to have a problem detecting an object's parent if it has a parent as well... and that basically canceled out my attempts to disqualify certain 'no-hit' enemies from the target pool. Part of the issue stems from the awkward collision hierarchy I gave enemy parent objects early on; but it's odd nonetheless. Basically, I had multiple lines trying to get it to see if an enemy had this extra parent and none worked.

Code: Select all

//reset target if invalid
if targ.object_index=enemy_nohit
or object_get_parent(targ)=enemy_nohit
or object_is_ancestor(enemy_nohit,targ)
or targ=enemy_nohit
{targ=0}

The solution? Giving the parent a simple unique variable. Why is the solution to persistent glitches usually so simple and embarrassing? -_-

[Rozyrg]

I just added "Super Medal Mode" where an enemy's bullets change into medals when they die (depending on how many waves you've fully destroyed.) Yeah, yeah... I know no other shooter in existence has that feature. >_> Obviously, this drastically increases scoring potential; but it also lowers the difficulty due to the continuous elimination of stray bullets.

It's too ridiculous and fun to leave out; but I'm not sure where to fit it in. It really only fits with the highest 'fake bullet-hell' difficulties; but it makes even those fairly easy. I might just have a special (unlockable?) setting specifically balanced for it.

[StevenM]

I just added "Super Medal Mode" where an enemy's bullets change into medals when they die (depending on how many waves you've fully destroyed.) Yeah, yeah... I know no other shooter in existence has that feature. >_> Obviously, this drastically increases scoring potential; but it also lowers the difficulty due to the continuous elimination of stray bullets.

It's too ridiculous and fun to leave out; but I'm not sure where to fit it in. It really only fits with the highest 'fake bullet-hell' difficulties; but it makes even those fairly easy. I might just have a special (unlockable?) setting specifically balanced for it.

I remember trying a scoring system like this once. In that game all bullets on screen would turn into medals that homed in on the player when they destroyed an enemy; the background had a few secret medals that could only be obtained by shooting around certain spots; and the player would automatically drop their medals onto the playing field against if they were shot - the medals would just scatter and disappear after a few seconds unless you managed to collect most of them again. In this way the game changed from 'player gets shot, loses a life' to 'player gets shot but at least one medal protects him for a hit'. It was pretty par for the course considering the source material but for a SHMUP it made for an interesting mechanic, especially if the lack of medals actually hurt your chances of survival on top of your final score.

The point is don't berate yourself for using a system that's tried and tested. Just experiment a bit with it, brainstorm - What if I changed X? What if Y did this instead of this? - and get a second opinion from playtesters if you're really concerned.

[Rozyrg]

I appreciate the feedback. Just didn't want to feign originality there.

I tend to try out features as I think of them, regardless of whether they've been done to death already. Once I've worked out the how to in my head, there's no going back. If they're fun (and don't tax my poor laptop too much.. which I'm concerned it might in this case), I leave them in.

In this case, I definitely like it too much to take it out, I just have to rebalance the game around it somewhat. Most the challenge is built around an enemy's bullets being a persistent obstacle...*staying* there whether they die or not. Their health is scaled with this taken into consideration too, so it's mostly an issue with smaller, low-health enemies. If they can be taken out before their first volley reaches you, their threat is completely neutralized. I do have a few ideas to deal with this issue, though. >_>

Here's a quick rundown on some of the differences so far:

- Right now, you select it as the 7th difficulty setting; but Medal mode must still be the active game type.
- On the first test run, I accumulated over 700 medals.
- On a later run, I managed a score of over 60m before dying at the stage 2 boss. The cap for the first stage is normally around less than a third of that.
- Enemies fire more bullets (wider spreads, denser bursts), which are faster - even the former hardest setting slowed them down- with less of a delay between shots. Anything you let live WILL give you a hard time.
- Medals level up as you complete a set number of waves: 5 and up gives you the small kind, 35 medium, and so on.
- Dying halves your completed wave counter (rounded down)
- The counter is not reset between stages (yet.) This is mostly out of laziness; but I kind of like it like this.

[BPzeBanshee]

A thought occurred to me while reading your stuff about the bullets-to-medals system - why don't you just make it so that only certain enemies have bullets turn to medals on death?

That way, small popcorn enemies can still be a threat since their bullets can remain while larger enemies will cancel their bullets and get you score/relief.

[Rozyrg]

Yeah, I guess that'd be a pretty obvious solution.

I tried that and a few other things but I think I settled on making you have to complete the enemy formation (wave) for the bullet cancellation to take place. Things stay nice and hectic and you still have a chance to reap the rewards of all the chaos.

I had a heck of a time implementing that 'simple' change too. *sigh* -_-

[StevenM]

I tried that and a few other things but I think I settled on making you have to complete the enemy formation (wave) for the bullet cancellation to take place. Things stay nice and hectic and you still have a chance to reap the rewards of all the chaos.

That's a pretty good balancing mechanic, actually - you can't collect unless you've killed the enemies, so it prioritises clearing the screen. On the other hand, with something like a 'bust one enemy = all bullets onscreen become medals' system you could milk the enemy waves for loads of points - though that could be pretty tedious in the end. It depends on what scoring system you're going for, really - you're the expert here.

[Rozyrg]

On the other hand, with something like a 'bust one enemy = all bullets onscreen become medals' system you could milk the enemy waves for loads of points - though that could be pretty tedious in the end. It depends on what scoring system you're going for, really - you're the expert here.

It's certainly a bit crazier that way. I'd have left it the same if it didn't make it so easy to stomp through everything. I think a little extra challenge and tension is a worthy tradeoff. I wouldn't say I'm the expert, though. As many shooters as I've played over the years, I never really thought about scoring systems until I had to build them.
--
I've added another wrinkle to the system to mix things up a bit - when you don't clear a wave (a single enemy or more from the formation escapes off screen), your wave completed count goes down... the penalty scales with the increase in medal value, too. At the lowest value, you lose one, at the highest you lose 5.

edit: Also added a (very) basic rank system for this mode. It kicks in when the highest value medals appear (due to wave clears) and again when you hit the maximum number of clears. Dying is the only thing that knocks it back down again. Nothing fancy; but ok for the time being at least.

[BPzeBanshee]

On the other hand, with something like a 'bust one enemy = all bullets onscreen become medals' system you could milk the enemy waves for loads of points - though that could be pretty tedious in the end. It depends on what scoring system you're going for, really - you're the expert here.

It's certainly a bit crazier that way. I'd have left it the same if it didn't make it so easy to stomp through everything. I think a little extra challenge and tension is a worthy tradeoff. I wouldn't say I'm the expert, though. As many shooters as I've played over the years, I never really thought about scoring systems until I had to build them.
--
I've added another wrinkle to the system to mix things up a bit - when you don't clear a wave (a single enemy or more from the formation escapes off screen), your wave completed count goes down... the penalty scales with the increase in medal value, too. At the lowest value, you lose one, at the highest you lose 5.

edit: Also added a (very) basic rank system for this mode. It kicks in when the highest value medals appear (due to wave clears) and again when you hit the maximum number of clears. Dying is the only thing that knocks it back down again. Nothing fancy; but ok for the time being at least.

Actually that seems a better idea than mine, bullet cancelling after removal of the complete formation. Reminds me of Galaga where the Challenge stages reap you extra score by successfully killing off every last enemy of the formations. Not sure about the penalty for the medal system but that's something one can find out only through tried and true gametesting.

One thing I meant to say to you a few days ago when I discovered the issue myself, but when you added TATE support did you compensate for the modifying of the view_xview/view_yview variables? I used them as a base for spawning enemies at positions outside the game screen in preparation for possible stages that were larger than the screen size (ie. most of the Xeno Fighters stages) but upon rotating the screen all of the enemies were 40 pixels off where they were meant to be. My solution for this was a simple change of variables to accomodate, which will be obvious with GMOSSE MK-V when I finish it up for release next week.

[Rozyrg]

Hmm...I haven't seen anything like that. It would've been especially noticeable for Flying V since I had to significantly increase the view_xview/view_yview values compensating for it's higher resolution. Did your make their creation coordinates relative to something else? That's all I can figure is what is possibly going on. Then again, most of mine use room_width/2 as a base for their starting X location, so maybe that's why mine aren't affected?

Not sure about the penalty for the medal system but that's something one can find out only through tried and true gametesting

Yeah, I'm not sure about it yet either... especially as it would seem at odds with the basic rank system in place. Purposely let an enemy escape every now and then and you don't have to worry about it going up all the way, despite the fact you'd be missing out on some points. It will indeed require further testing. >_>

[BPzeBanshee]

Hmm...I haven't seen anything like that. It would've been especially noticeable for Flying V since I had to significantly increase the view_xview/view_yview values compensating for it's higher resolution. Did your make their creation coordinates relative to something else? That's all I can figure is what is possibly going on. Then again, most of mine use room_width/2 as a base for their starting X location, so maybe that's why mine aren't affected?

Well if you used room_width etc it wouldn't have an affect but if you used view_yview like I did then the issue would be obvious as soon as you rotate the screen and start the stage. For GMOSSE originally I used hardcoded x/y variables which as I mentioned are bad if you have a room larger than the view, or if you're coding spawning of waves via code on such stages instead of placing them in the Room Editor.

Anyhow, sorry for derailing. Here's hoping Flying V turns out well.

[Rozyrg]

Yeah, I did the hardcoded pixel-by-pixel coordinates once (ZpULTRA), and that was all I needed to never do it again. "Hmm... 5 more pixels to the left?" *tests* "Nope. Ten?" *tests* "Nope. Twelve?" *tests until sanity slowly cracks*

I always have any XY coordinates (enemies, text, everything) relative to the room's boundaries now. If I were using views more than just for TATE, I'd probably have them relative to those.

The good news about FV is I'm running out of obvious things to fix/tweak. One thing I want to do next is drop the current, outdated soundtrack, so as soon I have something else to play over the stages, I'll probably put a new version up. The last thing I added in was a proper pause menu.

[raigon50]

SAFSDZGBTZSGVFD!
Ok, sorry. My internet has been dead for about week and a half now, and I haven't been able to get online until a couple of days ago. HURRAY FOR THE INTERNET!

Anyways, I've ran into a bit of a snag involving music, and I saw this coming for some time now but I dumb enough not to take action. I've recently moved to another town. The problem is, I used to visit my dad's house every weekend, where I would work on your music (I have FLStudios 8 XXL at my dad's house, but only FLStudios 8 Fruityloops Edition at my mom's.) The synth I used to make all the SEGA sounds is FruityDX10, a XXL ONLY instrument. Since I've moved, I'm only going to my dad's house every other weekend, so I have little time to actually work.
Don't worry, I'm trying out different FM Synth VSTs (instruments) to see which one works best, but it may be awhile before I do any music for you again. Lucky I got all of the XYX songs done before the move.
I just thought I should give you a heads up.

Also, I'm thinking about releasing the songs for XYX in an album. You fine with that, or do you want me to wait till you're done with XYX?

[Rozyrg]

A week? Damn... that sucks. I thought about asking if any musical experiments relating to Flying V had come to fruition; but as I'm not 100% on what I really even want for it, it's hardly vital. The other week I even pulled some random jazz/gospel cd (some local group I think) trying out different things.

Yeah, that's cool about the album, just as long as everything's finalized and I have the chance to whip up some serviceable cover art first. What I was working on before decided I needed to rescan the lineart at a higher resolution would probably work, come to think of it, though.

[raigon50]

Yes, I got something good.

Alright, so I've been playing around, seriously, for an hour. In that time I made 3 songs.
http://www.youtube.com/watch?v=fgxStCy1hLY
http://www.youtube.com/watch?v=tKG0nic9Juo
http://www.youtube.com/watch?v=JFNNLHvY-p0

It's probably my favorite Soundfont now.

[Rozyrg]

Holy hell, that sounds amazing! The MJ samples, too.... that's just fantastic.

Also, a friend has given me the very interesting proposal of making a video game mock up for a music video. Assuming this goes anywhere, we're thinking it'll look something like Street Fighter. Hopefully I'll be able to show it off, too; but no promises. >_>

[raigon50]

Cool! Sounds interesting.
Right now, I'm exporting a song for Flying V. Sadly, I forgot that using to many Soundfonts in a song causes the usually 2 or 3 minute process to take forever (No joke, I've been waiting 2 hours so far, and it's only at 65%). It sounded really good, so it's worth the wait. I'm looking forward to using that "sexy" Yamaha YM2612 FM bass sound (you know the one) some more. lol

[BPzeBanshee]

Cool! Sounds interesting.
Right now, I'm exporting a song for Flying V. Sadly, I forgot that using to many Soundfonts in a song causes the usually 2 or 3 minute process to take forever (No joke, I've been waiting 2 hours so far, and it's only at 65%). It sounded really good, so it's worth the wait. I'm looking forward to using that "sexy" Yamaha YM2612 FM bass sound (you know the one) some more. lol

'Nuff said. Looking forward to it.

[Rozyrg]

Some further tweaking.... I've made the Dual shot different for each ship now:

Fire Eagle - Wide forwards and back spread, similar to Star Soldier's
Gold Hawk - Narrowing "sword" shot for focusing firepower on a single target
Jade - Seeking shots are clustered together more, also for focusing firepower
Sapphire - Independently targeting (previously the default)


I've got more tweaking to do with these of course, though.

I also removed the cheapo blur effects for all bullets, as simply clipping out their draw stage (whether it consisted of anything or not) gave a slight performance boost. You can still see them on explosions if the option is enabled, though. Since the enemy bullets' 'shadows' were also done this way, I've redrawn their sprites so they now have outlines. Since it doesn't really have much bearing on performance now and I've gone to the trouble of redrawing most of this stuff so it doesn't need blurs to look decent, I might remove the "fx level" option altogether soon.

Got some new music, too (thanks, Raigon ), so I'm hoping to put up a new version soon.

[BPzeBanshee]

Out of curiosity, have you had any issues with the targeting shots spinning around and around like they can't make up their minds? Obviously I'm doing something wrong in GMOSSE for Busterhawk's self-aiming option type do be doing that but was wondering if you had similar issues.

By the way, this might be an interesting read. Their experimental test games exported with it are also quite interesting since despite being in alpha the games pretty much work perfectly with minor issues (Grave Maker is the only one that works with absolutely no issues). The moderators' concerns there seem a bit unfounded if not arrogant though - not EVERYONE makes games in Game Maker where it's all about having awesome-friggin-surface-fancy-glowing-draw-stuff-at-1080p projects and then wondering why it's lagging on their machine. On the bright side though, since it's using the new runner (and it exports to both HTML5 and Windows) hopefully the complaints they've had for ages about the sound engine on Windows will be addressed and using the interal sound handling will be decent instead of complete shit like it is now.

Looking forward to the update for Flying V regardless. HTML5 Gryzor games sounds like a great idea amirite.

[Rozyrg]

The targeting issue could be a few things.

1. if the projectile adjusts it's orientation by checking the point_direction to the target enemy and determining whether it needs to add or subtract it's direction to match that, it could be the good old 0/360 issue. If you're not familiar with it, basically, once the projectile's angle gets close enough to 360 it can never judge the correct way to turn or what will be the shortest adjustment to make. I have a quick fix if that is the case, though.

2. if the enemy's origin point is outside of it's hitbox. I noticed this happening with player ships (I almost never put the box at the origin) and large boss type enemies with a single vulnerable area. The solution in my case was adding a few variables to the ship/enemy creation and modifying the targeting scripts to account for this.

3. target not properly resetting when selected target is destroyed or becomes invalid, or the lack of 'no available target' behavior.

Granted, I've had something similar happen when I though had accounted for all these issues anyways, though.. such as the Cobra's (XYX) missiles stuck in 'searching' mode for a few seconds. Hmmm.

And yes, that HTML5 stuff looks promising indeed! o_o

[BPzeBanshee]

1. if the projectile adjusts it's orientation by checking the point_direction to the target enemy and determining whether it needs to add or subtract it's direction to match that, it could be the good old 0/360 issue. If you're not familiar with it, basically, once the projectile's angle gets close enough to 360 it can never judge the correct way to turn or what will be the shortest adjustment to make. I have a quick fix if that is the case, though.

I thought I already put in measures for that, turns out I solved that problem already with one of the other orbs but didn't make it general to all of the orb types. I say it was quite a bitch to actually track it down to replicate the event too, it was kind of a fluke I even noticed it to begin with. Turns out it's the second obj_bh_type3 orb's alternate program that makes it go schitzo due to the image_angle weirdness on top of what it's already doing. Thanks for the tips though.

Personally, I always use the origin point but I haven't made complex boss designs (simple orbs LOL). I'll probably have a go at doing fancy stuff like that for a future boss.

How do you get the target to reset properly when a selected target becomes invalid anyway? I tried doing this with the new heatseeker missiles and I kept getting errors as soon as something died. S20-TBL's original missiles that got implemented into GMOSSE have this but apparently it caused frameskipping to happen on lesser machines and I had trouble understanding how to modify the code (simpleton brain here).

[Rozyrg]

Go figure... I post that and then I end up with my own targeting headaches later. Not the same issue; but still very nasty and hard to correct. -_-

Anyways, I just use something like..

Code: Select all

if (instance_exists(targ))=0
or(targ.no_hit!=0) //variable for changing collision
or(targ.x<0-25)
or(targ.x>room_width)
or(targ.y<0-25)
or(targ.y>room_height)
or(targ.can_die=0) //variable for triggering explosion/death sequence
or targ.priority<0
{targ=0} 

From there it can choose another target, as I have if targ=0 as a main prerequisite for the selection process.

If the projectile is doing the targeting independently,though, I find it looks a bit better if you cancel any further targeting instructions and just have it go along its merry way in the same direction. This also limits the 'straggling' behavior somewhat, too.
-------------------------
Some random things:


Just to give you an idea of the kind of scores that are possible with the new mode..


Inspired by the recent bullet pattern creation thread, these are from a little test game I made to see if I could generate bullets without using multiple objects. This uses one object per burst/patten, which keeps everything in order using 1D arrays. I've heard GM's data structures were possibly faster; but I haven't figured them out well enough to test that.

With this method, I can squeeze about 150~200 more bullets on screen using my "good" PC, while on my laptop, the sweet spot is around 100~150 or less regardless. Not that I really need any more than that, really.

[Rozyrg]

New video of Flying V (courtesy of SSK.. thanks, man!)
And another!

- You now receive a 15X bonus for destroying an enemy before it can fire (you can see this in the video as "NO SHOT!") I might change this to increase up to that amount after doing this consecutively, though.

- "Fx Level" has been replaced with a more generic "Performance" setting. All this does is limit how many bullets can be on screen and, at the minimum setting, reduce the number of sub-sections of segmented enemies such as the trains.

- I'm still experimenting with alternate bullet generating methods. I managed to get the one I was working on implemented; but the results would go from a little better to clearly worse than what I had before when the array got big enough. Basically, as long it got a chance to reset (when all bullets had been removed), it would be ok; but bosses that are constantly spewing bullets are another matter entirely. Back to the drawing board, I guess.

[BPzeBanshee]

- You now receive a 15X bonus for destroying an enemy before it can fire (you can see this in the video as "NO SHOT!") I might change this to increase up to that amount after doing this consecutively, though.

"QUICK SHOT!" sounds more logical to me but then again the Raiden Fighters games was my gateway to shmups. Making the shot increase up to the 15x bonus also fits in more with how Raiden Fighters JET does it - a good method indeed.

- "Fx Level" has been replaced with a more generic "Performance" setting. All this does is limit how many bullets can be on screen and, at the minimum setting, reduce the number of sub-sections of segmented enemies such as the trains.

Does this make the game easier though? :S

- I'm still experimenting with alternate bullet generating methods. I managed to get the one I was working on implemented; but the results would go from a little better to clearly worse than what I had before when the array got big enough. Basically, as long it got a chance to reset (when all bullets had been removed), it would be ok; but bosses that are constantly spewing bullets are another matter entirely. Back to the drawing board, I guess.

I never understood most of that bullet generating method thread, probably cause I have no programming experience with C++ besides a "Hello World!" and massive installation problems with MinGW.

1-dimensional arrays, you say? So what, like one object draws a whole shitload of bullets using these arrays to dictate each sprite's position and hitbox each instead of just having one instance for each bullet...

[Rozyrg]

Does this make the game easier though? :S

As of now, it does a bit; but less bullets and enemies =less goodies to collect, so that's less potential points. It's partly why I'm trying to squeeze a few more bullets on screen, though, so I don't have to do things like this just to keep it from crawling. Granted, there aren't that many at any given time - the level 3 boss fight might peak at around 150.

I never understood most of that bullet generating method thread

Honestly, me either; but the idea of doing it a different way intrigued me enough. Even though I'm sure they're not really the same thing, the talk about Data-oriented methods got me curious enough to get my feet wet with GM's data structures. I have yet to integrate any of that successfully yet; but I might use something like that for XYX's map builder, though, since the speed advantage is supposedly more from accessing data rather than if you're constantly changing it.

1-dimensional arrays, you say? So what, like one object draws a whole shitload of bullets using these arrays to dictate each sprite's position and hitbox each instead of just having one instance for each bullet

Yep. Basically, each bullet is just a handful of variables, like:

Code: Select all

bullet_x[var]
bullet_y[var]
bullet_type[var]
bullet_speed[var]
bullet_direction[var]
bullet_img[var]

The bullet controller object uses loops to cycle through them all, change their positions, check for collisions and, if they aren't marked as 'dead', draws them. Considering potentially multiple hundred loops happening per step, it's pretty easy to see how this could get ridiculous; but it seems to have at least minor advantages if you can figure out how to manage it all.

..that's a pretty big if, though. I thought I finally had it behaving tonight; but it just went haywire again. >_>