The system relies on external files. All data is external
even enemy programming 'n stuff
well that's great!
but it's just that iterating through objects using fastloops is too slow.
That looks brilliant, and your style is as good as ever.
and i'm... jealous
If you e.g. spread a value over 10 objects and iterate through them, MMF2 must actually make 100 comparasions to filter out the objects.
Maybe when the New Faster fastloops objct is done that will be solved :)
yes, that's right, sadly
MMF2 should have a built in object looping like OINC's Peer Looping
so that it only makes 1 check per loop
I'm having a lot of trouble with that. This game is heavily optimized, I use less code for all this than a lot of simple beginner MMF2 projects, but I can't optimize it much further, and the lack of "for each" and sub-events is still eating too much CPU.
I did investigate some time ago about making a loop-object extension
I couldn't work out a way, and there's no guarantee it'd be any quicker due to function call overhead
**** Eliyahu Leaves Chat ****
(but above certain number of objects, yes it would definitely be quicker)
Mind you, I only investigated in MMF 1.5, perhaps it's more feasible in MMF 2
The main problem are not actually the loops, it's that each time you compare an object's ID (on every iteration in every loop), every ID of every object is compared.
cause then I'd only need to use the "on loop" command once.
* Nifflas is having great hopes for MMF3 in this area
* ChrisB too
vertigoelectric I like the logo for fangame!
Thanks. I can't seem to get it to appear for me.
it shows fine for me on the project page
**** Eliyahu Enters Chat ****
why does nifflas has a grey dot?
While you were gone Niff got a ball
that means VIP
I'm going to make a game over the summer, but I don't think I can do it in MMF2 for performance reasons (and because I eventually want to have multiplayer, and I want to use technologies that I can guarantee are mature, no offence to Ja mie)
Hopefully I can feed my experience from that into MMF3 somehow
The summer which is over at the end of this month ;)
Sounds like a good idea.
I personally hope that MMF3 will be a little bit of a fresh start. If you think about it, a lot of actions and conditions in MMF2 points right back to Klik & Play. They're all not thought out in the most optimal way. Mixed 0-based and 1-based, and a sound system that results in digital pops if you don't do your own voice allocation.
It's controversial to say, but I'd want CT to drop the backwards compatibility.
I said much the same before MMF2 :) I think it needs to be done sometime
For once, think through all the features, and see if they can be arranged in a more optimal way. It's better for both beginners and advanced users.
I would have to agree, the 0-based mixed with 1-based indexes get annoying and extremely frustrating
I think it'd also be much faster to develop starting from scratch. Y+F have had to contort the runtime on a few occasions to get things to work
Just looking at the difficulties porting to Java and HWA proves it, it is taking a long time
Getting things largely right the first time has its disadvantages :)
I know it's a lot for two programmers to do, but even so, it drags on a looooong time.
I think that the extension difficulty pretty much killed the backward compatibility of MMF2 anyway... although the release of the first bonus pack did go some way to reversing that.
sure but I don't think Y+F had planned in 1993 many of the things that ended up in MMF2 even
a lot is user input
It seems that a difficulty is that it has to focus on two different audiences, now... the beginner game-makers who need things like the insta-save/load features and more automation, and the high-end users who need more optimization and things that are more similar to programming concepts.
and I havent used any of MMF2's sprites in at least a year
It's tricky, but in some cases I feel that trying to make things too simple can actually get in the way of making things simple.
The two groups do meet in several places though (object loop is the most basic example of something that would be useful for beginners for simplicity and experts for performance)
If you special case too many things you are sure to have an impedance mismatch when mixing them together
and since programming is mixing lots of different things together, you are worse off :)
it'd be nice if TGF3 turned out a lot different than MMF3
They do... it's more things like... on the forums, the first thing that a lot of people are advised to do is not to use bits of MMF2! (Such as the infamous standard platform movement, the mouse movement...)
more beginner oriented
I believe CT should just ship a bar of chocolate with MMF 3 and then everyone would be happy
but it might melt
**** RickyRombo Enters Chat ****
A couple of valiums...
I don't actually think the platform movement is too bad, it just needs to be used in a non-obvious way (invisible detector object)
I'd even go as far as saying ditch the pre-built movements and come up with a way to get movements as event sheets (that uses it's built in variables that's not seen outside the movement sheet). This way, beginners would still have their pre-defined movements, but the community could develop them in MMF itself rather than the CT staff having to hard-program them.
hmm? Seems I've joined in the middle of a conversation...
then you would load these event sheets?
That implies that... well, there's a thought. Custom actions for objects...! Taking the place of the use of fastloops for function calls...
Yeah. You'd probably be able to load event sheets from external files, or from a "movement" directory where you can add many ones.
If they would carry their own variables, you'd not have to allocate alterable values for it.
* RickyRombo thinks he is slowly becoming a spammer
I would go as far as allow extensions to be built that way, if possible (basic extensions nonetheless)
Event "includes" would be wonderful, I think that's probably MMF's single largest weak spot at this point...
* RickyRombo :o
I personally don't see a conflict between advanced and easy to use. Everything just needs a slight re-think.
The event editor can be very powerful, it's just held over from a much simpler time and hasn't expanded enough
I still think that a text-mode event editor would be great. But then it gets clouded up with scripting and the whole thing collapses...
and save to xml :D
I would veto the text mode event editor if possible
save to xml, definitely :)
I don't like text editor idea
**** RickyRombo Leaves Chat ****
I will not rest until MMF is compatible with version control
he's coming back
* Eliyahu reinstalls RickyRombo
that doesn't mean I can't say bye!
* Andrew hopes ChrisB can go without sleep for longer than most
Well, the save-to-XML gives many of the same advantages I'm thinking of - just a way to be able to move things in and out and edit them more... manually...
6 years and counting :(
**** RickyRombo Enters Chat ****
ok I can't do it
the installation came back to success!
I can't leave
I'm using SVN just now for MMF2, just for a project backup (and it sort of makes sense for my external files).
How separate would you want to make it - a file for every frame, event sheet, object?
well, depending on how event sheets are implemented (separate please!), they could all be quite different...
Andrew - you need a /join command so when you quit you can join again
It'd be nice to have all the graphics stored externally, even.
or you could just click a button
I actually think Construct has some neat ideas about that. I don't like Construct and will stick to MMF, but I think they have some interesting solutions that could be taken (since they obviously got plenty from MMF)
RR if you quit you wouldnt be able to type commands =/
why isn't my fangame logo showing up on the projects list page
but could "offline"
yeah, certainly, though they suck hard in other areas :)
my fangame logo shows on the project page but not the project LIST page
I haven't looked at them in a while, Nifflas, but I was secretly planning to try to remake a recent project in it and see what was different... forgetting the Construct embargo for the moment, can you say what you didn't like about it compared to MMF2?
Oh, well I wouldn't know because I'm BANNED :(
for no reason too
LB there was a reason
Andrew is there an 0.9c?? Eliyahu is lying to me I think
oh, now it's working
It doesn't seem stable enough, but my main problem is that they decided to make it 1-based. That's just wrong.