next »

[Hayden Scott-Baron]

My girlfriend is a 2D artist, and is interested in using MMF2.  I was wondering whether anyone experienced with MMF2 could help out and let us know the pros and cons to the software.  (I have mild interest in MMF2 also, but I'm focusing on Unity right now, and PC-only development is less appealing right now)

I notice that most games written in Multimedia fusion seem to be very low resolution. I've been assured that this isn't a limitation of the software, but... then why is it that almost every game has been very low res? Is there some sort of problem with high resolution graphics of 640x480 or 800x600?

How much code is typically involved? Is it all GUI driven or do you tend to write a lot of script? Is there a lot of existing code that can be copied and pasted?

All of the previous software that I've used (Blitz3D, Unity, etc) has its fair share of problems, so I was curious about what difficulties we might face if we use MMF2.

[moi]

There aren't many limitation, except that some type of games would be very difficult to make (I'm thinking something non convrntional).
And there is a 30 days fully functional demo so try it!

[policedanceclub]

Although at higher resolution you will run into performance issues quite easily.

[Hayden Scott-Baron]

We'll definitely be giving the 30 day trial a shot. She's after making a platform adventure game, so it should hopefully be okay for MMF2.

Although at higher resolution you will run into performance issues quite easily.

What constitutes 'higher resolution' ?

[Terry]

Any resolution, based on my experiences with MMF - aside from Nifflas' stuff and Game With A Kitty, I haven't seen any MMF games that don't have performance/screen tearing issues. Something to bear in mind, at the very least.

As an alternative, Game Maker doesn't appear to have the same issues at all, and is significantly cheaper too.

[soundofjw]

The nice thing I like about MMF is that Clickteam is putting a lot of money into contractors to build up the extension base - to do lots of cool things. Granted, some of these seem trivial, but others of them are simply awesome.

The performance issues will hopefully be resolved sooner than later as HWA (Hardware Acceleration - basically using D3D to render instead of 2D API stuff) development is going rather smoothly (it seems to have slowed down as there haven't been any beta updates in a while and the main developer was on vacation, but he is also very much a stickler at only releasing plausible things, so I see this as OK.).

I think investing in MMF2 will only be a good decision as it's a constantly growing product. Bonesaw and The Spirit Engine 2 are two games that I know of that are almost 100% opposite gaming styles to show you what's possible. Skydiving Academy was made with the new Hardware Accel. stuff (includes pixel shaders). Knytt, Knytt Stories, WaDF, and anything else with Nifflas' name on it are also MMF games.

Seems like a good idea.

[Hayden Scott-Baron]

I didn't realise that "Mr First Skydiving Academy" was done using MMF2, or even a beta version. This certainly gives me a lot more faith in using MMF2 to produce stuff, as that game runs at 800x600 and has no problems with full screen mode either.

Is the beta available to everyone? It's encouraging to know that anything that was developed could theoretically later be upgraded to have DX8 acceleration.

[Tobasco Panda]

Also, I know Konjak's latest games (Noitu Love 2 and Chalk)are made with MMF2. So with a bit of work and love you can pull off some pretty impressive stuff!

[Hayden Scott-Baron]

Thanks for the heads up, it does sound like MMF2 should be fine for her. She's planning to get the "Power User's Guide to Windows Development" book, by John Darby, so hopefully it will put her in a decent position to work with it.

[Tatsu]

I'm a big fan of MMF2; it's what I do all my prototyping in these days. I've gotten comments that the scripting portion looks more like a spreadsheet than it does traditional programming. I think the scripting is incredibly intuitive, but due to the format there are a few things that are more difficult than they should be, like iterating through multiple instances of the same object, and nested loops can quickly become difficult to read.

Overall I highly recommend it, though; once you get it down it's great for rapid development for any kind of 2D project.

I'd like to know how your experiments with Unity goes. Someday I'll probably run across a project that I want to be 3D, and MMF still may not support it well enough by then.

[medieval]

Yeah. The instances thing quite put me off.

But it's quite a neat and practical tool for game development

[___]

MMF2 would be *perfect* if it just didn't lag so damn much for me.

[Farmergnome]

skydiving academy was made purely for hwa - 800-600 with decent performance, where to start, first off mmf2 is awesome, every negative point im about to list, remember foremost, its awesome.... AWWWEESOMMME.


The editor is either love or hate, if you dont like writing your code, and can manage the quirks involved with kliking, then cool, its bearable, be prepared for the occasional crash and you pulling your hair out wondering wtf is wrong when your file sizes get fairly large too.

HWA is still beta last time i checked, so alot of the cool shit doesn't work yet (fullscreen?), but its worth learning as I have faith clickteam will kick some serious ass on it.

The way game frames are setup in that you have to recode your engine for each frame makes developing certain genres very hard, anything that would span across several scenes that you would need to recode core functionality, or write your own level editor, loading system, depending on what you want to do.

Theres more minor shit, but ill skip that, PM me if you want the full list haha.

Okay good points!

Its probably the best prototyping tool I know off, especially for someone whos a artist primarily, the whole program is realy setup to knock out some sweet games realy quick in, keeping everything fairly artist friendly.

Java support comming is promising...

Theres lots of examples provided around, and extentions keep pushing mmf2 forward with extended functionailty, realy the most overlooked thing I think.  Theres extentions out there for alot of shit, takes the effort out of doing alot of the ground work.

Another thing that no ones really mentioned yet is mmf2 compiles clean, all your icons etc are handled easily, all your graphics that are actives are compiled into a clean exe, music also, theres usually no wierd runtime quirks, its pretty much to guarenteed to run on most windows comps, saves you time fixing wierdass runtime errors, and keeps the whole application nice and compact.

pixel shaders 

[Hayden Scott-Baron]

skydiving academy was made purely for hwa - 800-600 with decent performance
[...]  HWA is still beta last time i checked, so alot of the cool shit doesn't work yet (fullscreen?), but its worth learning as I have faith clickteam will kick some serious ass on it.

What is this HWA thing you speak of?
Skydiving Academy worked full-screen on my windows laptop, although might it have been using a maximised window 'over' the rest of windows?  Either way, I had a much better fullscreen experience than I do with most MMF games, where fullscreen doesn't work at all.

[Gainsworthy]

Wonder where Pencerkoff is.

[Farmergnome]

skydiving academy was made purely for hwa - 800-600 with decent performance
[...]  HWA is still beta last time i checked, so alot of the cool shit doesn't work yet (fullscreen?), but its worth learning as I have faith clickteam will kick some serious ass on it.

What is this HWA thing you speak of?
Skydiving Academy worked full-screen on my windows laptop, although might it have been using a maximised window 'over' the rest of windows?  Either way, I had a much better fullscreen experience than I do with most MMF games, where fullscreen doesn't work at all.

hardware accelerated, its in beta atm, same with java, check the clickteam forums to download the latest versions of both.

[Pencerkoff]

Wonder where Pencerkoff is.

Hello this is Pencerkoff

I would love to add many comments on this topic but I have been busy with Georgia.  I am speaking at rally tomorrow, wish to me luck.

As for MMF2, my largest complaint is that there is no good way to debug your program .  There are so many little quirks in MMF2 that you go crazy trying to backtrack your code to where problems are.  If your program or game is simple, sure MMF2 is fine.  For any not simple approach, it all go to hell.  I heard MMF2 was programmed in Georgia.

-PENCERKOFF

[deadeye]

Pros: 
* It's really easy to get started (noob friendly).

* There's a huge user community with lots of tutorials and examples.

* The spreadsheet-like interface is easy to understand.

* It outputs fairly optimized .exe's that run well on most computers (as opposed to GM games).

* Compared to GM, MMF is much easier to use, and has more functionality without having to get into scripting.  In GM, to get more out of it you have to learn GML (or whatever it's called).

Cons:
* The huge user community has a large percentage of dumbshit kids that offer no help and no constructive criticism.  Try not to post on the TDC boards if at all possible, and just look up solutions to your problems in already existing threads or articles.

* The spreadsheet-like interface gets cluttered and unwieldy with larger projects.  Hunting for pieces of code to change becomes a hassle and a half.

* You basically have to code your game on one frame, and load levels dynamically from external files (which means you have to code your own level loader and level editor) or copy your entire game engine onto each level of your game.  Once you do make this copy, if you want to change code in your basic game engine, it means you have to go into each level of your game and change the code there as well.

* Compared to other similar game-making programs, it's expensive.


The biggest problem I had personally with MMF was understanding "spread values."  It's a method of picking instances that is completely ass-backwards and unintuitive.  You would think that this (in pseudo-code) would work:

+ enemy.sprite "falling" = 1
+ enemy.sprite is overlapping ground
 - set enemy.sprite "falling" to 0

Sounds reasonable enough.  If your enemy is falling, and it hits the ground, stop falling.  But doing it this way doesn't just pick the instance that is hitting the ground... it simply changes every enemy sprite's falling flag to 0, meaning they all stop falling at once, even in midair.

To get MMF to pick only the sprite that is actually hitting the ground, you have to do this convoluted method of assigning each instance a unique number, then run a loop checking against the number, then perform the action on only the instance with the number that matches the loop index.  Once you get it down it's easy, but it's still a pain in the ass to do.  God knows there has to be a better way of picking instances.

And it gets even weirder, because this method of picking instances isn't even consistent within the MMF framework itself.  For example, if I have a bunch of enemy sprites, and I have a routine like this:

+ enemy.sprite overlaps bullet.sprite
 - destroy enemy.sprite

Then it only destroys the one that is overlapping.  WTF.  Then why do I have to use spread value some of the time, and not others?  Makes no damn sense to me.

[___]

Cons:
* The huge user community has a large percentage of dumbshit kids that offer no help and no constructive criticism.  Try not to post on the TDC boards if at all possible, and just look up solutions to your problems in already existing threads or articles.

Haha, ain't that the truth.  Sometimes its the community that disturbs me the most about Clickteam products.  Granted, there are some nice and helpful guys somewhere in the mix, but the rest is just awful.  Its like a bunch of 6th graders in a food fight at lunch, who happen to make games... sometimes.

* You basically have to code your game on one frame, and load levels dynamically from external files (which means you have to code your own level loader and level editor) or copy your entire game engine onto each level of your game.  Once you do make this copy, if you want to change code in your basic game engine, it means you have to go into each level of your game and change the code there as well.

Oh yeah, while making Bonesaw, my copy and paste skills leveled up considerably.  I can ctrl c + ctrl v lightning quick.

The biggest problem I had personally with MMF was understanding "spread values."  It's a method of picking instances that is completely ass-backwards and unintuitive.  You would think that this (in pseudo-code) would work:

+ enemy.sprite "falling" = 1
+ enemy.sprite is overlapping ground
 - set enemy.sprite "falling" to 0

Sounds reasonable enough.  If your enemy is falling, and it hits the ground, stop falling.  But doing it this way doesn't just pick the instance that is hitting the ground... it simply changes every enemy sprite's falling flag to 0, meaning they all stop falling at once, even in midair.

To get MMF to pick only the sprite that is actually hitting the ground, you have to do this convoluted method of assigning each instance a unique number, then run a loop checking against the number, then perform the action on only the instance with the number that matches the loop index.  Once you get it down it's easy, but it's still a pain in the ass to do.  God knows there has to be a better way of picking instances.

And it gets even weirder, because this method of picking instances isn't even consistent within the MMF framework itself.  For example, if I have a bunch of enemy sprites, and I have a routine like this:

+ enemy.sprite overlaps bullet.sprite
 - destroy enemy.sprite

Then it only destroys the one that is overlapping.  WTF.  Then why do I have to use spread value some of the time, and not others?  Makes no damn sense to me.

Now, that sounds like a weird problem, but in your first example it may be your order of conditions.  If you put the overlap test first, then it should choose the right object.  Yeah, its crazy shit like this though that drives you mad.  Swapping the order of your condition checking can make or break your engine, and fast looping everything after awhile starts to bring everything to a nice solid 10 frames per second.

[deadeye]

* You basically have to code your game on one frame, and load levels dynamically from external files (which means you have to code your own level loader and level editor) or copy your entire game engine onto each level of your game.  Once you do make this copy, if you want to change code in your basic game engine, it means you have to go into each level of your game and change the code there as well.

Oh yeah, while making Bonesaw, my copy and paste skills leveled up considerably.  I can ctrl c + ctrl v lightning quick.

picking instances

Now, that sounds like a weird problem, but in your first example it may be your order of conditions.  If you put the overlap test first, then it should choose the right object.  Yeah, its crazy shit like this though that drives you mad.  Swapping the order of your condition checking can make or break your engine, and fast looping everything after awhile starts to bring everything to a nice solid 10 frames per second.

Since you're into MMF I'm sure you've heard of it (through drama and such) but there's a new game making tool in the works called Construct that I've been keeping my eye on.  Actually, more than keeping my eye on, I've been actively helping out with bugfixes and such for a while (and the Little Hitler in Toyland demo I made for the VGNG compo was made in Construct). 

It's still in beta, but it looks very promising.  For instance, you can code your engine (or even just parts of it) on one Event Sheet, then just re-use that sheet for any frame you want.  So no copypasta, no editing multiple frames, you just edit the original sheet and viola, it's done.  You can create as many such sheets as you like and include them on any frame you like, so you can have one sheet for controls, one for animations, one for sound, etc.  It makes organizing your program way easier.  The event interface is also way better, IMO... instead of a spreadsheet, you just hit "new action" and pick the objects you need from a popup.  No hunting for code or hovering over checkmarks to see what action is underneath:



So it's laid out more like a regular program script.  Also, it has this sweet sub-event system:



So you don't have to create a lot of redundant events that have similar conditions.  You just set up your main conditions, then as a sub-event set up your differing conditions.  And all the object instance picking is handled automatically, so there's no order-of-events conflubberation like in MMF.

It's got a lot of other cool features too, but it's currently in beta so it is kinda buggy still.  But definitely one to watch, I think.