Construct 2 - Realistic State after 1 gazilion downloads

Discussion and feedback on Construct 2

Post » Thu Mar 06, 2014 6:33 am

tomsstudio wrote:Currently been working on 2 native compilers for android and ios for the past month.


DUDE

Interest piqued

Like seriously

Please don't give up on this. PLEASE.

Do you have any plans on how to deal with the memory management problem with large games loading everything at the start? Or would that require collaboration with scirra to add more manual memory management options? Or perhaps a plugin?
Moderator
B
95
S
34
G
33
Posts: 3,007
Reputation: 27,876

Post » Thu Mar 06, 2014 6:49 am

Right now i'm trying to mimic c2 behaviors so they work as closely as possible when previewing in HTML5.

But since i'm working on compiling to MOBILE, performance is a must.

Saying that my priority after behaviors is audio, since that's the most frustrating feature i read on the forums.

I'm not going to give up on this, having a a'lot of fun coding and working out problems :)

And this compiler is purely for C2, so i don't have to worry about supporting other developers like other companies...

:)
B
10
S
2
G
1
Posts: 92
Reputation: 926

Post » Thu Mar 06, 2014 6:52 am

I want to share my experience with you guys. I'm not only a construct 2 "developper", but also very experienced in other "real" programming. I totally agree with you, that most people think, construct is just a comfortable tool to create games easily. This is totally wrong. Construct requires a lot of knowledge of programming. Not because it's so hard to get something working. But it's so hard to make it good.
I love reading the forums and most of the time I just inspect the capx file of other construct users. And I can tell you with a 99% chance if the person has programming knowledge. Most of the time it really hurts in my programmer's heart to see what people do.
But to be honest it was very difficult for me in the beginning, too. Even though I had a lot of experience in programming i struggled in the first month with the unfamiliar event driven system. And I still learn something new about construct every single day.

But back to topic: You said the performance of most games is horrible. Yes maybe it's true but this is not constructs fault. In the beginning I also had a lot of performance issues. But then I tried to find a way to fix them. And I never had these issues again and all my games are running on 50 fps.

And maybe this opinion is unpopular but flash is dead and html5 the future.There is no flashgame I couldn't build same quality with HTML5 (actually this is what I'm currently doing and I never had problems. Quite the contrary: People are impressed how good it works and how fast projects can be programmed with HTML5). When I look at friends/collegues using YoYo'S gamemaker I only hear complaints. Tests even proof that c2 is more performant than GM.
B
14
S
3
G
1
Posts: 255
Reputation: 2,027

Post » Thu Mar 06, 2014 7:04 am

Other issue is that events are going in one big loop,
so 100 events may be good for simple game,
but 10.000 events (I guess) could kill even desktop

also bullet behavior that isn't 100% smooth both on mobile/desktop

also saying "use Crosswalk" is not serious at this moment
and +18MB apk size can dramatically decrease number of downloads on Google Play

but yeah... anyway: there is no competition with anything similar to event sheets
B
18
S
7
G
1
Posts: 783
Reputation: 4,247

Post » Thu Mar 06, 2014 7:36 am

@szymek - If you code it right, C2 has no problem with thousands of events because they don't all have to run every tick. Using groups and subevents, it only runs the ones needed. For example:

Sprite.count > 0
- do stuff with sprite

If sprite.count is 0, the entire subevent tree is immediately skipped. Because of this, you can have thousands of events even on mobile as long as not too many of them are actually running each tick.

I have a hard time imagining a game that requires 10,000 events to run every tick.
Moderator
B
95
S
34
G
33
Posts: 3,007
Reputation: 27,876

Post » Thu Mar 06, 2014 7:45 am

szymek wrote:Other issue is that events are going in one big loop,
so 100 events may be good for simple game,
but 10.000 events (I guess) could kill even desktop

also bullet behavior that isn't 100% smooth both on mobile/desktop

also saying "use Crosswalk" is not serious at this moment
and +18MB apk size can dramatically decrease number of downloads on Google Play

but yeah... anyway: there is no competition with anything similar to event sheets


The 10000 events per loop is not really a C2 issue. If you code in any language you pretty much always have to make a game loop, and if you code the game such that each and every single iteration of your game loop calculates every single thing possible every time, then you will run into this same problem. You should only include necessary events on every frame, then stagger the rest of the events with a random based on how often they need to run. For example if they need to run a minimum of once per second, you can put a condition of Every X milliseconds, where X is 700+random(200). The things that you really need every loop iteration are control checks, possibly movement if you want it to be as smooth as possible, etc. But most things can be staggered.
B
11
S
2
G
3
Posts: 283
Reputation: 1,968

Post » Thu Mar 06, 2014 8:01 am

Juryiel wrote:The 10000 events per loop is not really a C2 issue. If you code in any language you pretty much always have to make a game loop, and if you code the game such that each and every single iteration of your game loop calculates every single thing possible every time, then you will run into this same problem. You should only include necessary events on every frame, then stagger the rest of the events with a random based on how often they need to run. For example if they need to run a minimum of once per second, you can put a condition of Every X milliseconds, where X is 700+random(200). The things that you really need every loop iteration are control checks, possibly movement if you want it to be as smooth as possible, etc. But most things can be staggered.


Couldn't agree more :)
You think you can do these things, but you can't, Nemo!
Just keep reading.
Just keep learning.
B
65
S
16
G
9
Posts: 1,429
Reputation: 12,718

Post » Thu Mar 06, 2014 8:20 am

@Arima @Juryiel

ok, agree :)

anyway when I tried Airscape demo on my 1,5 years old laptop, with 2x core AMD and 8 GB - the game wasn't 100% smooth (and it was not connected with low/high details)
B
18
S
7
G
1
Posts: 783
Reputation: 4,247

Post » Thu Mar 06, 2014 8:23 am

Interesting discussion that has been going on here.

I think it becomes a bit muddled when we talk about exporters. Construct 2 has just one exporter, which is HTML5. Then it has different ways to export that. With third parties providing ways to wrap it to appear native-like (mostly slapping a standalone browser engine on it as far as I understand).

Now new native exporters would require reworking all plugins, behaviors and... stuff. :mrgreen:
If I understand correctly that is kind of what @tomsstudio is doing at the moment, which seems to have a lot of potential.

At the very least Scirra should provide reliable wrapping solutions. The closest thing would be Node Webkit, but yes, that's obviously not for mobiles.

If my memory doesn't betray me, there was talk about Scirra hiring additional manpower in 2014. This could enable them to work on extending the exporting options or even actual native exporters, while the regular feature progression of C2 didn't need to suffer.

In any case I do believe we have a piece of development software with a great concept, potentially enabling anyone to create the game he wants to create. I say potentially because C2 isn't a toy, it's still serious software which you will have to learn. It's just much less of a hassle to do so.

So yes, I generally agree. It is also my opinion that the current focus of Scirra should be on providing reliable exporting solutions for mobile platforms.
B
23
S
6
G
10
Posts: 1,041
Reputation: 7,553

Post » Thu Mar 06, 2014 8:27 am

lwgames wrote:I started making small games with construct 2 one year ago, but only recently i got serious into it, and while i got into some bumps on the performance (web) the community helped with tips and i struggled on.

So far i released 1 game on google play, amazon and on web and i was thinking at how much issues i encountered to publish a semi-polished game.

1) I love how its promoted as an easy solution, but in reality if u don't learn all optimization and workaround tips, you wont be able to release any game, anywhere in a semi polish state. Performance, crashes and audio problems will haunt you to death.

2) The lack of native exporters makes this even worse. Cocoon js is the only alternative at the moment, even tho Ashley sends us to XDK which is months away from a stable compiler, but we get ignored and sent to knock on doors. Its ludei problem, its intel problem, its Scirra problem, go bother them.

Here is even more sad, that after Construct 2 sold very well, Scirra simply didn't bother to make a few c2 exporters, they just send us to third party sites, which most of the time don't care much about C2. Man power you say? Well look at yoyo games, they made their exporters and each costs money and let me tell you, their HTML 5 mobile games have more options and flexibility, and overall better performance.

3) I really hate this HTML 5 trip, i only use this engine because of the even sheet workflow. Why i hate it? Because even after 2 years listening to Ashley about promoting the so called HTML 5 rush, the reality is completely different. Imagine if instead of riding this hipster trip, they would've built on Construct Classic, make it as simple as C2 and but with tons of native exporters. That wouldve simply destroy GM and other competition.


What you can realistically make at the moment:

a) Small html 5 web games, they work ok, but they need a lot optimization on large games. Construct 2 its easy but in reality if you want to make a deeper game, well you need a lot of experience with it. "lets everyone create amazing games." right, lets everyone create simple games with mediocre performance at best. In the end its possible to publish a decent game, but dont expect to make any money out of it, unless u put some ads on it, and that's about it, making a few dollars at best per month. Licenses are very rare for html 5 web games, but not uncommon. There are few networks you can share to, but its very hard, HTML 5 is simply inferior to flash, but a interesting game with attractive graphics can score a deal.

b) use cocoon js and publish small games on android and ios. The performance its ok with optimization, but lots, lots of problems there. The loading times are huge, audio music crashes your phone on sleep, but at least the performance is good. Post on scirra forums for support and watch Ashely sending you to Intel to use a incomplete, alpha sdk.

c) make non games, 3-7 days cash grabs and publish your income reports for the world to see. These html 5 web games that publishers buy(tho now the bubble its over, so its harder to sell them than last year), require VERY SIMPLE gameplay, and a lot of optimization to even work on mobile. Not Scirra fault here, is just this stupid technology. When i hear that Flash will die because of HTML 5 i laugh my ass off. Did u even saw the quality of the flash games? Maybe in 3-5 years we will get to have possibility to create something similar to flash at that performance, but i doubt it. HTML 5 mobile games are worse than those "free to play" mobile cash grabs.


So you can create something, but you are sent to knock on doors, and hope that you experience will help on the way. I am afraid to even make a deeper game with nodeweb kit and see that the performance is horrible on laptops.

I am very disappointed after 2 years, the engine improved in performance, but the ability to actually release a polished game, well its there, somewhere, but to be honest ill thinking on moving away from this engine. Ill create 2-4 more small games, see whats going on the exporter subject, if still goes on this ignorance trip, ill just move to Unity and call it a day.

So what you guys think?


/frustration off


Unfortunately, I feel the same way. The only reason I stuck with C2 over YoYo or Unity was because it was easier to start up with. Now, almost two years later, my hobby now has real business concerns which may lead me to jump ship. I have already bought a GM license and I really like what I see, but, for now I will finish up the game I am working on with C2 and reconsider my options after it has been published.

Which reminds me. I have a HeroEngine license as well.
B
56
S
15
G
13
Posts: 826
Reputation: 17,665

PreviousNext

Return to Construct 2 General

Who is online

Users browsing this forum: No registered users and 13 guests