next »

[eigenbom]

SFML is good only for small projects.

Care to elaborate on that?

Yeah, it has hundreds of bugs that you wouldn't expect and even if you ask the creator of SFML (Laurence?) he won't give you any helpful answer, probably he doesn't know why these things aren't working too.
Also there is the second part of bugs that are going to be fixed/changed in SFML 2.0, but how long should we wait for this version?

So: it's bugged as hell and no one can help you with those bugs.

You mean small projects that win lots of awards? SFML's easy to use, but has quirks like every other library. It doesn't have a great community though, as compared to e.g., Ogre3d... And if you're a beginner then I agree, Laurent probably won't be too helpful.

I didn't say that it's impossible but it's a bad tool to make big projects. So, if you're not a beginner then Laurent will help you? He doesn't help beginners or maybe we're too stupid and don't understand the hidden meaning of his "Did you load that image?" questions?

I agree that it's easy to use, but that's all about SFML.

Well in my experience he's been quite helpful. And blunt. But give the guy some credit - he's probably been answering noob questions for 5 years. The fact that he's still active on the forum as well as the sole developer is just awesome.

He won't give you any helpful answer, probably he doesn't know why these things aren't working too.

Why would he? I'll write a couple of thousand lines of code and then ask you to debug it... To get the most help from him you need to supply a minimal working example that reproduces the bug, then if it indeed is a bug in the library and not in your code, then it'll get added to the issue list. Open-source software improves this way. We all benefit from doing a little work and helping each other out.

I've been using it for 6 months now on moonman, but haven't encountered any show stoppers ... so I'm curious to know what makes it a bad tool for big projects? Was it just the bugs? Or was it something else?

[PompiPompi]

Well, I am not sure. There might be crashing bugs in SFML 1.6.
Also one bug I encoutered was that scaling doesn't work properly in 1.6(it's a known bug, but I didn't know it).
I had to edit his source code to fix that.
Though hopefully 2.0 fixes many of these bugs.

[ANtY]

@eigenbom: 6 months and you call me a noob/beginner? Whatever...
Just don't assume that I was creating threads with my whole source code and description like "ain't working, pls fix it" 

Yeah, he's answering noob questions for 5 years, it's good, but it ain't gonna help me with my bugs that I found and didn't receive any help with them.

[PompiPompi]

@eigenbom: 6 months and you call me a noob/beginner? Whatever...
Just don't assume that I was creating threads with my whole source code and description like "ain't working, pls fix it" 

Yeah, he's answering noob questions for 5 years, it's good, but it ain't gonna help me with my bugs that I found and didn't receive any help with them.

To be honest, when I told about the scaling bugs, they just told me to use version 2.0.

[eigenbom]

@eigenbom: 6 months and you call me a noob/beginner? Whatever...
Just don't assume that I was creating threads with my whole source code and description like "ain't working, pls fix it" 

Wha? I didn't call you a beginner. I said Lauren'ts bluntness probably comes from answering beginner questions for 5 years.

As for myself, I've been using SFML for 6 months but I've been using c++ for 12 years (and programming for longer), have used countless similar c++ libraries (ClanLib, SDL, HGE, Ogre3d) over that time, and have lectured computer science at a university. Of course, time doesn't really indicate too much, and neither do qualifications, but they are a general heuristic.

I was actually really interested in the length of your experience with SFML and wasn't trying to insult you. If you had posted minimal working examples and had no help, then I agree that that's a major negative for the SFML community.

[ANtY]

It's just too small and there is no experienced users besides Laurent that could help you.

About my length of experience with SFML, I guess it'd 2 years.

[Klaim]

I understand you frustration, but considering the product and my use of it, I don't agree with you at all.

But I still understand the frustration. 

[harkme]

I've been using SFML for a little over a year now and it's been great. I was using SFML 2 since last summer and even though it was in development and rapidly changing, I did not encounter too many problems. Most problems were solved by searching the SFML forums. I have never actually posted there.

[eigenbom]

Fair enough, but it is what it is - a simple media framework.

You can either pick and choose small libraries and combine them, or use a monolithic library that does it all. Both approaches have their benefits and issues. I tend to switch between both, I'll get a bug in a monolithic system that's impossible to debug because of the size of the codebase, so I'll switch to a smaller library. Then I'll get annoyed that I have to write all this extra stuff, so I'll entertain the idea of using a bigger library. And around I go ad infinitum.

I can certainly see how some things in SFML would annoy gamedevs, like the lack of an AnimatedSprite class for instance. I guess in this sense SFML is good for people making simple/small games (like you said originally) or people who want more control and a solid cross-platform base on which to build on.

Also, we'll be seeing an Android port in the next year or two, so that'll be sweet..

[ANtY]

I can certainly see how some things in SFML would annoy gamedevs, like the lack of an AnimatedSprite class for instance. I guess in this sense SFML is good for people making simple/small games (like you said originally) or people who want more control and a solid cross-platform base on which to build on.

I guess that classes like AnimatedSprite would be a better help for ppl making small games, why write everything from the scratch to make a simple game? For "gamedevs" hmm... I don't think that making an AnimatedSprite class is a lot of work, there are much bigger things to write (GUI controls?).

About the Android, right it's cool, sad that at this time I won't be using SFML anymore and I'll have to port my game to Android using some other way before this happens

[eigenbom]

Well no, it's not much work to do a basic animated sprite thingamajig. It only sprung to mind, because it's standard in most other lower level game libraries. Yeh, GUIs, a pain in the butt. GWEN seems alright, but SFGUI seems overly complex. Luckily I'm a fan of really minimal guis in games... :D

[Fungineers]

C - adventure game studio, ren'py, inform, construct, construct 2

Sorry to bring up a slightly dated post in a quote again, but Paul, I was wondering why you rated Construct 2 so poorly?

It's what me and my friend have been using, and though we haven't made anything substantial, except for a few tests and a Ludum Dare game that barely missed the jam deadline for "un-fun-ness" reasons, I think it's a very polished tool, lacking only in the
(ease of)debugging department. 

[ஒழுக்கின்மை (Paul Eres)]

@Fungineers - main reason is that no good games have been made in it, to my knowledge

i judge languages on several factors, the main ones are:

1) flexibility: the variety of stuff you can do with it (3d, 2d, genre variety, shaders, etc.), how easy it is to customize or expand. can you make any game in it? could you have made bastion in it, or super meat boy, or aquaria, or portal, or minecraft? if so, then it's flexible

2) utility: how many libraries, pre-made stuff, and other "batteries" it includes; does it include a collision system, a sprite system, a vector library, whatever. or, if it doesn't *include* them, does it at least *have* them somewhere that you can download

3) platforms: how many different machines you can run it on (e.g. windows, mac, linux, iphone, android, xblig, browsers, and others). is it restricted to one particular operating system on one particular type of machine? if it's multi-platform, how much do you need to change the code when you port it to new platforms? etc.

4) community: how big the community is surrounding something, thus how easy it is to learn and get help on it and to talk with people who know how to use it, how many tutorials exist for it. how active are its forums. how often is it being updated. how friendly is its community. and so on

5) history: how many good games have been made in it, games which show off its power. does it have any blockbusters? does it have a history of great indie games being made in it, or not?

all of those are about equally important. construct 2 scores very low in #4 and #5, because it's so new. that's also the reason i'd score html5 etc. low. unless you are some kind of superhuman you don't want your *first* language/engine to be one that basically is untested and that almost nobody uses. if you're experienced, then sure, try out some experimental stuff. but i wouldn't recommend that a beginner use something that is that new (and my list was intended as a suggestion for beginners, not for experienced game devs)

the ones i gave A's or A+'s to score high in all of those five factors

[Fungineers]

@Fungineers - main reason is that no good games have been made in it, to my knowledge

i judge languages on several factors, the main ones are:

1) flexibility: the variety of stuff you can do with it (3d, 2d, genre variety, shaders, etc.), how easy it is to customize or expand. can you make any game in it? could you have made bastion in it, or super meat boy, or aquaria, or portal, or minecraft? if so, then it's flexible

2) utility: how many libraries, pre-made stuff, and other "batteries" it includes; does it include a collision system, a sprite system, a vector library, whatever. or, if it doesn't *include* them, does it at least *have* them somewhere that you can download

3) platforms: how many different machines you can run it on (e.g. windows, mac, linux, iphone, android, xblig, browsers, and others)

4) community: how big the community is surrounding something, thus how easy it is to learn and get help on it and to talk with people who know how to use it, how many tutorials exist for it

5) history: how many good games have been made in it, games which show off its power. does it have any blockbusters? does it have a history of great indie games being made in it, or not?

all of those are about equally important. construct 2 scores very low in #4 and #5, because it's so new. that's also the reason i'd score html5 etc. low. unless you are some kind of superhuman you don't want your *first* language/engine to be one that basically is untested and that almost nobody uses. if you're experienced, then sure, try out some experimental stuff. but i wouldn't recommend that a beginner use something that is that new

the ones i gave A's or A+'s to score high in all of those five factors

All right, I think that giving Construct 2 a C based on that is more than fair.

But I would just like to point out that in the platforms section it should score nearly top marks because the HTML5 games are playable on iOS, Android, Mac, Linux, and Windows

It will run very well too on all those platforms, once WebGL is implemented and turned on for iOS and Android.

Then for the history section, I'd like to show you this game made in C2 for a competition held by the creators of Construct, with the theme of rotation. http://www.kongregate.com/games/sqiddster/airscape

EDIT: I also agree that starting out with, and putting all your eggs in the HTML5 basket is sketchy. But thankfully both of us are Lua and Python native programmers, so we'd be alright if HTML5 doesn't turn out like we hope it does.(Overtaking flash and all )

[ஒழுக்கின்மை (Paul Eres)]

i just tried out that game on kongregate, but i think my computer and/or browser may be too slow to run it, because i got crazy slowdown (about 6 fps if i had to estimate) rendering it sort of unplayable; i got up to the first thing that you need to 'smash' and then it slowed down too much to continue playing

[Fungineers]

i just tried out that game on kongregate, but i think my computer and/or browser may be too slow to run it, because i got crazy slowdown (about 6 fps if i had to estimate) rendering it sort of unplayable; i got up to the first thing that you need to 'smash' and then it slowed down too much to continue playing

Yeah, I don't exactly know what happened, but the game's audio pops for me. It's a problem I haven't seen in any other Construct or HTML5 games though, so I'm gonna guess it was something the creator did. Also running the game in Chrome may help. It's the best performer with WebGL (which may be disabled, causing the problem).

[ANtY]

Same as Paul, but on my PC it's a slideshow already in the main menu 

[Tumetsu]

Game worked very well on my laptop with Chrome.

[Fungineers]

Yeah, I'm pretty sure it was a problem with something the developer did. My Ludum Dare 23 game ran on my iPad 3 and crappy laptop at like 300+ FPS with around 250 objects on screen.