I've been looking through and there are a lot of choices. I guess if you want to spend 100 dollars, Impact is the best, but I don’t since I'm a hobbyist. So far I like the look of Phaser. It’s fast and been around for a while, works well on mobile. I'm going to surrender on my Unity game at least for now. It’s a little too big of a project for me. I thought I would focus on 2d puzzle type games for a while. Something I could finish sooner and offload on some site. If you write your own engine, just give other advice, or reasons not to use an engine, or not to use html5. I'm not fussy. I could do Unity 2d, but the project size is a bit larger for the type of game.
Sounds like you want to jump onto html5. a 2d game is in no way too big a hurdle for single person. even a 3d game if you focus on a single level at a time. why not finish a level or two of the unity game. then make an html5 version of it. dropping unity all together. does not make sense. dont focus on popularity, finish some aspect of the game.
I really can’t finish some aspect of the game. I may not give it up, but for now I’m shelving it. It’s too large a project and I can’t push myself to work on it anymore. I should have learned by now, but I wanted to do a game with a larger story. Unfortunately, that meant getting stuck in the same trap I usually get stuck in. I’ll give it some more thought and see if there is a way I can possibly make it smaller, like getting rid of the minigames I normally put in an adventure. Right now, though, I just need a break from it. I need to actually finish some small games and get back more into a cycle.
As far as popularity of html5. It’s just a framework that suits the game size. I don’t want to use a sledge hammer on a finishing nail.
You could consider using Haxe and a framework like HaxePunk or HaxeFlixel. Haxe can compile out to Flash, native executable, mobile, and I believe HTML5. It probably will serve your purposes well - just be sure to follow the tutorials pretty closely if you run into issues.
I've looked at Haxe off and on. I just wonder if it could really work "well" on all those targets. The thing I find interesting about html5 engines like Phaser is that they use opengl ES pretty well, and run pretty fast. I looked at another multi-target engine and it seemed very slow running html5. Html5 will run on quite a few platforms on its own, so it somewhat eliminates the need for multi platform engines.
Hmm, yeah. It would seem HTML5 is actually considered a bit experimental for HaxeFlixel; my apologies. It would seem that it is possible, though, from their about page.
GameMaker Studio. Yeah, it's pricey, but it's simple to use and it exports to HTML5.
Yoyogames has been really aggressive with their price lately and have literally decided to give away GM Studio. So, it might be worth looking into again.
The only other HTML5 engine that I've played around with is EnchantJS. Still in development, but it's free on GitHub.
The problem with that logic is that things work differently on different browsers, operating systems, devices, etc. So, I can test and retest on different devices, or I can use an engine which no one will notice in load time anymore because it's still very small compared to the rest of the game assets. Granted, the functions are specific, but any of us are pretty used to calling other peoples functions by now. Also, it's just basically a waste of time to redo collision detection etc, and what about if you want to use physics. People have spent years on those physics engines. Do I write my own physics engine also? What about all the sound problems known for html5? Sure, I can write a very simple game with it, but I don't think that type of thing is as useful as it once was. We all are used to using other people's libraries now days.
Continuing with my experiments with html5 engines, I've started using Pixie. It's a nice alternative to a full engine as it only does rendering and is supposedly the fastest, according to Pixie. Phaser is built on it, but I've found that Phaser's choices kind of get in the way. Their simplest physics engine isn't simple enough so I'm constantly doing workarounds and it frankly doesn't work well anyway. If I just use the collision, it's extra overhead and still doesn't do what I want. There are also sound engines and physics engines around so I think adding only what I need would be more modular and re-usable, and also smaller.