Build an HTML5 game in 30 minutes

You're right. I'm 100% beside you on the value for time spent, but as you've mentioned there's also the 'learning value' that comes form building it from the ground up. (Think of how much I'd have to spend to learn it in a University course seen another way, I'm actually saving a few thousand!).

Going with an engine also carries with it the weight of engine choice as well. Your investment is in learning how to do it from within that engine - versus - knowing how it's done at the 'raw-guts' level.

There is likely a point at which I'll be better off cutting and moving across to an engine. (Though with the dozens of engines available deciding that IMPACT is the right one to go with isn't quite clear cut, for example, there's the ISOGENIC engine on its way, https://github.com/bebraw/jswiki/wiki/Game-Engines is the list on GitHub)

What about Rocket Engine? Maybe I should wait for that? What about Unity3d? Or Flash? Cocos2D (on github, widely used, BSD license)

I'm honestly not trying to be facetious - but these are actual problems that run through my head when trying to decide. So I suppose maybe I got so stuck in the 'analysis paralysis' phase on the picking an engine phase and decided to look through what I could figure out on my own.

Each puzzle brought with it a single interesting question (http://chesser.ca/jsSprite shows the order, beware step 12 it has crashed many weaker computers).

There's also a number of interesting 'next steps' I'd like to try to figure out in a semi-assisted state (like on my own, but I get to ask experts when I've gotten really stuck). I'm really kind of excited about getting to look at something called 'quad-trees'

And then there's all the setting up NODE.JS and WEBSOCKETS for adding multiple mobs to the screen at once.

It's less about the finished product than it is about building valuable full-stack developer skills. With node specifically, I'm going to have to build an AWS instance and set up the brains of my operation... not to mention sending the correct data back and forth! (i'm getting a little giddy thinking about it).

Placed in a broader context - I'm employed full time doing other stuff and would love one day to ship a working project - but at this phase am not totally ready to rely on 'magic functions'.

If it's possible to give a short answer after all of that, you're 100% right & I agree 100% ... and I'm doing it the hard way because I'm a stubborn, dorky nerd (but my lights don't go out if I miss the ship-date).