The whole "entire level as giant map" idea has been done though. Actually take it further: two giant maps. One map contains the foreground as usual, the other map contains the collisions (i.e. floors, walls, ceilings, usually this map is half resolution since you don't need that much detail). And yeah, it isn't really optimal...

Kinda related, Metagun used PNGs for map data. Granted it *was* tile-based, each tile just one pixel, but I bet you could actually have a full-size map as one honkin' PNG, provided it was *only* for collision/spawn data. There wouldn't be much waste, because compressing sparse, few-color bitmaps is exactly what PNG does best.

I wouldn't dare do that for the foreground though. How would you animate it?

NSIS is not that hard to use, really.

But I feel like, in this day and age, we shouldn't have to choose between installer and archive. A self-contained executable, i.e. zip file that just extracts its damn self and runs when you double-click it, isn't rocket science.

I mean, a web page is basically that, plus invisible network magic, and we all take that for granted.

I could probably go and invent something like that, but instead I'll sit here in my armchair until someone else does.

I was wondering how long it would be before the pad vs. mouse turf war broke out. I'm proud of you, TIGS, for lasting this long. <3

On topic, once I got it working, I remembered how awful mice are for movement (as opposed to aiming). Remember mousemove in Quake? Yeah, no.

Maybe some weird half-controller half-mouse thing is in order. :/

I would definitely do that. I'm browsing ManyMouse source as I integrate it, and it's really quite a fine example of how to neatly expose an API without letting the details of multiple implemenations (there are five) leak into it. Super clean, clearly annotated code. Of course, it's anything but object-oriented

Unfortunately the one thing that really needed documentation, the event struct, has none at all. (Actually, the whole API has no direct doc, but e.g. ManyMouse_Quit" is kinda self-explanatory.)

From what I can tell based on the examples, the `item` field is 0 or 1 depending on x or y axis (and meaningless for non-motion events), the `value` field is some arbitrary int bounded by `minval` and `maxval`, and you may get notified of either absolute or relative motion, or both, depending on the available implementation.

Yeah, it looks like Icculus has it covered for all three platforms, for ten years already. Not only that, but

You don't have to build this as a shared library; we encourage you to just
 compile the source and statically link them into your application...this
 makes integrating ManyMouse much less complex.

Sage advice!

In theory it should be possible on Linux, but in practice...it might require manual configuration on the part of the user

You just described trying to do anything in Linux :D

A quick google led me to this http://icculus.org/manymouse/, no idea if it works or how nice it is but check it out.

Wow, that's....pretty much exactly, precisely what I need. What exactly was your search term? Teach me, sensei, for my google-fu is lacking.

Schrompf, it sounds like you'd enjoy the WTFPL.

And yeah, I really need to find some time to get up to speed with gyp/ninja. I don't see CMake going anywhere but downhill, unfortunately.

Can't talk about it, it's part of my top-secret scheme to world domination.

Uh, pretty simple, I want to use it as a control scheme for Vec. One mouse for crosshair, the other for moving around it in polar space. Degrades nicely to dual-stick gamepad controls (though mouse is richer and more precise than a stick), or mouse/WASD if we must.

Beyond that, it seems like a relatively unexplored area (as evidenced by the wholesale lack of prior art) with some potential for novel gameplay. Plus, I've always had a soft spot for symmetry.

There's some promising links at http://gamedev.stackexchange.com/questions/24046/is-it-possible-to-detect-two-different-mice-at-the-same-time-and-have-their-mov

Has anybody been able to successfully grab input from more than one connected mouse? Being able to distinguish one from the other, of course, I mean.

I've googled around but most of the results are for wonky OS addons that put multiple pointers on the screen, whereas what I want is to control one game with two different mice.

I'm not afraid to get down and dirty with device APIs, but I'm not sure where to get started with that...and I'd prefer to stay cross-platform.

Halp?

I'm browsing forums instead of makking gamm.

Also, I'm putting all my crap in one monstrous, unmodular header instead of splitting it up into ten headers like a good programmer would do. So, instead of clawing my way out of #include hell, I'm being lazy till I feel like working on the input scheme.

I guess it all evens out.

There is also (well, of course!) a subreddit http://www.reddit.com/r/proceduralgeneration/ that I didn't know about until now.

Procedural generation of what? It's a rather broad space.

I, for instance, am into procedural graphics (shaders, filters, particles) and audio (real-time synthesis and algorithmic composition) but a little less wild about generating levels, terrain, or NPCs.

What excites you?

Anybody for a pre-jam light lunch, say around 3pm? I've heard the cherry pie at Mel's is legendary. Perhaps bum around YBG prior to heading to Hryx's place.

My phone number is under contact at the url below, and my handle is my gmail.

I cut my teeth on SVN but gradually shifted to git. I'll never go back. The learning curve is brutal because (a) DVCS follows entirely different paradigms that are not intuitive; and (b) the commands are named stupidly and have no built-in abbreviations. HOWEVER:

Once you get critical mass with git comfort level, you can be more productive than with SVN (if you're on Windows, do yourself a favor and use TortoiseGit). Plus, GitHub is totally dope. I can't speak to Mercurial (hg) but word on the street is it's even better.

I use Perforce at work and I hate it. Its only saving graces are being reliable and scaling with enormous codebases, which may be important to some enterprises, but is certainly not for me. It's clunky, prone to bad state, has a learning curve on par with git's and it's easily to accidentally attempt to pull down the entire repo.

Bejujular!

I made a reservation at Bourbon & Branch (a speakeasy) for Monday 8pm but it's looking likely (not guaranteed) that one of my buddies will no-show. Drop me a PM if you're down for sipping a fancy cocktail with a random TIGger and his random-er college mate.

I tried setting up appveyor to get some auto Windows builds for my CMake-based project...

A game needs a story like a song needs lyrics.

"They're not artists because nobody can play the guitar!"

?

Heh. It's a quote out of an impromptu interview that was notably sampled by Skrillex. The upshot being, no, obviously you don't need to play guitar to qualify as a musician, whatever anyone's specific tastes may be. It doesn't mean everyone has to love bro-step, but refusing to recognize it based on on such narrow criteria is (IMHO) rather petty and short-sighted.

https://duckduckgo.com/?q=they%27re+not+artists+because+nobody+can+play+the+guitar

tl;dr I agree, 100%!

[Combat Flight Sim + Sim City]

"Mortal Kombat Flight"? I dig it. On that note...

Combat Flight Sim + Sim City

Two players, asymmetric. Avoid AA, find the capitol and bomb it.

A game needs a story like a song needs lyrics.

"They're not artists because nobody can play the guitar!"