I don't know if I have a set "plan" for WFA. I picked up on coding WFA after Static/Zitheral gave me the code, and my son had an interest in quake3. He really got into it. (This was before the advent him receiving his x-box). I also got the code for True Combat sitting around, which I still pick up from time to time.
I wanted to optimize code. I wanted to refactor everything. So I did. It now runs less intensive code with smaller requirements than anything before. So I moved onto adding all the weapons from Q2WF. They ported easily, and performance STILL was better than before I touched it.
So I moved onwards to the AI.
I wanted to fix all the problems with the bots. I wanted the AI to be perfect. I wanted a complete simulation of a real player, or as close to it as I could get. If you ever looked at the code you'd see that I split out most of the bot's AI from whatever it was, 8 files, into a few dozen. I just took what Mr. Elusive did and expounded upon that in so many different ways.
Oh I hit some roadblocks, like the bots using the jetpack, but I think I got that down now, at least in theory. Once I get the jetpack settled down the grapple becomes more "WFA" like instead of just a substitution of movement, the grapple becomes an extension of the AI's movement. It's some complicated shiznit.
Once the Q3 engine (and mostly BSPC) was opene source I had already gathered most of the information by reverse engineering what did what, and where stuff went and how it worked... and qfusion helped prep me for the engine.
But always, always in the back of my mind, some stuff really irked me because I always thought that single player was the key to getting new users to play and get experienced with WFA. New users to the game would otherwise get slaughtered and turned off by it. Or they'd practice on a server, all alone, without being able to experiment with all the weapons and specials.
I wanted to change that, so I went all-out with the AI. A shy, new person can play match play single player, *then* decide if they want to hit a pub. That's what Q3 was in the first place, that's what WFA should've been.
WFA is a hard game. And difficult to code, at the beginning. But even now my son picks it up and plays on the lan server every once and a while, when he sees me working. He jumps in on his linux box and plays around with me. (He loves the arsonist)
I'm making the sentry dockable because he wanted to see what it'd be like "from the sentry's point of view". Ain't too much going on there, but it allows me to experiment with the ui and cgame, and show stuff that would otherwise only be known by mappers (like showtris of a sort) and shows off the engine. I just gotta get it so you can select which enemy to track or steer it / guide it yourself. Basically more AI to deal with.
The only thing stopping WFA from being on it's own is yea, the models and sounds of the players. I looked around for open-sourced models, looked at Open Arena, and went "bleah" because there's not too much I can do in that area. I'm not a modeler. I can tweak what's already there, but it's not my specialty.
I can map, and created a few maps but they were ones that others already came up with, I just made them more efficient for the engine and the bots.
I started picking up the WF tools that Static/Zith gave me access to and created a basic map right now, and by basic, I mean no textures.
Just the blue and white. It's like citadels but not really. I'm still getting used to the editor itself.
The tools are tough to work with when you've been so used to GTK/Q3Radiant, yet the idea is there. What's driving me is the fact that a lot is coded already, and Aries and Rav have made some amazing looking stuff.
But will it be argued that it's a TF2 clone, when in fact TF2 is sorta like a grown up WFA kid? I dunno, but I'm gonna try helping them out as best I can; the demands on a dev machine, however, are enormous compared to what's needed for Q3-based games.
I don't know if WFA will have a future after this release except as a study of what can be done with the Q3 engine, because by god, it HAS to be the most intensive use of the engine out of any other Q3 derivitive.