Oh goodness.  How could we decide it.  Building a mod is a crazy thing that requires so many different aspects coming together to form a conglomerate whole.

Phase 1: Firstly there's the concept.  This is the initial time when a few friends get together and come up with a game they want to play.  Usually it involves thinking about what games they've enjoyed playing in the past, and what subjects they'd like to include.  For us it's rather obvious: star wars. 

Phase 2: The next step is content creation.  This actually goes on throughout the mod, however the starting of it is extraordinarily difficult.  Creating an object for a game is an intense process in itself.  First step is modeling it (requiring someone who knows how to model), then - assuming it is clean - it gets unwrapped and skinned.  After that it's structured in the game's heirarchy and an initial layer of code is applied onto it.  Usually this process involves a minimum of three people (modeler, skinner, coder) though depending on the difficulty of the object and the skill of the team it can be a little as one or as many as five.  For those groups that start off having never done it before, it's a massive mountain to climb before they become a well oiled machine and skilled enough to throw things into the game.

And, sadly, about 99% of all mods die at this point.  Probably the most difficult thing in mod-building is finding all the needed skillsets to get objects in-game.  Multitalented and experienced people are thus modding gold when it comes to recruitment, as they can be put to work almost immediately and can blow through content creation far faster than those that haven't built content before.  I don't mean to discourage anyone, that is.  Indeed, if you're keen to get in on game building then modding is the best environment for you.  The best exercise is doing though, so actually trying and learning how will take you miles.  There are great tutorials out there for how to build a mod from start to finish, and nothing says actual commitment like being able to demonstrate samples of your work

While the content is being created, mapping also begins here.  Mapping in itself is its own art form.  Especially with BF2/2142 and its buggy editor.  There are dozens of bugs, problems, issues, and pitfalls that catch mappers up, and each one usually ends with the mapper scanning through archives of forum posts in an effort to figure them out.  What's more, the mapper generally requires some knowledge in modeling (at least for the export process) so that they can work with the content team in getting stuff into their maps.

On top of this the mod leader generally needs to know everything.  Seriously.  Every step, from the first poly to the last line of code and all the social skills required to explain it to everyone.  They are the backbone of building a mod and the person everyone falls back on in times of need.

Phase 3: Once objects start appearing and a map is laid down to play on, in-game balancing can finally begin.  This is where the coders take over and start shaping the raw content into a playable form.  This also has its share of problems to overcome.  Firstly, having more than one coder means the project starts getting spread out.  Systems to aggregate work and keep everything at the latest version become necessary.  This has a few options, from the mod leader keeping track of all the resources in and out to CVS, and any flavour in-between.  Indeed, FS uses a system of successive builds which merge together with an autoupdater program I wrote.  It has its pros and cons, but was created with the singular purpose of rapid development in a distributed environment in mind.

Then there are also issues of trying to translate concept into actual gameplay.  Some things have to be left behind as they are just impossible to do in a game.  Other things sprout up as coders figure out how to do new tricks and cool features.  Occasionally something has to be dropped down again to be redone as it is found not to fit with the current scheme.  All the while the content team is still churning out new stuff so it's an ever increasing push to get things into the game.

Phase 4: Finally, once maps, statics, characters, weapons, and vehicles are all added and balanced (you're looking at the process of model-skin-heirarchy-export-code done a hundred or so times minimum) a beta tester team is formed up.  While some testing has obviously been done by the team beforehand, there is nothing like pouring players into a map to show all the bugs.  And oooh, there are bugs.  Usually it takes about a month of hard squeezing to get them cleared out.  Sometimes longer.

Phase 5: And, after that, we release the game.

As you've probably guessed, we're currently in phase 3.  We're adding stuff on a regular basis and we've also hotwired a few steps along the way to speed up our development process.  Our tools allow us to test stuff that comes hot-off-the-editor.  We have an alpha team whose job is to try out countless variations and versions of our work, and provide feedback on its play so we can fix/balance and move on.  This means things will be done by the time we get to phase 4, rather than throwing it all together then trying to hammer out a hundred bugs at once.  Finally, our team has become a well oiled machine.  This is definately one of the best teams I've seen working on a mod, with a lot of dedicated individuals.  Especially some of our oldies, like Blitz and Bandu

So yeah.  It's hard to put a time scale on this.  We're constantly adding new content whilst simultaneously testing and tweaking it, however it's still a climb to go.  As one of our other devs once put it "The first release is always the hardest because it's all about building the thousand things needed to actually play.  The second release is always better because after the first release you can sit back and say 'what can we now do to make this mod even cooler?' rather than 'what do we need to make it playable?'"

Fear the Woodzilla