done.txt

I seem to get stuck in a trap where I'm overwhelmed because I don't know what
to work on and the game gets no closer to being done. Often, I tear stuff out
or refactor things that end up later getting deleted. It feels like I'm taking
a random walk through the design space and getting no useful feedback during
the process, so I'm not even necessarily making progress.

Part of it is when I'm not sure what to work on next, I try to play it and see
what's not fun. But it's actually really hard to tell *why* it's not fun when
entire systems aren't there.

Another trap I get stuck in is, "I should add more content so I can see if this
system is working." "No, I should get all the systems in place since changing
a system unbalances everything."

I believe the way out of that deadlock is to realize it's possible to add new
systems and content that don't unbalance everything. Adding a new dangerous
monster or move displaces other monsters and moves at that depth. Adding a new
skill takes points away from other skills the player could have.

The state I really want to be in is that I have a complete, OK game, and then
I get to incrementally tweak and refine it and add new features. But I'm not
there, so acting like I am (constantly rewriting new dungeon generators,
deleting various game mechanics, etc.) probably isn't the right path.

It's like I'm trying to get to a good complete game by focusing on good and
incrementally fixing the "complete" part. But there's no way to evaluate where
an incomplete game is good or not (beyond telling if the code itself is nice).
That's not the right path. I should get to complete first, then ramp up good.

In other words, make a shitty complete game, then make it good. Just like
writing, don't polish the first chapter over and over.

Here's a plan:

## Phase 1 - Vertical slice

- Stake out a minimal list of features/systems I think the game probably does
  need.
- Implement them. They don't have to be great, but just passable. The code can
  be ugly.
- Stake out a minimal list of content needed for a full playthrough. Again,
  doesn't need all the bells and whistles. Monsters don't need all their cool
  moves. Items don't need all their neat prefixes. Just enough so that you can
  technically play to the end. Some depths may only have one monster.

- Add winning, and the basic framework to play a complete game.

This is the *first draft* of the game. Now is *not* the time to revise any of
the above. Just barf it all out. No deleting.

No adding fancy features to make it more fun. No non-essential features.

## Phase 2 - Playable

*Then* start playing it. Try to play all the way through with a minimum of
changes. Do not add features. Avoid removing or changing features. Do tweak
content to roughly balance it.

The goal here is a bare bones but playable game. It might be boring but
shouldn't be wildly unbalanced.

## Phase 3 - Good mechanics

Now we should have enough context to actually evaluate which game mechanics
do and don't work. Remove or replace things that aren't fun. Try not to spend
too much time on this.

Delete dead code.

## Phase 4 - Refine

This is the payoff. Now we add new gameplay behavior and content to enrich the
world and play experience. More monsters, equipment, items, effects, moves,
tiles. More sophisticated level generation, etc. Tweaks to the user interface.

Ideally, this phase then lasts forever.

---

So, for phase 1, what is the minimal set of stuff needed?

- Stats. They affect all of the other combat and play mechanics.

- Skills. This is the fundamental system for character advancement, which is a
  key part of a playthrough. We don't need all the skills. Just enough to
  produce a strong enough hero to win. Definitely don't need skills that cover
  all "class" archetypes. Probably:

  - Archery so that bows can be used at all.
  - Enough magic skills so that mental attributes have value. It should be
    possible to win without heavily investing in strength.

  - TODO: Consider unifying attribute and skill points? Hmm, doesn't feel very
    fun. But it might make warriors (high attribute, low skill) a more
    meaningful type of character without having to make up a bunch of warrior
    skills. Basically, it would let players choose whether to focus on beefing
    up stats or unique abilities.

    Let's try it. This means "strength", etc. are attributes but also skills
    that the player can increase, so that we can use the same UX for skills and
    boosting attributes. This also lets us have skills that require certain
    attributes as prereqs.

- A few more monsters at greater depths.

- Final boss and winning.

For now, we don't need to worry about:

- Hunger. Maybe redo it in phase 3, but it's not causing any problems now.
- Monster AI. The sleep/awake stuff is meh, but it's working.
- More level gen stuff. The encounter stuff is still a work in progress as is
  all the biome stuff. But it's not preventing someone from playing the game.
- Lighting. The open river stuff encourages limited light ranges, but it's not
  needed.
- Better water mechanics. What's there now is temporary, obviously, but it
  works.
- More affixes. Could potentially disable them entirely for now, but adding
  them later will require rebalancing monsters. That may be OK.

---