Deth's Journal
"There Is Nothing To Understand."
Ôdishon (1999) (Audition)
I've provided a link to the IMDB for this; but i urge you not to check it out there. I urge you, in fact, to not look at *any* cover-art related to this film at all.
This is a movie that should be walked into like a dark alley - apprehensive, nerve-inducing. It's meant to shock, to startle, to horrify. I urge you all to see it, as it is VERY good. There's nothing in western cinema that even comes close. Hostel, Saw, Halloween - these WISH they were as good.
I've provided a link to the IMDB for this; but i urge you not to check it out there. I urge you, in fact, to not look at *any* cover-art related to this film at all.
This is a movie that should be walked into like a dark alley - apprehensive, nerve-inducing. It's meant to shock, to startle, to horrify. I urge you all to see it, as it is VERY good. There's nothing in western cinema that even comes close. Hostel, Saw, Halloween - these WISH they were as good.
On Project Bob, they allow bots. I think a pretty great bot to make would be a "fishing" bot. So, let's take a look at what it'd take to make a fishing bot that - *at least* - didn't get itself killed all the time.
The first step would be the boring one I'm about to skip - writing up all the triggers to observe the state of fishing (of which there are 2 primary parts). This is essentially a number-punch, and isn't critical to figure out right this second.
The second step is to figure out *what* the states should be. My guess:
1. Line
a. No Tension
b. Light Tension
c. Moderate Tension
d. High Tension
e. Emergency Tension
2. Hook
a. Emergency Set
b. Weak Set
c. Moderate Set
d. Good set
e. Perfect Set
This encompasses all the possible states. Fun fact: it'd probably behoove me to *reverse* the numerical values of hook set, so i can reuse code. Also, i could easily find a 2-digit number that described the state, like this: "11" would be "No tension, Perfect set", while "15" might be "No tension, Emergency Set". there are a total of
The next stage is to figure out what to do in each state. In this case, the two states are dependent on each other, but in an odd way. If the line is taunt, i cannot set the hook (it would increase the line tension). If the line is slack, i can set the hook.
So, a simple rule might be: "If the line tension is moderate or less (ie, 3 or less) and hook is not set above Emergency, Jerk hard. If line tension is above moderate, reel slack.
Another idea is to use heuristics: "jerk hard twice, then reel steady unless tension goes above 3, in which case we reel slack until it is below 3."
The final stage is actually the trickiest: we have to gauge our success. Why do we gauge our success? Simple. Without knowing how well we are doing, we cannot know if we're doing the *right thing*. For example, maybe a better rule is to reel steady if the tension is below emergency tension. If we fish for a few days and find that we're not getting the values we want, we might tweak it, and see if that improves (or lessens) our chances. It might even be useful to set up a random factor.
All told though, the main thing i need is:
1. A mud client with variables.
2. A mud client that allows (reasonably) complex input parsing
3. A stable, and reliable, mud client.
I'm not sure what I'll do about all that. On the one hand, i could set this up *fairly* easily in ZMud. On the other, i could leverage my other computer by finding a mud client light weight enough to do this. On yet another, I could set up a *separate* mud client to do this on this machine - an option which has much to recommend it.
I'll have to ponder the specifics, a bit. To be sure though - the fishing idea is a good one, as it'll quickly gain levels and allow for easy setting up of a "baseline" character.
The first step would be the boring one I'm about to skip - writing up all the triggers to observe the state of fishing (of which there are 2 primary parts). This is essentially a number-punch, and isn't critical to figure out right this second.
The second step is to figure out *what* the states should be. My guess:
1. Line
a. No Tension
b. Light Tension
c. Moderate Tension
d. High Tension
e. Emergency Tension
2. Hook
a. Emergency Set
b. Weak Set
c. Moderate Set
d. Good set
e. Perfect Set
This encompasses all the possible states. Fun fact: it'd probably behoove me to *reverse* the numerical values of hook set, so i can reuse code. Also, i could easily find a 2-digit number that described the state, like this: "11" would be "No tension, Perfect set", while "15" might be "No tension, Emergency Set". there are a total of
The next stage is to figure out what to do in each state. In this case, the two states are dependent on each other, but in an odd way. If the line is taunt, i cannot set the hook (it would increase the line tension). If the line is slack, i can set the hook.
So, a simple rule might be: "If the line tension is moderate or less (ie, 3 or less) and hook is not set above Emergency, Jerk hard. If line tension is above moderate, reel slack.
Another idea is to use heuristics: "jerk hard twice, then reel steady unless tension goes above 3, in which case we reel slack until it is below 3."
The final stage is actually the trickiest: we have to gauge our success. Why do we gauge our success? Simple. Without knowing how well we are doing, we cannot know if we're doing the *right thing*. For example, maybe a better rule is to reel steady if the tension is below emergency tension. If we fish for a few days and find that we're not getting the values we want, we might tweak it, and see if that improves (or lessens) our chances. It might even be useful to set up a random factor.
All told though, the main thing i need is:
1. A mud client with variables.
2. A mud client that allows (reasonably) complex input parsing
3. A stable, and reliable, mud client.
I'm not sure what I'll do about all that. On the one hand, i could set this up *fairly* easily in ZMud. On the other, i could leverage my other computer by finding a mud client light weight enough to do this. On yet another, I could set up a *separate* mud client to do this on this machine - an option which has much to recommend it.
I'll have to ponder the specifics, a bit. To be sure though - the fishing idea is a good one, as it'll quickly gain levels and allow for easy setting up of a "baseline" character.