19 Dec, 2008, David Haley wrote in the 21st comment:
Votes: 0
Lyanic said:
Why is a PC necessarily more interesting data than an NPC?

Because the PC has a player controlling the actions, whereas NPCs don't?

Lyanic said:
Why have one set of rules for PCs and another for NPCs?

I don't think anybody suggested different rules: the suggestion was just to not group them the same way, because players are more interesting (less predictable; they're your friend and you want to see what they do; etc.) than NPCs.
19 Dec, 2008, Lyanic wrote in the 22nd comment:
Votes: 0
DavidHaley said:
Lyanic said:
Why is a PC necessarily more interesting data than an NPC?

Because the PC has a player controlling the actions, whereas NPCs don't?


That's essentially just restating the acronym: PC, NPC. It doesn't give a valid reason for why the data of one is more interesting than the other.

DavidHaley said:
Lyanic said:
Why have one set of rules for PCs and another for NPCs?

I don't think anybody suggested different rules: the suggestion was just to not group them the same way, because players are more interesting (less predictable; they're your friend and you want to see what they do; etc.) than NPCs.


Not grouping them the same way is a form of applying different rules. But, it doesn't stop there. Just about every single MUD in existence applies different rules to NPCs than are applied to PCs (and I'm not talking about the commands coming from a socket connection vs. an AI module). In fact, Diku was practically built around treating NPCs differently from PCs.

As for the other point, I find players boringly predictable at times and NPCs don't necessarily have to be if they're well written. The only real distinction in terms of unpredictability is communication. NPCs are very predictable in that aspect, where PCs aren't. There's not much that can be done on that front until we get a system that passes the Turing Test, though.
19 Dec, 2008, David Haley wrote in the 23rd comment:
Votes: 0
I don't understand where you're coming from. A player is almost by definition more interesting than an NPC because you make a game for players, not for NPCs to run around by themselves. Therefore, you care a lot more about what players are doing than each individual NPCs in a horde.

Lyanic said:
Just about every single MUD in existence applies different rules to NPCs than are applied to PCs (and I'm not talking about the commands coming from a socket connection vs. an AI module). In fact, Diku was practically built around treating NPCs differently from PCs.

I consider representational questions to be significantly different from applying, say, different combat rules. I fully agree that doing that is confusing and undesirable. As I said, though, I believe these are completely different issues.
19 Dec, 2008, Lyanic wrote in the 24th comment:
Votes: 0
DavidHaley said:
I don't understand where you're coming from. A player is almost by definition more interesting than an NPC because you make a game for players, not for NPCs to run around by themselves. Therefore, you care a lot more about what players are doing than each individual NPCs in a horde.


Going back to the previous part of the argument - if you keep identical data for both PCs and NPCs, thus making them indistinguishable, then one set of data should not be discernible from the other, rather less one being more interesting than the other. That was my main point. And yes, the game is made for the players. But, should immersion in the game not be an important factor? This argument stemmed from discussing how PCs are represented on display versus how NPCs are represented. Yes, it matters what an individual player is doing - but why should it matter to anyone other than that player? At least, why should it matter more so than what an individual NPC is doing? Again, I'll rule out communication, but I got the impression we were discussing more along the lines of position/movement/combat, anyway. We would also be talking about representation from a third person point of view. In this case, a player is not likely to see that representation of himself/herself. Now, it just boils down to how one player sees another player represented. Why should this be handled differently from how the NPCs are seen? Does it not ruin immersion to have something like the following?:

NPCs:
Garin, the Knight
Melcor, the Knight
Dast, the Knight
Milnis, the Knight
Andrus, the Knight
Terrak, the Knight
Baph, the Knight
Zent, the Knight

PCs:
Player1, the Knight
Player2, the Knight
Player3, the Knight
Player4, the Knight
Player5, the Knight
Player6, the Knight

'look'
<Display for Player1:— —>
Standing in the Castle Courtyard
There is an open, grassy field, completely surrounded by gray stone walls.
8 Knights are standing here
Player2, the Knight is standing here
Player3, the Knight is standing here
Player4, the Knight is standing here
Player5, the Knight is standing here
Player6, the Knight is standing here

My point is simple: If you group the NPCs, then it makes sense to group the PCs, too. Otherwise, it just looks silly and it ruins immersion.

*Note: Insert whatever names you wish for the Players. I got tired of making up names.

Also, I'll go ahead and make the counter to one of the likely arguments I expect: Players care because they want to play with other players. But, what if they can't tell the difference? What if the player logs into the game, starts playing, is interacting with other characters (who happen to be NPCs), and is having fun? Later, this player discovers that he or she was the only live person on the whole time. Does this invalidate the fun that was had? Sadly, and irrationally, for some players it will.
19 Dec, 2008, David Haley wrote in the 25th comment:
Votes: 0
Lyanic said:
Otherwise, it just looks silly and it ruins immersion.

I don't think this has anything to do with immersion. It has to do with information management, and not getting spammed to wazoo due to being in a huge battle.
Lyanic said:
But, what if they can't tell the difference?

You answered this yourself:
Lyanic said:
Sadly, and irrationally, for some players it will.

It doesn't matter that in your view of rationality, it shouldn't matter if they're playing with people or not. If it matters to them, and they won't come back because of it, then it should matter to you as the game admin.

Incidentally, I challenge the assertion that in some kinds of games you can have as much fun playing with NPCs as with PCs. People typically provide more challenge or at least interaction than NPCs. This will be the case until AI improves dramatically, which I don't see happening any time soon.



Anyhow, this is all kind of irrelevant. The OP explicitly wanted to be able to do this:
Gehrig said:
I wanted the PC's to still be able to attack on their own. As a single player. It seems that grouping mobs won't work in this case. Just to clarify, the PC would only order their particular squad. Also, a PC would have the option to go out on his own (but shouldn't in most cases). They would still function as a single PC.

I guess you could make the case that he shouldn't want to do this, and that could be an interesting discussion to have, but I would argue that that discussion should take place in another thread so that we can focus on the representational questions here.
19 Dec, 2008, Lyanic wrote in the 26th comment:
Votes: 0
DavidHaley said:
Lyanic said:
Otherwise, it just looks silly and it ruins immersion.

I don't think this has anything to do with immersion. It has to do with information management, and not getting spammed to wazoo due to being in a huge battle.


Except my way is diminishing spam. You're advocating MORE spam.

DavidHaley said:
Lyanic said:
But, what if they can't tell the difference?

You answered this yourself:
Lyanic said:
Sadly, and irrationally, for some players it will.


Irrationality is not a valid reason for anything.
19 Dec, 2008, David Haley wrote in the 27th comment:
Votes: 0
Lyanic said:
Except my way is diminishing spam. You're advocating MORE spam.

By calling it "spam", you are saying that people don't care about it. But, obviously, some people do seem to care about it. So… :thinking:

Lyanic said:
Irrationality is not a valid reason for anything.

Human beings are irrational and that is a fact that must be lived with. (Just ask anybody in marketing or consumer-related activities in general. Or psychology. Or HCI. Or …)

Anyhow, you are of course welcome to craft a game that only caters to people with your sense of rationality. (I don't agree that you're actually describing "objective rationality", but that's neither here nor there.) But part of designing a game is figuring out what people think is fun, and if their fun depends to some extent on "irrationality", that means you need to deal with that. :shrug:

Anyhow, I vote for returning to the topic at hand…
20 Dec, 2008, elanthis wrote in the 28th comment:
Votes: 0
This is intellectual masturbation. This is the kind of question people in academics ask, not the kind of question people who actually want to make a real game ask. :/

There is a very, very clear distinction between PCs and NPCs. PCs exist because a player is trying to have fun playing your game, while NPCs exist in order to entertain the PCs. Without PCs, NPCs have no reason to exist at all. Likewise, without players a MUD is missing the rather key "multi-user" component.

NPCs are worthless without PCs. An NPC's meaning is defined solely by how it is perceived by players. Without a player observing your NPCs, all the NPCs are are electrons moving around through circuits and racking up your electric bill. Without a player, an NPC is just numbers and algorithms. Put differently, numbers and algorithms are meaningless unless they produce an effect that is observable by a player. And, therefor, highly detailed NPC simulations that don't produce any different results than simplified simulations would are meaningless.

It's the classic "if a tree falls in a forest and nobody's there, does it make a sound?" question. Sure, the tree falling causes vibrations that travel through matter and which can stimulate the inner ear and trigger a signal the brain… but those vibrations aren't a sound unless they actually get to somebody's ear. Huge sets of data and tons of detailed combat simulations that nobody ever sees do not make an interesting combat at all; it's just a bunch of numbers and math if they aren't being visualized by a person.

Sharing code/calculations between NPCs and PCs is wasteful. NPCs don't need nearly as much data to represent them, simply because nobody actually sees that data. All that matters about an NPC is what it does, not how it does it. Player's don't care if an NPC has 10 strength and 12 dexterity; all the players care about is how hard the NPC hits and how often it dodges. All those minute stats are interesting on PCs because the players get to see them and get to tweak them – they get to min-max equipment or skills to affect a ton of different numbers, all of which filter down into a small number of meaningful values used in actual combat/skill tests. Players don't tweak NPCs. They don't min-max NPC equipment. They don't look at an NPC's character sheet. They don't actually get to _play_ with the NPC numbers. Players just get to fight them, and all the combat system is doing nothing more than turning all those PC stats in a handful (at most) of die roll modifiers… so why does the NPC need anything more than those modifiers at all?

If NPCs are fighting "off screen," why do they need to be burning up CPU and memory for the simulation when NOBODY CAN SEE IT? If a player is in the middle of a huge conflict between 10,000 soldiers on each side, he can only possibly pay attention to the handful of allies and enemies in his near vicinity… so why waste time on the simulation of all those other thousands he can't even remotely pay attention to (or worse, flood his screen with more information than he can possibly take in quickly enough) ?

Most NPCs don't even really need interesting behavior, to be honest. Even long before RPGs and MUDs were invented, the 'mook' was a staple of action fiction. Large numbers of uninteresting, identical, unchallenging enemies that slow the hero down. They aren't there to challenge a hero. They're not there to help the villain escape. They're not there to do anything remotely useful. They're there to establish the above-average skill levels of the hero. They're there to provide a little narrative common sense for the villain (obviously they need people to do all the mundane boring stuff while they do all the interest hero-killing and doomsday-machine-building). They're there to add some fun action before the big dramatic showdown. They're there because they make things more entertaining.

When the players are over-shadowed by the NPCs, the players stop being the heroes. They stop being in the center stage. They stop having fun. Nobody wants to log into a MUD just so they can be the mooks to a bunch of AI-driven heroes. Nobody wants to be a storm trooper when there are jedi and sith running around all over the place. Nobody wants to be the speed bump when the titantic forces come rolling through town. That isn't fun. That isn't entertaining. Outside of some mild interest from psychology majors, you're not going to get much of a player base when you make NPCs equals to the PCs.

You must also remember that MUDs are a text-based medium. In a graphical game, you an actually render 1000 soliders on screen at once, a player can actually take in all that information in a tiny fraction of a second. In written medium, though, you can't describe the actions of each and every soldier. It'd take hundreds of pages. You don't see novelists do that; their books literally say things like, "and then the northern platoon sprang from their covered holes and rushed the enemy flank, sewing chaos amongst the surprised soldiers." 100 ambushers, any number of attacks and parries, many seconds worth of action… all in 21 words. To even begin to give that scene the level of detail that you can do in a graphical medium you would need an entire chapter's worth of words, and you'd end up making the player of your would-be game take an hour to read through 30 seconds worth of action.
20 Dec, 2008, Kayle wrote in the 29th comment:
Votes: 0
elanthis said:
This is intellectual masturbation. This is the kind of question people in academics ask, not the kind of question people who actually want to make a real game ask. :/

You really like the masturbation references huh? :tongue:

elanthis said:
There is a very,<snip snip goes the novel>30 seconds worth of action.

Well said, and I found myself completely agreeing with everything said. Which is unusual…


[Edit:] To add context to the second quote. :P
20 Dec, 2008, Cratylus wrote in the 30th comment:
Votes: 0
lol teabag
20 Dec, 2008, quixadhal wrote in the 31st comment:
Votes: 0
I think the point that's being lost here is that the OP was talking about having large scale battles between groups of game objects. It doesn't matter if there's a player at the helm, or not.

To put it simply… nobody cares about peons. If you have an army of 5,000 infantry charging onto the field, with 2,000 cavalry flanking… nobody cares if all 5,000 infantry have unique names or stats or weapons. Even the players taking part just don't care for that level of detail.

What we DO care about, is the generals. The guys leading the batallions…. the people at the point of each cavalry company. Nobody wants to see 500 messages of the form 'George the rifleman, from 12th company of the South, ***TICKLES*** you with his buck shot'. What I think people would enjoy seeing is something more like 'Lyanic leads the 12th company of the South into a small skirmish, killing 13 soldiers of DavidHaley's 4th company of the North.'

Think about the board wargames you've played… if you field an army of hundreds, you don't control hundreds of miniatures. You control groups. If you're playing an RTS game, you may make dozens of units, but chances are you drag a selection box around them and order them all to do something at once.

Now, you can try to make the argument that players might want to just play the part of some random ground pounder, but how many do you think will really sit there and let the NPC leader control their actions? No, if a player isn't acting as a leader in a large battle, they're probably acting alone, and as such they are not likely to accomplish anything of military value.

From a game mechanics perspective, suppose a lone wolf player tries to assassinate an NPC general. To be fair, that player has to roll well on dozens or hundreds of stealth vs. perception checks by all the NPC's in that general's company. If he somehow goes undetected, to NOT end up fighting the general PLUS half his army, he's have to also win initiative rolls against the dozens or hundreds of soliders who would leap up to defend their leader as soon as he went to make the kill.

In short, he might pull it off, but would probably die (or be captured) in the process. Is that a useful thing? Depends… I guess if it keeps an entire unit off the field, it might be worth it. Is the player going to like the fact that their chance of success is close to NIL? Is it worth modelling? IMHO, no… but YMMV. :)
20 Dec, 2008, Lyanic wrote in the 32nd comment:
Votes: 0
elanthis said:
This is intellectual masturbation. This is the kind of question people in academics ask, not the kind of question people who actually want to make a real game ask. :/


People in academics ask these questions for a very important reason. It's called thinking outside the box. If people in academics didn't do it, then there would be little to no scientific and technological advancement. We wouldn't be having this discussion right now, because this forum wouldn't exist, MUDs wouldn't exist and the internet wouldn't exist. Also, for the record: I am in academics AND I actually make a real game.

elanthis said:
Sharing code/calculations between NPCs and PCs is wasteful. NPCs don't need nearly as much data to represent them, simply because nobody actually sees that data. All that matters about an NPC is what it does, not how it does it. Player's don't care if an NPC has 10 strength and 12 dexterity; all the players care about is how hard the NPC hits and how often it dodges.


Why do you assume the player doesn't care if the NPC has 10 str and 12 dex? That could make the difference between life and death. Or, it could make the difference between learning a new skill or not learning a new skill? Or, the difference between completing a quest and not completing a quest? Etc? Why would players only care how hard an NPC hits and how often it dodges? That assumes that A. NPCs are only good for slaughtering and B. Combat is narrowed in scope to nothing more than hitting and dodging. And again with the argument on being "wasteful" - it's not 1985 anymore. Hard disk and memory space are no longer precious commodities.

elanthis said:
All those minute stats are interesting on PCs because the players get to see them and get to tweak them – they get to min-max equipment or skills to affect a ton of different numbers, all of which filter down into a small number of meaningful values used in actual combat/skill tests. Players don't tweak NPCs.


Says who? My Mage wants to add a +6 str bless to the NPC's existing 10 str, thus allowing it to carry my heavy bags for me?

elanthis said:
They don't min-max NPC equipment.


See above. My Crafter spent many days setting the NPC army's swords and mails with precious gemstones, imbuing magical affects, thus making them near invincible in combat!

elanthis said:
They don't look at an NPC's character sheet.


Once again… Maybe I used a Scroll of Read Aura to determine if the NPC is good breeding stock for a monster farm.

elanthis said:
They don't actually get to _play_ with the NPC numbers.


Why not? Perhaps I have an ability wherein I abandon my normal body and take control of the NPC, playing as it, with all its stats and abilities?

elanthis said:
Players just get to fight them, and all the combat system is doing nothing more than turning all those PC stats in a handful (at most) of die roll modifiers… so why does the NPC need anything more than those modifiers at all?

Most NPCs don't even really need interesting behavior, to be honest. Even long before RPGs and MUDs were invented, the 'mook' was a staple of action fiction. Large numbers of uninteresting, identical, unchallenging enemies that slow the hero down. They aren't there to challenge a hero. They're not there to help the villain escape. They're not there to do anything remotely useful.


This mindset perfectly exemplifies everything I find wrong with not only Diku, but most games (MUD or otherwise) in general. I immediately quit any game where NPCs don't have any interesting behavior. I DO want a challenge. And why are they there if they're not doing anything remotely useful? If they're just cannon fodder that my character is supposed to mindlessly slog through, then I'm going to get bored very quickly.

elanthis said:
When the players are over-shadowed by the NPCs, the players stop being the heroes. They stop being in the center stage. They stop having fun. Nobody wants to log into a MUD just so they can be the mooks to a bunch of AI-driven heroes. Nobody wants to be a storm trooper when there are jedi and sith running around all over the place.


The player also isn't going to feel any sense of accomplishment if he or she starts out as the penultimate hero and every other character around is a mook. Ideally, the player starts as a mook and works his or her way up to become the hero. That necessitates having NPCs of varying degrees on the mook to hero scale. And I can give you an excellent example of why this is so - the MMO, Star Wars: Galaxies. They had people start as traders and such, put in lots of effort, and slowly work up to becoming a Jedi. Sure, there were complaints, but they had a relatively decent number of subscribers. Then, they decided to let everyone start as a Jedi. What happened? Their subscribers dropped off substantially. Why? Because there was no challenge anymore. There was no higher level to strive for. On top of that, they just pissed off the people who put in the hours to become Jedi on the old system (although this is only an issue when the game design is changed mid-stream).

As for the rest of your points on grouping and filtering output from large numbers of NPCs - I agree with you, and have done so all along.
20 Dec, 2008, David Haley wrote in the 33rd comment:
Votes: 0
Lyanic said:
I immediately quit any game where NPCs don't have any interesting behavior.

To be perfectly honest, I think that this is a remarkable straw man statement. Nobody is saying that NPCs should never be interesting. People are saying that PCs are (in most cases, at least) inherently more interesting than anything that AI can provide. Obviously, more interesting NPCs are always a good thing. But the focus of a multiplayer game, in the end of the day, is in giving people interesting ways to interact with the world and each other.

That said, relatively less interesting NPCs do serve a purpose, which is as elanthis said: their reduced competence makes the hero's increased competence stand out. (Again, this doesn't mean that all NPCs are or should be less interesting!)

Anyhow, I think this is wandering far away from the original question, and is rapidly turning into a verbal sentence-by-sentence tennis game. :thinking:
20 Dec, 2008, Lyanic wrote in the 34th comment:
Votes: 0
DavidHaley said:
Lyanic said:
I immediately quit any game where NPCs don't have any interesting behavior.

To be perfectly honest, I think that this is a remarkable straw man statement. Nobody is saying that NPCs should never be interesting.


It's not a straw man argument at all. Perhaps the word 'any' is being misconstrued, but I'm not misrepresenting anything. Elanthis stated that most NPCs are meaningless, don't need interesting behavior, don't need to be challenging and don't need to do anything remotely useful. I disagree. I think EVERY NPC should have interesting behavior and at least be challenging, useful or have meaning. And, I really will immediately quit any game where NPCs don't have interesting behavior. There is no problem in the logic of this.

DavidHaley said:
Anyhow, I think this is wandering far away from the original question, and is rapidly turning into a verbal sentence-by-sentence tennis game. :thinking:


Yes, and for that I apologize. There are some things I just can not let go, though. I won't be responding on this thread anymore. That way it can either go back to discussion of the original question or become inactive.
20 Dec, 2008, David Haley wrote in the 35th comment:
Votes: 0
I see and recognize your point, and actually agree with you in the context of games where there are individual NPCs and that is the granularity you care about. I believe though that the idea Gehrig is going for is to simulate very large battles, where individuals are not as important as the bigger picture. You're not really out to talk to privates 1 through 10,000 and learn their stories and personalities; you want to have a huge battle.

So going back to that original goal, I think the best approach is to group the "foot soldier" mobs together, possibly having groups of groups (battalions). If officers play some kind of role, they should be more visible: maybe killing an officer confuses the rest of the group and reduces its combat effectiveness? (Recall that we can treat the whole group as a single unit.) Players can travel as part of a unit (presumably, as an officer of their own unit) or individually. I'm not sure how much a single player could accomplish against a whole battalion, but it could be very interesting for things like commando missions.

Actually, I think that would be a fairly good way to give players something "special" to do. I'm not sure how fun it would be to be just one soldier trying to land on a beach in Normandy along with hundreds or thousands of other soldiers. It might be more interesting to let players take on surgical strike sorts of missions. It doesn't have to be entirely realistic (the main goal being fun after all) but it's a nice way (IMO) to do something "special" while the larger battle rages on nearby. And the player could occasionally take some role in that much larger battle, for example as an officer giving direction to the various units.

In the end of the day, I think that the takeaway message from all of this discussion (it actually was useful to some extent, IMO :smile:) is that you (you == the game designer) need to establish what players care about in your game – or rather, what your game lets them care about. You then need to represent information in such a way as to maximize their ability to process said information to let them make decisions relating to what they care about.

If you have commando style missions where NPCs are teammates in very small units, it makes a lot of sense to spend time making them somewhat intelligent and interesting. You don't need to do the same for every foot soldier, especially if the only interaction the player cares about is moving them around on what is essentially a strategy map.

Don't simulate anything that does not have visible effects. It's a waste of time, and other than amusement for the person making the system (which is valid enough reason in some cases) it serves no purpose at all.

Especially when trying to build very large simulations involving thousands of agents, you really need to represent the high-level picture, and not try to replicate the world exactly as is.

(Consider how you might go about designing some kind of economic simulation. Do you really need to model each agent in your cities as a separate decision-making entity? Well, probably not…)

(Sorry for the bullet-point-ish post…)
20 Dec, 2008, Igabod wrote in the 36th comment:
Votes: 0
I've gotta say that this thread has been both extremely fun to read and extremely thought-provoking. It has inspired some ideas for me to contemplate putting into my mud that aren't really even related to this subject. Sometimes thread derailing isn't such a bad thing.
20 Dec, 2008, David Haley wrote in the 37th comment:
Votes: 0
Thread derailing is the nature of forums, and IMO at once their most valuable and most annoying feature. :smile:
21 Dec, 2008, elanthis wrote in the 38th comment:
Votes: 0
Lyanic said:
This mindset perfectly exemplifies everything I find wrong with not only Diku, but most games (MUD or otherwise) in general. I immediately quit any game where NPCs don't have any interesting behavior. I DO want a challenge.


Never did I say that there is no challenge. Only that individual mooks are not themselves challenging. The challenge in those kinds of battles has nothing to do with individually interesting NPCs, but in having very interesting group tactics. A one-on-one-hundred battle kind of scenario.

And then of course you have the (relatively) uncommon important NPCs that actually do have uniquely interesting behavior.

Quote
The player also isn't going to feel any sense of accomplishment if he or she starts out as the penultimate hero and every other character around is a mook. Ideally, the player starts as a mook and works his or her way up to become the hero. That necessitates having NPCs of varying degrees on the mook to hero scale.


Obviously. Never was anything else stated. I am in FULL agreement with the "when everyone is super, nobody will be" line of thinking.

I also realize that if a PC is always playing second fiddle to a freaking AI, people aren't going to be interested. People want to work up the ladder, but they usually want to compete for first prize against other players, not NPCs. If they wanted to see if they can beat NPCs, they'd be playing Oblivion or something, not a multi-user dungeon.
21 Dec, 2008, Cratylus wrote in the 39th comment:
Votes: 0
Quote
Sharing code/calculations between NPCs and PCs is wasteful. NPCs don't need nearly as much data to represent them, simply because nobody actually sees that data. All that matters about an NPC is what it does, not how it does it. Player's don't care if an NPC has 10 strength and 12 dexterity; all the players care about is how hard the NPC hits and how often it dodges.



I agree with elanthis that simplifying groups and
NPC's to only-what-is-necessary is good in terms
of efficiency. It's kind of self-evident, really,
I'm not sure I need to get into why it is so.

I don't find that argument compelling, however. In
part because efficiency is not my top priority.

My priority when I'm messing with code involves
maximizing my amusement and maximizing the opportunity
for a learning event. I believe the technical term
is "intellectual masturbation," and I'm here to defend it.


Defense 1: Amusement
——————–
I like the idea of an NPC with all the features
and characteristics of a body that could be occupied
by a PC. It's not terribly important to me that
each NPC be shaved by Occam into only its absolute
essentials. The existence of those unnecessary
pluralities provides me with amusement.

Do you remember J.F. Sebastian in Blade Runner? I
am lucky that I do not suffer from his disease or
his loneliness, but I strongly sympathized with his
hobby. He used his skills to make toys for himself,
that sang or talked or walked around, for his own
amusement. I can imagine the tenderness he might feel
as he constructed a marching soldier, adding to it
touches of verisimilitude that served no purpose of
efficiency, but were not meaningless for being superfluous.

It is my preference for NPC's to inherit the same
base classes/functions/etc as the bodies inhabited
by PC's because it pleases me to imagine the virtual
world in this manner.


Defense 2: Simplicity
———————

"Simplicity?" you ask, "Doesn't that directly contradict
the argument you just made for amusement-over-efficiency?"

Somewhat, but hear me out. Consider the efficiency
argument that asserts that a player does not care about
an opposing NPC's skill points/stat levels/etc….and
that such numbers are not needed in the NPC. According
to this view, only the simplest numbers that calculate
combat results are really needed.

I believe this is somewhat sound if the player is never
going to confront another player in combat. However, it
is my opinion that sometimes it is appropriate to have
PvP. In such cases, it would be important to me that the
logic of "to hit" and "damage taken" be the same whether
I am fighting an NPC or PC orc warrior. Since my preference
would be for the rules of combat to be the same, using
a separate set of code to exactly duplicate the combat
mechanics of PvP would be reinventing the wheel.

It would mean that in order to suit my preferences, I
would have to come up with a system whose predictability
is indistinguishable from that of the system already
in place, but is somehow *simpler*. Why not just
simplify the existing combat as much as possible, and
just use that throughout…rather than two separate-
systems-that-might-as-well-be-the-same?

As an example, let's take an npc, make two
instances of him in a room, one with platemail and
a longsword, one with leather armor and a dagger.
In my codebase, this equipment difference causes
all sorts of important distinctions in the given npc's
survival in combat, dependent on what the attacker's
weapons and skills are. To replicate this avoiding the
standard combat/body code each soldier inherits,
I'd have to come up with a really complex system
whose only merit, if such it be, is that it isn't the
system already available.

It seems like unnecessary extra effort, in my case,
given my preferences.


Defense 3: Serendipity
———————-

A book I loved as a kid was I, Robot. It introduced me to
the idea that simple rules applied to complex systems can
have outcomes so unexpected as to amuse and educate.

Something like this is another reason I am attached to
the idea of NPC and PC code being similar. Wandering
NPC's have taught me a lot about how the virtual world
affects them, and how they affect it, and because their
code is so similar to players this has allowed me to
improve how things work in general.

It may be that simplifying automated simulations to their
most basic possible variables is more *efficient*. I
would suggest, though, that running simulations using
variables as close as possible to the production state has
the potential to be more valuable, due to the simple
fact that we're humans, and we don't think of everything,
and sometimes the most exciting discoveries are heralded
with the surprised comment "Hmmm….that's funny…"

I suppose there are folks who have achieved a status
so godly as to be beyond surprise when coding. Count me
among the happy newbies then, who still have adventure
and discovery as their company in their hobby.

I'm a fan of efficiency when it serves my interests. I
am not interested in serving as an instrument of efficiency.
I'd be inclined to start simplifying groups when
my platform/performance made it necessary, and not as
a general guiding principle.

Unless, of course, I was getting paid for it. I am doing
this as a hobby, you see.

In terms of large scale "armies" type stuff, I think that
simplification is pretty much *necessary*, since most muds
don't run on machines in Sandia National Labs. However, if
it weren't necessary, I would not advocate it for every case.
Some people just don't want it. Big deal.

-Crat
http://lpmuds.net
21 Dec, 2008, Lyanic wrote in the 40th comment:
Votes: 0
Cratylus said:
A lot of stuff…


I know I said I wasn't going to be posting on this thread anymore, but BRAVO! I completely agree with every word of your post, Cratylus. It's like you're in my head. You even argued some of the points that I wanted to, but couldn't find the words for earlier.

P.S. The thread derailment continues! *hides*
20.0/44