Worlds VS Games, etc {was GMuds, UO}

[Nathan Yospe]

On Tue, 29 Jul 1997, Koster, Raph wrote:

:[An aside; since we do client-side prediction for a lot of stuff, we 
:can "swallow" latency up to around 500ms without visible effect except 
:when the prediction screws up. For those who have played, the lags 
:you're seeing are from other causes, mainly server-side load. Now back 
:to the topic...]

I'm curious about this... I'm working on a _secure_ client negotiation on
login, that will confirm or fail to confirm one of several client types
(using keys for the clientside safe confirm) but all it really comes down
to is, how much information filtering capability does the client get? No
other data is available. (In case you were wondering, the secure clients
have a copy of the text database, edited on login to bring it up to data,
and do all of their text conversion filtering on the client end... this
makes way for graphical clients at some point in the future, as the client
can also negotiate for full vectorized positioning and update algorithm
access. Hopefully, a final version can someday be adapted to VR, once I
solve the problems with distribution.) Nevertheless, I hesitate to
actually add this type of thing to the game... and it doesn't even handle
reflexes, which are server end.

:However, we have limb-based calculations, equipment damage, 
:health-fatigue relationships figured into speed, variable speed 
:attacks based on skill and weapons, weapon weight, range, etc etc. All 
:in all, it's about as sophisticated a system as most muds use. It's 
:just not really visible to you. You can feel the effects of it all 
:with experience, of course, and "grow" into the system through 
:experience. Can you lop off an arm in the midst of a fight? Nah, I 
:don't like limb-based health, it makes fights too quick and 
:contributes to "trophying" which I find repellant. :)

How coordinate flexible is it? Degrees of freedom? I do use fully limb
based combat, to possible excess. My theme is sci-fi, the motto of combat
is "Duck and cover! And I mean COVER!". The result of this is a majority
of players ending up with regenerated or prosthetic/cybernetic limbs. I do
hide it from the players, of course, except for the text flow across their
screen... which is all events, no numbers. As for trophies, why not
discourage it in-game somehow... make it an integral taboo in the culture?

:Hmm, don't mean to get up on a hobbyhorse of defending the system. I 
:too wish that we had more interactivity in it, and more ways to see 
:feedback, information like "Grogan's blow hits your breastplate and 
:dents it deeply... you have trouble breathing now that your chest is 
:constricted, and a haze falls across your eyes as you are obliged to 
:gasp quickly and shallowly (increased fatigue from here forward, 
:penalty to hit, possible subsidiary damage to chest if fight goes on 
:too long)." But interface-wise, ugh. Not in a graphical system, not 
:easily, and not for the general public. :(

Dunno about "not for the general public"... I sincerely HOPE they can get
used to it... cause I'm doing it. As for graphics support... I guess its
gonna need to take the form of a little character icon in the corner,
showing your current condition/behavior until we get VR that includes some
sort of tactile feedback. Taking that breastplate off might be a good
idea. (The other alternative is a red message flashing over their screen,
"Your chest hurts! You can't breath! Your breastplate is crushing your
chest!")

:UO has an economic system that goes from raw materials to finished 
:goods, and players can make any step of it.

:UO also has an ecological system that handles creature repopulation, 
:behaviors, etc.

:Both of these are intrinsic to the game, far more so than combat or 
:spells or guilds. They are the simulation layer under that, and 
:despite repeated calls for "mud evolution" not many take up the 
:gauntlet to work on this sort of thing further. Yet I remain convinced 
:that adding a simulation layer to muds is where the genre needs to 
:go.

Most of this list seems to feel the same, myself included. (To a degree -
I use utter destruction and fallow, to be recreated later under a
different name, after a month or so real time, as an alternative to this
in the case of intricate "story/puzzle/planned" elements. On the other
hand, my engine makes 90% or more of my world up as it goes along, off of
general templates, and evolves this based on local events, but with a
strong effort to renormalize - my ecosystems have artificial equilibrium
points.)

:Now, I know the arguments against it; it's expensive (man hours, 
:money, computation, space, etc). You can trick the player into 
:thinking it is there when it really isn't, far more cheaply. (And in 
:fact, we have used "cheats" many places where portions of the model 
:were deemed less important to actually simulate, but we needed the 
:appearance). But having a solid sim layer enables so much... and it 
:renders future growth possible. For one thing, the next direction 
:which I would like to take it is towards conquering that last barrier 
:of "staticness"--changing the setting based on simulated environmental 
:factors. Given a good model, there is no reason why roads could not be 
:formed by players as they walk on the grass repeatedly and kill it. 
:And so on.

I've never bothered with economies, finding them far less important, and
far less interesting, than environments. On the other hand, this is
probably why I set my game in the middle of a warzone, and the world as an
extended two galaxy backdrop for said war. The war itself is also set
toward equilibrium. You start out losing, you progress, gain ground, but
the good guys, so to speak, are always losing. And if you are not actively
a combatant, you are threatened by a warfront that seems always to be
getting closer, but never quite reaches you. Yes, it is dynamic, but
everything is designed to simulate equilibrium. Which is, IMO, the only
way to get a static (desirable from a maintainance perspective) world that
is always different (desirable from a game design perspective.) The real
trick is having a file of "protoworlds" that can be added to the galaxy
subtly when balance is disturbed, or destroyed in a massive invasion, or
reconquered (each proto comes with its own renormalization program), which
allows PC led armies to _attempt_ to defend it (and possibly succeed, if
that was allowed for, otherwise the enemy turns out to be limitless in
number)... quite a lot of fun. The "Fall back to the transport station,
and prepare the wormhole for destruction once we leave!" reaction is quite
good... allows for the pack of cut and bleeding soldiers, returning to
base with drawn and depressed faces... just wish there were a way to pull
off the full effect. (At the moment, that sort of event actually does set
default mood to depressed, or defeated, depending on character, and this
does get relayed.... changing it requires explicit effort, eg '>cheer up',
"You attempt to smile, but the effort is too great.", '>sing redemption',
"You start singing the battle hymn, and hear the voice of seargent Kingsly
join in behind you. Several of the troopers begin singing along. You feel
yourself walking more quickly." - this type of thing, and the filtering
that goes on to make it possible, are the reason for my move to client end
filtering.)

:We use a classless, levelless system. Skills atrophy from disuse. It's 
:not to everyone's taste. It seems to be working, so far. Whether it 
:will retain the totally goal-oriented players is yet to be seen, since 
:it does not provide a long string of milestones, but rather obliges 
:the player to create milestones for themselves at intervals.

Sounds quite a bit like my own system. Are related skills codependant?

:Not sure what this last sentence means. Does it mean that NWN and DSO 
:are focused on being games?

:As far as it being a good mud... well, it's I suspect, up to par with 
:run-of-the-mill muds in most ways. In other ways it's a heck of a lot 
:more ambitious. Bt then, doing ambitious stuff is why I came here 
:where the funding is. ;) In other traditional areas of "mud 
:measurement" it's gonna fall way way short. A discussion of what 
:exactly the metrics are for something like this would be quite 
:interesting. How do we evaluate and judge muds?

Hmm. Muds go in so many directions, and each one has its own goals... My
system is far less programmable than MOO or Cold, has to be modified in
the actual executable to change anything beyond world-feature type
programming (which uses an internal language), and has no formal database,
aside from the lexical and prototype-object databases, which are both
heirarchal, and the world object listing, which is entirely geographical
and decentralized, and has no method for retrieval aside from those
invoked at the object to be retrieved (which requires a method of
identification in the first place... either a watcher object or a line of
sight ID, so to speak...) By the criteria of many mud bases, Physmud++ is
an utter failure. But I don't know of many muds that boast weighted
partial neural nets for skill balancing, or for environmental alteration,
or a (decent) language parser, much less an output language parser, or the
ability to model physical interaction with what amounts to open order
formal evaluation (if that much detail is requested, a lagrangian (not
laplacian, which is what it seems that addledbrains fella thought I said)
evaluation will take place. In most cases, a transform approximation
suffices.) I think I have so far done fairly well by my own criteria, and
when I finish, have every confidence that _I_ will be pleased with the
results.

:UO is certainly focused on being a world first, and a game second. The 
:social aspects also fall secondary to this. Hence the lack of 
:easy-to-implement, obvious social enhancers such as long-distance 
:communication, embedded mail system, and global chat spaces. All of 
:these things are major social enhancers, but (usually) outside the 
:fiction and reductive of a game OR world experience.

One nice thing about sci-fi is that I have something to replace these in
game context... the other measure of retribution I took, however, was to
create a set of chatrooms that players (not their characters in the game
world) could frequent, sort of a socialization room. The one I expect to
be most popular is modelled as "BackStage", complete with "Dressing Rooms"
and "Makeup" (Dressing Rooms are character creation, makeup is character
_persona_ creation, where a history (past) and personality (which is used
for subconcious mannerisms - a bit of color, but sometimes actually
relevant, in cases of being incited to anger, for example.)

:One reason btw why we went with this approach was that a focus on 
:world tends to capture the "explorer" types as Bartle defines them, or 
:in Bettelheim's terms, encourages open-ended play. Or to put it in 
:other words, having a varied, evolving setting (even though it only 
:evolves in that "middle layer" of NPCs/creatures/economy) encourages 
:roleplay, encourages exploration, encourages alternate styles of 
:achievement, and rewards it with changed circumstances rather than 
:with a milestone.

But it also set you up for disaster if you don't have either a way to
retain euilibrium, or to reset conditions. Say, for example, half the game
was burned down in riots. How long does it take the game to recover? Or is
there a mechanism for recovery? Or is there a "reset" switch to undo the
damage? Or is it impossible to burn half the game down?

:The problem with "game" style design in a mud setting is that you run 
:out of game. Games are finite. In a fiscal sense, you wanna keep folks 
:around as long as possible, of course, to get their money, and the 
:more "infinite" the game is, the better. Remember that most mudders 
:only play for around 3-6 months, and even dinos tend to give up after 
:2 years or so.

Would having an infinite world alleviate this? I'm not entirely sure.

:It is difficult for a player of any <game> (using game in a broader 
:sense now, as in game design, as opposed to "game"-style <game> 
:design, boy I hope that made sense) to make the transition between 
:methods of approaching the game. For one thing, not many games have 
:the flexibility to be played in truly different ways. One of the 
:reasons why Sid Meier is a master game designer is that he has a knack 
:for open-ended play that has milestones that can be freely ignored. 
:Yet it is rare to see a Civ player who plays once for conquest and 
:again for cooperation and again for mastery of a particular area and 
:again for social stability etc etc etc... the game design supports it, 
:the *individual player* does not. But the *audience* does.

Ah... the difference is quite significant, though. If you have two
conquest oriented players, and ten stability players, you have a fierce
war between them. Ten to one to one, and no chance for any of them to win.
(Ten outpower either one, but the ten stand no chance of achieving their
own goals.) Which is not particularly a _bad_ thing, mind you.

:> > The goal of 'building a
:> > world/environment' is something that appears to have evolved 
:amongst the
:> > readership here (not that its a bad thing by any means! Its a 
:huge
:> > conceptual step forwards, IMHO).

:It's also a really old conceptual leap forward. :) I'm glad to see 
:that this list is embracing it, but I remember arguing with Orion 
:Henry and Mike Sellers (who did Meridian 59 and is now freelance I 
:think) and others about it on the newsgroups a LONG time ago. Like, 
:over two years. "What is the MUD State of the Art?" I think is the 
:question I posed back then... it is STILL a valid question to ask, 
:because I suspect that if the mud community gathers as it has on this 
:list, and actually manages to make all the disparate great ideas come 
:together, we'll see some mud evolution.

Some of the great ideas are mutually incompatible. Miro and I have had
some interesting discussions of the merits of superstructured vs
unstructured world simulations... both extremes are closer in effect to
each other than to the semi-structured (Diku & LP) median. Things like a
dinner with a guest being slipped a love philter. You either have _food_,
eating, dishes, and disolution all supported, and a filtering system to
hide/show the action, which is what I do, or you do it all by emotes and
private emotes (Miro's method), but you cannot do it with minimal support
like a Diku or LP... The love potion itself has to have realizable effects
(this is programmable in my internal language, and since it is derived
from aqueous fluid, the coding is minimal) or the victim has to be a good
sport and play along. (Love potion code, at least the version I coded to
follow this, results in an inability to focus on anything else when the
subject of the potion is present in the room, and constant "thoughts"
about the subject, slipped in whenever a trigger (think eliza) pops up.
There are also penalties to the victims ability to carry out action
against the subject.)

:That vision is exactly what I wanted to bring to UO. And lemme tell 
:ya, on a personal note, it's great for me to log into this game and 
:try to go make a living as a tailor who wants to be a bard, have the 
:character respected and in demand for the character's skills 
:(everybody wants to look special, so everyone wants custom dyed 
:clothes), be frustrated because there's a shortage of dyes in town, 
:ponder getting backing to bring a trade caravan into Trinsic to see if 
:I can make a killing on dye pots, and go kill a bear in the woods that 
:I KNOW won't be there tomorrow. There's something oddly liberating 
:about how different it feels to take for granted sim-based design 
:rather than static environments. How many of you are working on this 
:sort of thing in a text environment, where it could be pushed so much 
:further than in graphics? (The possibilities boggle the mind there)... 
:I'm curious, because I'd love to see what designs you come up with.

I'm working on it in an environment, nominally text at the moment, but in
a lot of ways, I am finding text more difficult than graphics for things
like decay - remember, while similar tiles are acceptable in a tile
graphic environ, and polygon graphics are crude enough to get away with a
lot these days, text implies a demand for literary style. Balancing the
literary considerations of design with the dynamic flexibility I desire
has been a terrible chore. I'm cracking it, slowly, getting the kinds of
descriptions I have posted before, but the markup language and information
filters have grown and grown, and they are sucking up cycles like soda.
That's the main reason for going with a client, and the truth is, the
stuff the client recieves could as easilly translate to graphics as text.
I know. I've actually done a few rooms as graphics... 3D wireframe
graphics with simple textured color wrapped around them.

:A thought:

:Game design is an art and a craft, and like all arts and crafts, it 
:has techniques and approaches, and that implies that it can support a 
:criticism; said criticism exists though it is not very sophisticated. 
:Mud design is also an art and a craft, and it also has techniques and 
:approaches, but there is no criticism, no self-evaluation, no 
:standards defined, no study of what has gone before. And without 
:self-critique, it cannot improve except in fits and starts. If this 
:genre is to evolve into more than game design, which I firmly believe 
:it has already begun to do, then it will have to support at least the 
:critical apparatus of game design, and preferably the critical 
:apparatus of many disciplines that most people do not bother to link: 
:server design, and writing, and hypertextual theory, and art (for 
:graphics are coming *and will dominate*, it's not worth fighting 
:over), and psychology and sociology... Game designers today generally 
:do not know even the short history of computer game design; we must as 
:a community educate ourselves and each other if we want the community 
:and its art and craft to grow.

I agree fully with this. It seems to me that this list has contributed far
more to that end than any of the newsgroups, or the JoMR, or the various
pages in existance. Partially, I think, because of Chris L's knowledge of
early muds, and because of his constant stirring, partly because we
encompass so much of the innovation energy of the field today, and are
unafraid to criticize each otehr, or to scrap our own designs, or to
borrow and share ideas, in spite of radical differences in implementation.
I think my rather unusual method of simulating physical systems, which
encompasses a working chemistry and both gaseous and fluid dynamics as
well as solids ranging from moldables (metals included) to crystals to
composite organics and ceramics, rubbers and foams and so forth... all
started with a discussion between J C and myself (with input from C G and
others) about limb based combat on the previous incarnation of the list.


"You? We can't take you," said the Dean, glaring at the Librarian.
"You don't know a thing about guerilla warfare." - Reaper Man,
Nathan F. Yospe  Registered Looney                   by Terry Pratchett
yospe at hawaii.edu   http://www2.hawaii.edu/~yospe           Meow