Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[jamoecw]

i was thinking a day was in game day, not out of game day.  stone takes forever to haul, my legendary miners need at least 3 or 4 haulers each.  i think we are in agreement that it would be nice so long as one can repair the damage, which means engraving artificial walls and a way of dropping walls (or anything that acts like a wall like a flood gate) into water from one open z-level up.

[GoldenH]

Well I consider the current system for Constructions to be wholly inadequate. Filling in isn't just needed to make rooms closed, it is needed so that I can 'blank' an area and then just designate to Dig again whenever I feel like it... instead of having to deconstruct everything which is a huge hassle.

But yeah an in-game day is pretty short, I think it takes longer even to hollow out a new dining room, but it doesn't take a real time day to clear a whole Z level, about an hour at 30 FPS is right. It takes longer to bust through an aquafier (I busted through a two level one for the first time last night, yay) The reason I do things like this is because i hate the Z-level pathfinding algorithms so I like to dig down and find an empty layer for me to play in.

It helps that I don't bother to haul stone. I will make a dump zone in an area to clear stockpiles, but that's it.

[G-Flex]

Deconstruction needn't be a bigger hassle than digging. It works exactly the same way. The only difference is that it takes way too long and doesn't require a specific job, both of which can be changed easily enough.

[bloodnok]

I am posting, newly registered, as someone who is fascinated by the idea of DF but finds that about four hours of play leads to keyboard-snapping frustration. What I find most frustrating about DF is the way there's clearly a good game in there trying to get out.

[FWIW, I play NetHack, Crawl, POWDER, etc. I don't care about graphics, and a modicum of interface perversity is par for the course.]

As I see it, a few things get in the way.

First is the sheer mass of micromanagement. Someone upthread mentioned MOO1, which to me (although it gets it wrong in some places) is, in places, the perfect example of how a 4X game should work. Let the player control policy, not implementation. Job priorities are a hardy perennial wishlist item; another one would be using the job manager to define desired stocks of certain items. Don't make me order batches of doors; let me say "maintain a stock of twenty doors with priority n" and never think about door manufacture again.

[FAOD, this isn't my sole micromanagement gripe. The game is absolutely riddled with places the player ends up doing implementation repeatedly when they should just set a policy and only come back to it when the policy needs changed. A good look at the way the Settlers 2 economy gets on with the job without human intervention wouldn't hurt.]

The second thing on my list is rampant simulationism. This may seem a strange gripe, rather like complaining about water's rampant wetness, but I'm going to run with it. My first subgripe is the diversion of development effort into pointless detail. For example, from the actual point of view of fortress management, a dwarf might be in one of at most six states of health; healthy, with a temporary injury and fighting on, with a temporary injury and seeking or receiving medical treatment, permanently crippled but still able to work, permanently crippled and a dependent of the fortress, or dead. So why model thirteen kinds of suture in exposed guts? The history of the world is modelled, but when a bronze colossus turns up, who cares if it killed Arnar Bibbleface in 524 with a slightly rusty steel mace? Not I - it presents the same challenge either way.

My second subgripe is cognitive dissonance from inconsistent simulationism. There are umpteen kinds of plant each of which brews its own booze with the specific characteristics of each booze modelled... but brewing does not require water. There's a complete model of hydrostatic pressure, but tap into a river below a waterfall and comedy ensues. The lovingly detailed "strange mood" subgame can produce an artifact pigtail sock, but no-one has reflected on the quantity in which socks are ordinarily produced and used.

Third is simulationism versus gameplay. DF has done surprisingly - incredibly - well with gameplay through emergent properties, but gameplay does seem to be losing out. There's a reason there's a hardcore that still plays the 2D version, where the river/chasm/magma gameplay was explicitly scripted in.

[Anu Necunoscut]

I am posting, newly registered, as someone who is fascinated by the idea of DF but finds that about four hours of play leads to keyboard-snapping frustration. What I find most frustrating about DF is the way there's clearly a good game in there trying to get out.

[FWIW, I play NetHack, Crawl, POWDER, etc. I don't care about graphics, and a modicum of interface perversity is par for the course.]

As I see it, a few things get in the way.

Welcome to the forum.   I differ from you a bit on a few of these criticisms and agree on some others, but it's all down to taste, so take the following with a huge grain of salt:

First is the sheer mass of micromanagement. Someone upthread mentioned MOO1, which to me (although it gets it wrong in some places) is, in places, the perfect example of how a 4X game should work. Let the player control policy, not implementation. Job priorities are a hardy perennial wishlist item; another one would be using the job manager to define desired stocks of certain items. Don't make me order batches of doors; let me say "maintain a stock of twenty doors with priority n" and never think about door manufacture again.

It's a fine line, though, no?  And the two are obviously different sorts of games--DF is inherently more personal in scope than MOO1, where you throw hundreds of millions of people around on any given game turn.  Keeping that in mind, many standard 4x features just wouldn't feel right for the game.  Cutting down micromanagement for mature forts, however, would be great--"Maintain x stock of y" would be a great feature.  The game shouldn't too easily "play itself," however, and happily many mechanics are in place to upset your policy decisions with individual implementation problems, like unhappy dwarves, the necessity of having an experienced bookkeeper/manager, etc.

The second thing on my list is rampant simulationism. This may seem a strange gripe, rather like complaining about water's rampant wetness, but I'm going to run with it. My first subgripe is the diversion of development effort into pointless detail. For example, from the actual point of view of fortress management, a dwarf might be in one of at most six states of health; healthy, with a temporary injury and fighting on, with a temporary injury and seeking or receiving medical treatment, permanently crippled but still able to work, permanently crippled and a dependent of the fortress, or dead. So why model thirteen kinds of suture in exposed guts? The history of the world is modelled, but when a bronze colossus turns up, who cares if it killed Arnar Bibbleface in 524 with a slightly rusty steel mace? Not I - it presents the same challenge either way.

To a point that's true, but the same style of argument could be used to justify removing a lot of great flavor from the game.  The hilariously violent current combat system could be replaced with hitpoints, groups of similarly employed dwarves with abstracted "units," free bedroom/workshop/entrance design with a few templates, etc.  They wouldn't be that functionally different, but something valuable (to me anyway) would be lost.  For surprising and delighting the player, amputations and unexpected weapon-monster interactions, the fluid mechanics, etc., seem to me very useful.  A lot more of DF's "flavor" should eventually go to improving replayability and having an impact on how you play a unique fort/site/world, which all currently play much the same.  To use MOO1 as an example, the random range of a race's personality in that game actually determines how you should respond to them diplomatically rather than being only window dressing.  Ideally that's what should happen in DF.

My second subgripe is cognitive dissonance from inconsistent simulationism. There are umpteen kinds of plant each of which brews its own booze with the specific characteristics of each booze modelled... but brewing does not require water. There's a complete model of hydrostatic pressure, but tap into a river below a waterfall and comedy ensues. The lovingly detailed "strange mood" subgame can produce an artifact pigtail sock, but no-one has reflected on the quantity in which socks are ordinarily produced and used.

This comes down, I think, to features being added as Toady feels like adding them, and them being balanced as necessary/feasible.  While complexity without overall consistency in sims can seem a bit wacky, you don't want the player to be forced into learning arcane real-world mechanics for mundane aspects of the game.  Imagine having to micromanage forge/smelter temperatures precisely based on what sort of metal you want worked.  If sim complexity adds to variety and potential for player creativity, it's good.  If it provides only tedium, it's bad.  Still, it's a hard balance to strike.

Third is simulationism versus gameplay. DF has done surprisingly - incredibly - well with gameplay through emergent properties, but gameplay does seem to be losing out. There's a reason there's a hardcore that still plays the 2D version, where the river/chasm/magma gameplay was explicitly scripted in.

The easy answer to this is that replacing scripted antagonists with procedurally generated antagonists is difficult and can't be effected easily all at once.  A lot of the challenge in 2D comes down to bugs (+pig tail sock+ chasing, malevolent pachyderms) and a lack of effective player control.  There's also the fact that scripted challenges tend to be exciting only once or twice, before a formula is learned that forever beats them dead.  Once that happens, they become simply tedious.  A lot of 2D's value, I think, is that you are guaranteed interesting stuff in your site, and each new breached feature adds some new potential antagonists.  I think 3D could potentially reclaim and surpass a lot of what makes 2D fun.

[zwei]

What do you think is scaring people away?  The building placement?  Designations?  The embark screen?  Or maybe its finding the right tile sets and setting them up.  We are hoping at some point to build easier commands and tutorials to help bring in more players.  We have to identify the main culprits first.  So what is frustrating you the most about Dwarf Fortress?

Lack of tutorial.

Lack of common interface/'verb' to do stuff.

I.E. There are quite few command trees to do some tasks. For example, in order to build something, I can 'd'esignate area, i can 'b' uild something, i can make stock'p'ile, and there is some stuff like 'i' zones, i can 'q'uerry existing object to 'build' room.

Digging stairs and constructing stairs uses different keys for each stair type.

Some workshops working automatically (butcher) with no aparent way to make use of them (A-HA, i can set animal to be slaughtered and someone will take care of it!)

That kind of stuff. It really takes some serious research and wiki in order to figure out how to do, well, anything interesting.

I just wish there was 'w'iki or 'h'elp command in all menus, object queries and such which would launch browser pointing to relevant page on some wiki.

[GoldenH]

Even something simple like "you can eat this" or "Used at X workshop" in the description is helpful. You don't need to link to the wiki, if you're going to use the wiki it's easier to have it open in it's own window.

"Okay, I've got the caravan open... i see fisher berry wine, I can drink that... gnomeblight? Uhm, guess not... What, what's this Sunshine stuff? It sounds like it would be inedible like gnomeblight, but oh look my dwarves are drinking it, neat!"

the same with the large/small/narrow etc. "Dwarves have no use for this' in all the descriptions of things you can't use would be nice.

There is a lot about dwarf fortress that automation would not reduce choices but would reduce babysitting. For instance you could decide to always have 20 Doors for instance but you probably don't want x amount of everything lying around, it would waste a huge amount of stockpile space and you'd have to manage it based on your prospective immigration opportunities. If dwarves had decent z-level pathing (or even if it just counted 1 z level = 10 spaces or if it added x+y to find the distance instead of  finding the distance between two points. A toggle would be nice. Both these would take about ten seconds to implement) then they wouldn't insist on running down ten flights of z levels when they have stones literally right next to them and then you wouldn't have to be careful do forbid anything below them. etc. If you could copy room Designations then you'd still have to change the rooms when you want to but at least you wouldn't have to get a repetitive stress injury first. If you could designate diagonally you wouldn't have to build a million things. And why build a bridge when constructing floor tiles works so much better? (unless you want a drawbridge of course) And so on.

[bloodnok]

It's a fine line, though, no? And the two are obviously different sorts of games--DF is inherently more personal in scope than MOO1, where you throw hundreds of millions of people around on any given game turn.  Keeping that in mind, many standard 4x features just wouldn't feel right for the game.

I think you misunderstand me slightly. I don't want DF to be like MOO1 and I certainly don't want it to be like a normal 4X: I want to look at the way MOO1 is not like normal 4X games. Instead of this nonsense of ordering individual buildings built on colonies of millions of people - and having no real idea what the effect of changing the build order is - you say "build factories" and they go and build factories. Policy, not implementation.

Cutting down micromanagement for mature forts, however, would be great--"Maintain x stock of y" would be a great feature.  The game shouldn't too easily "play itself," however

I think the fear of "playing itself" is misplaced. I don't want the game to make decisions for me (albeit that some sensible defaults might help the process along a bit - for example, a job priority system would probably default to doing Clean Fish before Fishing), but I do want to automate away anything I can do which is repetitive and requires no thought. For example (once I can check fortress stock at all) I could maintain x stock of y by checking the fortress stock every fifteen minutes or by manually tracking whenever I use a y, and ordering new ones built immediately. Boring, repetitive, requires no skill on my part - automate it. But I don't expect to be able to say "build me a farm" (that said, a macro language for standard constructions wouldn't hurt, but that really is wishlist stuff).

To a point that's true, but the same style of argument could be used to justify removing a lot of great flavor from the game.  The hilariously violent current combat system could be replaced with hitpoints,

Well, quite. As I mention, dwarfs only have about half a dozen distinct states. A dwarf is healthy, hurt, or dead; if hurt, they will or will not make an eventual recovery. Now the modelling of injuries in hideous detail is amusing once, and thereafter it contributes basically nothing, and represents a vast diversion of development effort. Furthermore the current system is very hard to balance compared to a simpler one - to write a creature that will, say, provide a reasonable challenge to well-trained dwarves is tricky at best without copying an existing one. Look at the way totally random creatures have become King of the Beasts. Carp?

This comes down, I think, to features being added as Toady feels like adding them

Well, quite, rather than being added where they are most needed to paper over known gaps in the game. (Blah, blah, yes, it's his game, he can work on what he likes. But the question was what _I_ don't like). For how many years, now, have dwarfs been incapable of responding sensibly to being on fire, and fire itself a slow-burning but irresistible contagion which isn't actually very like fire?

While complexity without overall consistency in sims can seem a bit wacky, you don't want the player to be forced into learning arcane real-world mechanics for mundane aspects of the game.  Imagine having to micromanage forge/smelter temperatures precisely based on what sort of metal you want worked. If sim complexity adds to variety and potential for player creativity, it's good.  If it provides only tedium, it's bad.  Still, it's a hard balance to strike.

But I don't think there is any effort to strike a balance there at all; sim complexity is added without any real consideration of the gameplay effects.

The easy answer to this is that replacing scripted antagonists with procedurally generated antagonists is difficult and can't be effected easily all at once.

But, again, is there ever going to be a real effort to ensure the procedurally generated antagonists are equally varied and interesting? Procedural content can go too far. Consider a roguelike; the individual levels and monsters therein are procedurally generated, even the depth and branching of the dungeon may vary, but you never turn up to find the dungeon consists of two rooms, three asthmatic goblins, one cheese ration, and the Amulet of Yendor.

The current - and even the projected future - situation where interesting terrain features are scattered all over the world and a given site has only a small proportion of them strikes me as analogous to that. So what's the answer? Use Site Finder to find the spot in the world with them all. Why not just _generate_ a spot with them all and have done with it? The terrain in the rest of the world is basically irrelevant either way.

Sure, eventually there might be _so many_ fascinating terrain features that any site will have a good selection and a site with them all would be obviously mad. That strikes me as being a very long way off, though, and right now a reversion to the 2D model where you just get a site with all the good stuff wouldn't hurt.

[Granite26]

I find that considering Fortress mode to be a playable stub of a bigger game helps with a lot of those conceptual issues.

Edit:

Or, to belabor the MoO analogy, adventure mode is the space cobat sim and fortress mode is managing a single planet.

To me, complaints that some planets are rich or poor or radiated, and you've got to pick the perfect one doesn't make any sense, nor does worrying about the excessive modeling of the fleets and space monsters that show up at your door.

[bloodnok]

Seeing DF as a stub of a bigger game is a bit "jam tomorrow". I've got to play the game that's available now, and knowing that these issues might be fixed before either Toady or I die of old age is scant comfort.

To me, complaints that some planets are rich or poor or radiated, and you've got to pick the perfect one doesn't make any sense, nor does worrying about the excessive modeling of the fleets and space monsters that show up at your door.

In this analogy, that's not quite the complaint. The complaint is that there are lots of planet features which individually make for a more interesting game. In the 2D version, you just got a planet with all of them. Now you can either search the galaxy for a planet with most of them, or have a less interesting game; and for all that now every planet is generated in agonising detail, once I'm playing, I'm on one planet and the contents of the others are largely irrelevant.

The excessive modelling is an issue if it diverts development effort from things that actually would make playing the game more pleasant; if the depth of modelling is jarringly inconsistent (like the examples above; brewing does not require water, a master sock-maker does not know that socks are made and used in pairs); or if the modelling is not actually a very good model and produces silly results (fire acts more like a slow disease than like, well, fire) in spite of enormous effort.

[Granite26]

All I can say to that is that it takes a lot longer to build the foundation of a skyscraper than to build a whole house.

[G-Flex]

a dwarf might be in one of at most six states of health; healthy, with a temporary injury and fighting on, with a temporary injury and seeking or receiving medical treatment, permanently crippled but still able to work, permanently crippled and a dependent of the fortress, or dead. So why model thirteen kinds of suture in exposed guts?

Because those are all broad categories and not actual states. A dwarf with one leg has a different handicap from a dwarf with one eye.

The history of the world is modelled, but when a bronze colossus turns up, who cares if it killed Arnar Bibbleface in 524 with a slightly rusty steel mace? Not I - it presents the same challenge either way.

Keep in mind how unfinished the game is here. You don't care because right now, nothing outside your fortress really matters at all when it comes to fortress mode. Changing this appears to be a rather high-priority development goal. In the future, if a bronze colossus has been terrorizing some civilization from day one, that civilization should care if you decide to kill it. Even in adventure mode, stuff like that does sort of matter and will more in the future; a mayor might decide to enact judicious revenge on a creature for killing his grandfather.

My second subgripe is cognitive dissonance from inconsistent simulationism. There are umpteen kinds of plant each of which brews its own booze with the specific characteristics of each booze modelled... but brewing does not require water. There's a complete model of hydrostatic pressure, but tap into a river below a waterfall and comedy ensues. The lovingly detailed "strange mood" subgame can produce an artifact pigtail sock, but no-one has reflected on the quantity in which socks are ordinarily produced and used.

Fair enough, but again, I feel like you're judging it without enough regard for its current development status. "Brewing should require water", for instance, is a rather commonly talked-about thing and will likely occur. Water pressure quirks are a little more difficult, since modeling flows more accurately (although they really aren't THAT bad right now) would probably incur a serious cost in computational complexity.

Third is simulationism versus gameplay. DF has done surprisingly - incredibly - well with gameplay through emergent properties, but gameplay does seem to be losing out. There's a reason there's a hardcore that still plays the 2D version, where the river/chasm/magma gameplay was explicitly scripted in.

Most of the advantages of having those features in the 2D version are coming back anyway, what with the continuous attacks from the underground and more common (universal, really) underground features.


I guess what I'm saying here is that your points seem well-made enough concerning DF as-is, but not DF as-planned. Keep in mind that the game is supposed to simulate an entire fantasy world, essentially. It's easy for you to say things like "why bother generating the world beyond my embark square" right now for Fortress Mode, because it currently doesn't matter so much, but that foundation is necessary for how it will work in the future, which will be quite different. Things that seem irrelevant right now won't be forever. This is bound to happen with a product in early development; you have to keep in mind what the game is actually striving for before complaining about features you deem useless. It's like complaining that the pipes in a house are useless before the water's turned on.


And quite frankly, I like the ridiculous detail going into stuff like creature bodies and injuries. It adds detail that, in addition to being flavorful, actually adds depth to the game. High levels of abstraction work fine for some games; DF is not one of those games. It's interesting to see how a battle plays out in actual, quasi-realistic terms and have to deal with the consequences on those same terms. It makes a much more interesting story when a warrior has an arm mangled, and has to drop his shield and use only his weapon hand for the rest of the battle and succeed anyway, than it does for him to become "crippled" in some generic fashion. Realistic detail in a game like this is good, and will likely be more consequential and interesting in the future. It's hard to create an interesting emergent story out of "he lost half his hit points" compared to "he got his arm muscle torn with an axe and almost bled out before the battle ended, but was saved by getting emergency medical attention".

[Anu Necunoscut]

I'm removing your quotes of me so this post isn't a minefield of quote pyramids. 

I think you misunderstand me slightly. I don't want DF to be like MOO1 and I certainly don't want it to be like a normal 4X: I want to look at the way MOO1 is not like normal 4X games. Instead of this nonsense of ordering individual buildings built on colonies of millions of people - and having no real idea what the effect of changing the build order is - you say "build factories" and they go and build factories. Policy, not implementation.

Yeah, that pissed me off in MOO2, or Master of Magic--I'm the galactic emperor, FFS!  I'm not going to waste my time selecting every farm building on every world in a multi-parsec empire!  It is more adequate in DF however, to spend a little time with each individual dwarf.  Scrolling through each dwarf's labor screen constantly is not at all fun, however.  Optimizations are necessary in other areas--for example, a placed bed should default to making the enclosed space containing it a bedroom, with the standard ability to (q)uery it into something else if need be.  As is, laying out living quarters absorbs a prohibitive amount of game time.

I think the fear of "playing itself" is misplaced. I don't want the game to make decisions for me (albeit that some sensible defaults might help the process along a bit - for example, a job priority system would probably default to doing Clean Fish before Fishing), but I do want to automate away anything I can do which is repetitive and requires no thought. For example (once I can check fortress stock at all) I could maintain x stock of y by checking the fortress stock every fifteen minutes or by manually tracking whenever I use a y, and ordering new ones built immediately. Boring, repetitive, requires no skill on my part - automate it. But I don't expect to be able to say "build me a farm" (that said, a macro language for standard constructions wouldn't hurt, but that really is wishlist stuff).

Here I completely agree.  Macros are present in the d# versions, and the fan utility Quickfort (found somewhere in General Discussion) allows you to build, mine and place items from a spreadsheet template, which is great for modular forts.  But yes, this stuff should be in-game, eventually.  I think Toady has some conception of these problems, given the auto-loom and auto-butcher features, it just needs extension as you say to more tedious maintenance aspects of the game.

Well, quite. As I mention, dwarfs only have about half a dozen distinct states. A dwarf is healthy, hurt, or dead; if hurt, they will or will not make an eventual recovery. Now the modelling of injuries in hideous detail is amusing once, and thereafter it contributes basically nothing, and represents a vast diversion of development effort. Furthermore the current system is very hard to balance compared to a simpler one - to write a creature that will, say, provide a reasonable challenge to well-trained dwarves is tricky at best without copying an existing one. Look at the way totally random creatures have become King of the Beasts. Carp?

I don't agree here.  It's amusing more than once for me, to begin with, to see limbs and heads fly everywhere in a heated battle.  I can see why some don't care if Urist McConscript lost his fourth finger, right hand, but I like the organ system, spine injuries, eye gouging and severs very much as they are.  More than this, having weapon/armor/skin materials interact will provide a lot more complexity and flexibility in combat--your corps of uber marksdwarves isn't going to be too effective against magma men, for example.  Your sword-wielding adventurer isn't going to slice up a colossus.  This could be coded in with standard rock-paper-scissors balancing, but it'd be nice to have something more.  As for buggy critters, that's to be expected I think, and enjoyed or not as the case goes. 

Well, quite, rather than being added where they are most needed to paper over known gaps in the game. (Blah, blah, yes, it's his game, he can work on what he likes. But the question was what _I_ don't like). For how many years, now, have dwarfs been incapable of responding sensibly to being on fire, and fire itself a slow-burning but irresistible contagion which isn't actually very like fire?

Fire looks pretty good spreading on a grassland (or blowing up booze), but your point here stands.  I'm sure we all have our ideas of where Toady -should- be working on the game.  Given my experience with these sorts of indie projects, it's amazing he's still working on it at all.  That doesn't mitigate the criticisms, which often hit home, but it does provide some context I think,

But I don't think there is any effort to strike a balance there at all; sim complexity is added without any real consideration of the gameplay effects.

Well, the fluid mechanics have quite celebrated game effects and several utterly practical uses.  What aspects are bothering you?  The preponderance of plants?  This comes mostly down to flavor and happy dwarves, no?  Do engravings of fort history feel as worthless to you?  I agree they aren't core mechanics that deeply influence your play, but they're not awful to have around.

But, again, is there ever going to be a real effort to ensure the procedurally generated antagonists are equally varied and interesting? Procedural content can go too far. Consider a roguelike; the individual levels and monsters therein are procedurally generated, even the depth and branching of the dungeon may vary, but you never turn up to find the dungeon consists of two rooms, three asthmatic goblins, one cheese ration, and the Amulet of Yendor.

It will seem like it will never happen until it happens.    But what you say is true--challenges must be heavily sought out by the player, as opposed to being provided in a balanced way by the game in a way that they are present in most areas.  That is a problem.  Who hasn't wandered into a Nethack level that consists of mostly Leprechauns and Nymphs, though?

The current - and even the projected future - situation where interesting terrain features are scattered all over the world and a given site has only a small proportion of them strikes me as analogous to that. So what's the answer? Use Site Finder to find the spot in the world with them all. Why not just _generate_ a spot with them all and have done with it? The terrain in the rest of the world is basically irrelevant either way.

 

The idea would be to have enough diversity of elements and enough varied challenges that result from their unique combination in a given site/world, such that there is no "boring" site, though there may be easier/harder sites.  Right now you have a choice of laughably easy or moderately tough sites.  This needs work.  Certainly a mature fort even in a peaceful area should attract many dangers and powers from the outside world.

Sure, eventually there might be _so many_ fascinating terrain features that any site will have a good selection and a site with them all would be obviously mad. That strikes me as being a very long way off, though, and right now a reversion to the 2D model where you just get a site with all the good stuff wouldn't hurt.

I note, however, that very few people seem happy with sticking to the 2D version, which is readily available.  Since that good stuff is static, its dangers are static, and the formulas for dealing with them remain the same.  If DF allowed for more and more interesting and challenging features to be accessed by digging deeper, unique to a site/biome/world/whatever, that to me would be more satisfying than a scripted site where everything is expected after one solid playthrough.

[bloodnok]

I think there's two main things here before we're just going round in circles. One is the "jam tomorrow" syndrome. I'm kind of used to the small-roguelike/7DRL scene where every game is in development - but you still have to judge the game released now, not the game you might at some point in the future be able to play. DF might realise its grand vision and then simulating the entire world will make sense - then again, it might not (and frankly the thrust of development, which appears to be adding an endless series of minituae picked at random, doesn't inspire confidence). I only really want to comment on the game I can play now, and in _that_ game the entire-world-simulation is a bit pointless.

2D versus 3D - I'm not trying to say the 2D fixed-layout was a bed of roses, far from it. However, I think - if you were trying to optimise playability of the game as is - there'd be a happy medium where the embark site itself is procedurally generated with a view to an interesting game not to being a coherent part of a world that presently doesn't matter much. Most of the useful-but-dangerous features would be guaranteed to exist somewhere, but exactly where and how would be an open question (apropos of which, some means to have skilled dwarves infer the geology without digging up the entire map would be nice); the resources on the map would be ordinarily artifically weighted towards self-sufficiency (unless you ordered up a challenge game) so you wouldn't find you had no flux or anything annoying like that.

Incidentally, I do totally agree with your contention that the nature of the embark site could more strongly inform the kind of gameplay you have.

[Footkerchief]

(and frankly the thrust of development, which appears to be adding an endless series of minituae picked at random, doesn't inspire confidence)

Which minutiae are you thinking of here?  There's a ton of minutiae in the upcoming version but I can't think of any that weren't motivated by an upcoming major feature (the Army Arc, in most cases).