Vague ramblings on things SCS...

[nexusdog]

In our own thread now, to avoid hijacking and heading off on tangents (sorry evilchatbot!)

Okay, let's go back to this issue with handling large populations of 'people' in games, starting with large scale battle rts's

What I was trying to say was that the sheer scale - volume of people, involved in the battles has to have some relevance to the likes of a city building game with it's own population.

Each population has an AI, however rudimentary or whatever. Essentially, they have goals or 'preferences' - some will favour walking, cars, buses. Next, there is how far they'll travel to get to their destination; be it work or the store for groceries or venues.

Compare that to the mass battles in the former example - what stops individuals from adopting flocking behaviour and engaging en mass onto one individual at a time, against one to one combat, or nearby enemies. This applies not only to 121 combat, but also other types of units - long bows, catapults or whatever else is included in the game.

I don't know how to describe this other than motivations - what is the drive internally for a sim to do what they do?

Actually, I think I've gone so far off from my original point that I no longer know what I'm even talking about now... There was a point, but I'm losing this! AHA, right found it.

The original point was how do other games handle masses of population onscreen, each doing their own thing - AI & pathfinding issues immediately spring to mind. Then after that, we have the issue of 'personality', a la the Sims.

Now, since mass battles have more people compared to say SCS (massive assumption, I know) how is it done and how would it translate to SCS's population handling?


AI - ?
Pathfinding - ?
Personalities - ?Did I miss anything?


The free version can generate maps 512 by 512 pixels (or 513 x 513 vertexes, which is correct for SC4 use). Lesse, 512 / 64 = 8 small cities by 8 small cities. So you may use World Machine to generate a fantasy map no larger than 8 x 8 small cities worth.

1km tiles? How much detail would you want, though, given the option? Using the DEM info you've mentioned, if a way to blend the two were possible, that would provide ample realistic terrain - it's simply the impetus to do so. There would have to be a massive pay off in terms of revenue to actually develop features we're rhetorically discussing.

But as an option, yes, I think it would be both progressive and forward thinking if EA & TM 'pushed the envelop' that far. God, I hate lingo..

But I digress on that point, and I'd have to discuss privately how to use World Machine for SC4, or cover it in a tutorial on my website www.mik-maq.com (http://www.mik-maq.com/).

Good plug! :D I'll head over right away and bookmark it.

Anyway, my goal was to show that proceedural generation is limited in the fact that the apperance of your terrain, and making it look the way you want strongly depends upon your ability to manipulate the math

Surely this is purely down to a way of making a front end user friendly, if someone were of that wont? Blimey, that's some ol' English there!:)

Now, I've just booted up WM and... I see what you mean. It means nothing to me, despite being laid out well, as I have no idea what any of it does. I just want to make a pretty landscape, easily. No hope there, then.

My 'Whatcom County' region isn't proceedurally generated

link? MD or CJ?




For the free version, that area is no larger than 512 pixels by 512 pixels, or the same as 8 x 8 small cities in SC4.

I think this would make an excellent tutorial - class, get your pens out, we're terraforming today!:)

(10 mins later...) Okay, you can play with the default setting and move around, but like you say, it's pretty rudimentary and there's no sense of depth of scale or detailing.

Go for that tutorial, Romaq, I'd be interested to see what's possible.

Regarding trading the expense of an artist for the expense of a tool. Regardless of a tool not knowing what beauty or art is, at some point you have to market a product you expect people to want. And your market is what determines 'beauty' and 'art' is, at least what beauty and art they are willing to pay for to enjoy.

I can only refer to what Spore is doing in this capacity - that's the reference I'm drawing on, from what I've seen possible and discussed in the vids.

EDIT...It's cheaper to hire an artist and a programmer make those buildings 'slightly randomized' things like base color, trim and surrounding props, than to have 'buildings themselves' proceedurally generated.

Again, I'd refer back to Spore - best of both worlds?

Phew, that was harder than I thought.

Next subject - how to stop global poverty...

[Romaq]

In our own thread now, to avoid hijacking and heading off on tangents (sorry evilchatbot!)

Ooops! I started a thread as well. Well, the one that gets the traffic is the one we'll reply in, I suppose. It matters not to me. ;-)

Okay, let's go back to this issue with handling large populations of 'people' in games, starting with large scale battle rts's

In 'battle/ army' games, you have units comprised of squads (around 5 to 10 soldiers) and companies (comprised of 5 to 10 squads). Companies have tasks they are assigned. Within each company, a company commander designates squads to accomplish parts of a task. Within each squad, individual soldiers cooperate to accomplish their assigned goal. There are exceptions to this, but they are exceptions that prove the rule. The amount of CPU spent on any given soldier, squad or company depends upon the scope of how closely you look at them. The more you look, of course, the more there is to see. But that would fall under 'flocking behavior' and other programming methods along that line.

What I was trying to say was that the sheer scale - volume of people, involved in the battles has to have some relevance to the likes of a city building game with it's own population.

Because the dynamics of military units provide cohesion (people doing pretty much the same thing for the same reasons) the math behind making that happen is relatively cheap on the CPU.

Now, applying this to SC:S... you *can* 'cheat' in the same way you do with the military by grouping people, and then running the math upon the group, not upon individuals. Do you group people by their job? By the homes they live in? By the venues they prefer? How far can you push the computer into calculating individual behavior while keeping the sim running at a sane pace? Do you use simple calculations on individual sim behavior, and then 'fill in the blanks' when the enduser queries an individual sim? Can you store events as a number from a dictionary, and simply have the tags follow the individual sim around? Then when the query is made on a given sim, those tags are looked up from a dictionary to tell a sim's life story? Dwarf Fortress (http://www.bay12games.com/dwarves/) is a fascinating look at Dwarves living out their lives within a 'sim', and it may point the direction TM is taking. I wouldn't know.

I would bet those are issues that TM has already resolved within their game engine, and I would also bet they are not telling us how they resolved it until the game is released, if even then. We'll have to examine sim behavior for ourselves and reverse engineer how TM did it.

Each population has an AI, however rudimentary or whatever. Essentially, they have goals or 'preferences' - some will favour walking, cars, buses. Next, there is how far they'll travel to get to their destination; be it work or the store for groceries or venues.

Compare that to the mass battles in the former example - what stops individuals from adopting flocking behaviour and engaging en mass onto one individual at a time, against one to one combat, or nearby enemies. This applies not only to 121 combat, but also other types of units - long bows, catapults or whatever else is included in the game.

The AI is programmed differently. In 'battles', there are different sets of behavior from individuals within a cityscape. With an army, you have 'us vs them', and you don't tend to have an entire army engage a single soldier of the opposition simply because units persue units: squads are tasked against an opposition squad or specific objective. Companies battle companies. There's nothing 'personal' within the AI. One soldier will shoot at another because it is a squad level task to do so.

There are differences, but the same concepts apply to 'sims at work'. My wife and I will travel to our job shortly. Most of the people in the building I live are of similar income level, and we all tend to work at roughly the same time. We may work at different jobs, but we'll tend to have the same types of jobs. We'll tend to go to the same types of venues. So take 1/3 of the building sims, have them go to the same place of work. Do the same with the other two thirds. Now have another random third go to the same venues, and come back home, and do the same thing with the other remaining two thirds. You'll end up seeing groups of people from the same building going to much the same jobs and much the same places, just not precisely so. Pretty much like real life, even though we all think of ourselves as individuals.

I don't know how to describe this other than motivations - what is the drive internally for a sim to do what they do?

That's what you determine as a programmer. That 'flock of birds' ai on that java page shows that. Avoid obstacles, persue food, avoid being eaten. Those are simply motivations you program in. "If you can see food, head that way. If you see something blocking your path, go around it."

Actually, I think I've gone so far off from my original point that I no longer know what I'm even talking about now... There was a point, but I'm losing this! AHA, right found it.

The original point was how do other games handle masses of population onscreen, each doing their own thing - AI & pathfinding issues immediately spring to mind. Then after that, we have the issue of 'personality', a la the Sims.

I don't know about The Sims. But if you are very much interested in these areas, I would strongly suggest you look for articles on http://www.gdmag.com for articles, and if you *can* you can buy the articles. If you are not in the position to purchase the articles (or simply get a subscription and wait for topics to repeat) then you have direction for doing your own online research. I don't mind discussing the topic at all, but there is a LOT of depth to it. It's kinda like talking about the whitecaps on the bay while there is a world full of ocean to explore.

Now, since mass battles have more people compared to say SCS (massive assumption, I know) how is it done and how would it translate to SCS's population handling?(snip)

The devil is in the details. In *this* case, you can pack ALOT of 'detail' controlling entire armies of individuals in a military sim simply because any individual sim has only a very narrow range of choices, and those choices are mostly dictated by ever higher levels of military control. If your programming breaks down from a General to a Major to a Captain to a Lieutenant to a Sargent to a Private having to fire a riffle in a particular direction, the processing for this sort of AI is relatively 'cheap'.

AI for SC4 is *also* relatively cheap: You don't have 'sims', you have 'aggregate numbers' representing sims doing stuff.

I'm not clear how much detail SC:S is packing into each individual sims. The more 'smoke and mirrors' TM is using to fake individual sims, the cheaper the CPU needed to make that happen. The more 'smoke and mirrors' you use to fake individual behavior, the more points at which the model breaks down and causes the player to realise that 'sim' is nothing more than a computer faking a person, and not doing it very well. That's the trade-off, and why a city's population level might be constrained far more than your normal city builder expectations.

1km tiles? How much detail would you want, though, given the option?

For SC4, exactly and precisely 16 meters per unit of detail. Detail does NOT equal size. How BIG do I want my area? How big may I have it before startup/ region load times become too long to be playable? For me, that's about 4,295 square kilometers, or 64 small cities by 64 small cities. As the region becomes larger, the time it takes to start the game and switch between cities becomes unmanagable.

Using the DEM info you've mentioned, if a way to blend the two were possible, that would provide ample realistic terrain - it's simply the impetus to do so. There would have to be a massive pay off in terms of revenue to actually develop features we're rhetorically discussing.

Not necessarily. World Machine is *free*, if you limit yourself to the small cities I've described. MicroDEM and 3DEM are also free. The USGIS data for 10m per pixel data is available free. The expense is YOU have to decide what and how to do.

But as an option, yes, I think it would be both progressive and forward thinking if EA & TM 'pushed the envelop' that far. God, I hate lingo..

Then don't use the market-speak lingo. Just talk. But you'll note in my earlier posts directed at TM, I had much to say concerning terrain and the ability to exchange data with MicroDEM and World Machine. The more easily SC:S can import data from those two tools, the less need there is for SC:S to duplicate what those two tools already do very, very well in the area of terrain generation.

Good plug! :D I'll head over right away and bookmark it.

I do try to be a good resource.

Surely this is purely down to a way of making a front end user friendly, if someone were of that wont? Blimey, that's some ol' English there!:)

There are many details. User Interface, data manipulation, and so on. I'm not clear on what you refer to here.

Now, I've just booted up WM and... I see what you mean. It means nothing to me, despite being laid out well, as I have no idea what any of it does. I just want to make a pretty landscape, easily. No hope there, then.

http://www.world-machine.com/help.html will refer you to a PDF file you may download and read in your native language (for you non-native english readers). The PDF will cheerfully take you through more than you never wanted to know about how to make World-Machine make those landscapes. But at the end of it, you should make some really nice ones and THEN post some pictures here, and then DEMAND that TM make it so you can port those landscapes into SC:S.

At least, *I* wouldn't mind if you were to do that. >:)

link? MD or CJ?

I'm not clear on the reference.

I think this would make an excellent tutorial - class, get your pens out, we're terraforming today!:)

(10 mins later...) Okay, you can play with the default setting and move around, but like you say, it's pretty rudimentary and there's no sense of depth of scale or detailing.

Go for that tutorial, Romaq, I'd be interested to see what's possible.

Well, the sense of depth and scale is all relative. Instead of modeling a region of the US, you could model the surface of an tomato seen under a microscope, and there is simply no means to tell UNLESS there is a context. And YOU have to provide the context. The tutorial I would create would have to be how to make something simple in World Machine's free version and import that into SC:S, which is outside the scope of this particular forum. Is that what you are looking for me to do? I would have to work on it this weekend, in between going to Seattle to harass EAGames over Ultima Online.

I can only refer to what Spore is doing in this capacity - that's the reference I'm drawing on, from what I've seen possible and discussed in the vids.

Again, I'd refer back to Spore - best of both worlds?

Phew, that was harder than I thought.

Next subject - how to stop global poverty...

Heh. We'll just keep it to SC:S possibilities for now, at least in this thread, shall we? ;-)

--Romaq

[nexusdog]

In 'battle/ army' games, you have units comprised of squads and companies. Companies have tasks they are assigned. Within each company, a company commander designates squads to accomplish parts of a task. Within each squad, individual soldiers cooperate to accomplish their assigned goal. There are exceptions to this, but they are exceptions that prove the rule. The amount of CPU spent on any given soldier, squad or company depends upon the scope of how closely you look at them. The more you look, of course, the more there is to see. But that would fall under 'flocking behavior' and other programming methods along that line.
You deal with this further on, but what I'm curious about now is, are you saying, commanders have AI that over-rules soldier behaviour AI? Or squad or company 'flocking' AI? From what I can understand, squads look for other like squads to engage, and so it goes for companies etc.

The population of a city sim... now how would they work? I'd speculate something along the lines of intelligence or workplace, or by wealth or even those social energies - but that wouldn't be the only influence on them, there would have to be additional elements of grouping and subgrouping. What immediately came to mind though, was travelling to work and workplace.

Because the dynamics of military units provide cohesion (people doing pretty much the same thing for the same reasons) the math behind making that happen is relatively cheap on the CPU.
So it is possible to apply to SCS, to speculate once more, despite their geographical location to each other - they're still following the same rules, no matter where they are located? What are the primary motivators, basically?

Now, applying this to SC:S... you *can* 'cheat' in the same way you do with the military by grouping people, and then running the math upon the group, not upon individuals. Do you group people by their job? By the homes they live in? By the venues they prefer? How far can you push the computer into calculating individual behavior while keeping the sim running at a sane pace?
Primary motivators... or is it due to individual 'personalities'? Hmmm... I'm intrigued. Or obsessed. :D

Do you use simple calculations on individual sim behavior, and then 'fill in the blanks' when the enduser queries an individual sim? Can you store events as a number from a dictionary, and simply have the tags follow the individual sim around? Then when the query is made on a given sim, those tags are looked up from a dictionary to tell a sim's life story? Dwarf Fortress (http://www.bay12games.com/dwarves/) is a fascinating look at Dwarves living out their lives within a 'sim', and it may point the direction TM is taking. I wouldn't know.
I like the idea of interrogating a dictionary, that way, small file sizes can be used for lots of info on an as needed basis.

Downloaded DF, though getting used to all that ascii is kinda hurting my eyes. :)

We'll have to examine sim behavior for ourselves and reverse engineer how TM did it.
Interrogating sims... sounds, deliciously sadistic. :D

With an army, you have 'us vs them', and you don't tend to have an entire army engage a single soldier of the opposition simply because units persue units: squads are tasked against an opposition squad or specific objective. Companies battle companies. There's nothing 'personal' within the AI. One soldier will shoot at another because it is a squad level task to do so.
Right. So, based on my experiences with squad level games, I can stretch them out across the screen and they'll still follow orders I give them, or they've been set - opportunity fire, for example (yes, my experience is with turn based games)

There are differences, but the same concepts apply to 'sims at work'. Most of the people in the building I live are of similar income level, and we all tend to work at roughly the same time. We may work at different jobs, but we'll tend to have the same types of jobs.
Yup, to me that's going to be the primary motivator I feel. Start with what needs to be done and then thru to what they would like to do.

However, there was the issue of 'getting their arses out of bed' mentioned. Personally, if they can't, I'm demolitioning the building they live it!

I don't know about The Sims. But if you are very much interested in these areas, I would strongly suggest you look for articles on http://www.gdmag.com for articles, and if you *can* you can buy the articles...
Gamasutra was into tech details, but I don't know if it's still going or not. The rate I devour reading material would easily bankrupt me if I subscribed... yeah, I'm a skinflint, I know.

The devil is in the details. In *this* case, you can pack ALOT of 'detail' controlling entire armies of individuals in a military sim simply because any individual sim has only a very narrow range of choices, and those choices are mostly dictated by ever higher levels of military control.
Choices, narrow choices - thats subjective on the actual game in question, surely? We don't know exactly the flow chart example for choices in the military sim example, but I still figure it's going to be very similar to those within a city sim simply because the choices are rudimentary; work, venue... and those societal energies - are sims going to have more than one that motivates them?

I'd like someone to comment, if possible, on the populations of C IV and CotN and how their behaviour (in the game, not the gamers!) could be examined and extrapolated upon.

I do try to be a good resource.
Without a doubt, so much so, I have trouble keeping up.

Right now, I'm trying DF but finding the icons and controls a little hard to get to grips with, not idiot proof enough, I'm afraid. I'm intrigued by it though, so hopefully the learning curve will start to level out soon enough.


I'm not clear on the reference. whatcom county - I assumed it was a city journal or similar, on ST or SC4D?

[Romaq]

You deal with this further on, but what I'm curious about now is, are you saying, commanders have AI that over-rules soldier behaviour AI? Or squad or company 'flocking' AI? From what I can understand, squads look for other like squads to engage, and so it goes for companies etc.

The answer depends upon scale and context, both 'real world' and 'in game'. Squads are given objectives from 'higher ups', orders they must carry out. Sometimes the orders are specific ("take posession of that hill" or "Launch a mid-range nuke with these numbers" /providing a target but no identity, no idea what we were shooting at/) In a game, if you are scaled out to see regiments, I would expect 'regiment wide' or 'company wide' calculations to be performed while squad-level calculations to be absorbed and thus unnecessary for the CPU to perform, nevermind 'individual' choices being calculated. The rest is 'smoke and mirrors', you figure a certain chance at sucess and have the companies 'jiggle' accordingly, as if it were the individuals making choices and not as a result of regimental or battalion level calculations. The point of using 'smoke and mirrors' is to find a cheaper way to accomplish an end goal of what you want people to see other than doing exactly what it is they think they are seeing.

And, of course, these are the same sort of techniques likely to be used within SCS to have you think there is a great deal more 'depth' to the game than what TM actually wants to have to calculate on a limited CPU budget.

(snip) So it is possible to apply to SCS, to speculate once more, despite their geographical location to each other - they're still following the same rules, no matter where they are located? What are the primary motivators, basically?

Primary motivators... or is it due to individual 'personalities'? Hmmm... I'm intrigued. Or obsessed. :D

That's an excellent question. Presuming I actually purchase SC:S (somewhat likely) I'll take the opportunity to design various 'test city cases' to see if I can 'game the game' into revealing its secrets. You know, making it solve a few unsolveble problems, diging at the underlying algorythms with an ice pick, tickling the feet of sims, using water dripping techniques and so on. All in good fun. THEN I may be able to answer your questions, but until that point only TM knows. And it's quite possible one may find out things in the course of exploring that TM didn't consider. That's the surprise of software. Sometimes it doesn't behave as expected when pushed to extreems.

I like the idea of interrogating a dictionary, that way, small file sizes can be used for lots of info on an as needed basis.

And, depending upon the design of the dictionary, it allows for people to create local language versions. It might be possible, for instance, for a custom Klingon dictionary. At least in theory, depending upon if and how TM creates such a thing.

Downloaded DF, though getting used to all that ascii is kinda hurting my eyes. :)

Yes, though there are tilesets for different ASCII tiles to use and graphics packs you may use to replace :) with something that looks like a working dwarf.

Interrogating sims... sounds, deliciously sadistic. :D

All I require is a bit of information. And, of course, capturing and interogating David Beebe would tend to annoy various legal authorities, so we'll simply have to content ourselves with getting the Sims of SC:S to expose the secrets of the game.

(snip)Gamasutra was into tech details, but I don't know if it's still going or not. The rate I devour reading material would easily bankrupt me if I subscribed... yeah, I'm a skinflint, I know.

Or you if you are a member of a consulting firm, you might be able to swing a free subscription to GDMag. There are a LOT of ads in there, game companies paying big bucks to push middleware and creation tools. If you are an audience for such tools, GDMag would like you to have a 'free copy' to expose you to those ads.

Choices, narrow choices - thats subjective on the actual game in question, surely? We don't know exactly the flow chart example for choices in the military sim example, but I still figure it's going to be very similar to those within a city sim simply because the choices are rudimentary; work, venue... and those societal energies - are sims going to have more than one that motivates them?

Choices in the military stem from 'posted orders and standing orders'. You are never to obey illegal orders, and you are given the basis to know what those orders are. You have standing orders, such as never leaving your post until properly relieved. You have various other orders that come up on an 'interrupt' basis. It's actually rather trivial to flow-chart choices for a soldier's life.

For a 'city sim', the choices allowed are narrowed by the game engine and what it is set to portray. I don't think individual sims really have all that much choice within the SC:S context as we understand it. Remember, your goal as a game designer is to provide the APPEARANCE of depth. You don't actually WANT it to be to be 'deep', since that cuts in on the CPU budget which is based on the 'minimum system requirements' you'll put on the box.

I'd like someone to comment, if possible, on the populations of C IV and CotN and how their behaviour (in the game, not the gamers!) could be examined and extrapolated upon.

SimTropolis has City Journals regarding 'city experiments'. They are not really 'for the fun of portraying a realistic or fantastic city'. Such cities are 'built to test the extreems of what the game engine allow, or to push the game logic to the extreems of the most illogical, or otherwise test the limits of game mechanics'. There may be other such 'city journals' posted here in TM for their games, but I've not looked that far yet.

Without a doubt, so much so, I have trouble keeping up.

Time budget... never enough time to look at all there is to see. And the more you look, the more there is yet to be seen.

Right now, I'm trying DF but finding the icons and controls a little hard to get to grips with, not idiot proof enough, I'm afraid. I'm intrigued by it though, so hopefully the learning curve will start to level out soon enough.

http://dwarf.lendemaindeveille.com/index.php/Main_Page (The Dwarf Fortress Wiki) has reference to tilesets and shows what they look like, as well as object tilesets that replace an ASCII character with a graphical character, and it shows you what they look like. This may prove useful. The controls are quite awkward. In some contexts, + and - on the number keypad are for scrolling. In others, the up and down key. Take it slow, and if you really are not interested in playing DF consider simply reading about it through the wiki and the Bay12Games forums. Play a 'slow' game and focus on your dwarves and how their thoughts and moods change rather than avoiding disasters. As they encounter disasters, in fact, you'll learn more of how bad thoughts work.

whatcom county - I assumed it was a city journal or similar, on ST or SC4D?

http://www.simtropolis.com/stex/details.cfm?id=17665 shows West Whatcom, a region I uploaded to the STEX. I've not gotten to the point of CJing it because life is what happens while one makes one's plans, and I've been livin' large lately. Hope this helps!

--Romaq

[nexusdog]

snip ...I would expect 'regiment wide' or 'company wide' calculations to be performed while squad-level calculations to be absorbed and thus unnecessary for the CPU to perform.

The rest is 'smoke and mirrors', you figure a certain chance at sucess and have the companies 'jiggle' accordingly.

The point of using 'smoke and mirrors' is to find a cheaper way to accomplish an end goal of what you want people to see other than doing exactly what it is they think they are seeing.
Okay, so groups of sims will be, for the sake of argument, filtered by job, then it's open to speculation. However, those social energies seem to me to have a bearing on our LCP's as well, though I'm not entirely sure if they'll be lower or higher in the priority list. That means no matter what motivates a sim, it can be filtered and the S&M element slapped onto these groups without all that extra cpu processing.

Just thinking, now, but this is where these buildings now come into their own - they have influence factors... kinda gravitational pull, if you like, for sims - venues. Even so, how much again is smoke and mirrors and actually has a real relevance.

Jiggling - averaging out and applying the numbers to sims of a certain group, yes? Go to work, go home, go to venue... sounds pretty easy life, doesn't it?

Yes, though there are tilesets for different ASCII tiles to use and graphics packs you may use to replace :) with something that looks like a working dwarf.
DF is beginning to get to me - I can't figure out half of the basics for survival, despite my best efforts so far. I mean, how do you accomplish something as simple as gathering wood to build something, or, how do they gather food or items left lying around? I can get miners to mine, trappers to hunt and fisherdwarves to fish, but other than that, I can't even build a bed!

Despite all that, I'm still battling with it, to figure it out, but it's damned irritating, in a good way. :)


Choices in the military stem from 'posted orders and standing orders'. You are never to obey illegal orders, and you are given the basis to know what those orders are. You have standing orders, such as never leaving your post until properly relieved. You have various other orders that come up on an 'interrupt' basis. It's actually rather trivial to flow-chart choices for a soldier's life.

We shall come back to this...

For a 'city sim', the choices allowed are narrowed by the game engine and what it is set to portray. I don't think individual sims really have all that much choice within the SC:S context as we understand it. Remember, your goal as a game designer is to provide the APPEARANCE of depth. You don't actually WANT it to be to be 'deep', since that cuts in on the CPU budget which is based on the 'minimum system requirements' you'll put on the box.
Yeah, and that's a shame really - it seems like the 'what can we get away with' clause coming into effect. I think that's something which makes me wince - tight, efficient programming isn't something which I get the impression is endorsed or promoted in the industry. Compared to the modding scene, games seem to be bloatware when it comes to serious coding/programming. There just doesn't *appear* to be the impetus or motivation to create lean, mean code.

http://dwarf.lendemaindeveille.com/index.php/Main_Page (The Dwarf Fortress Wiki) has reference to tilesets and shows what they look like, as well as object tilesets that replace an ASCII character with a graphical character, and it shows you what they look like. This may prove useful. The controls are quite awkward.
I like it, I really do, but it's finding the time to learn by mistakes, read up on it and squeeze in real life, unfortunately. I need the idiots guide, simple as that, to get myself started.

Actually, this is a classic example of how to promote and create games - if it's not easy enough to pick up in half hour, then what's the point in continuing? Repeated frustration and failure without any feedback (DF could do with an inbuilt tutorial, for example), devs need a good kicking for lacking the insight and omission of such a feature. Seriously.

I am seriously finding it hard to figure out what I was talking about, y'know? :D

[Romaq]

For the first part of your message, there are ways of 'faking' depth or taking shortcuts in computing the appearance of individual behavior. The exact form of these shortcuts (smoke & mirrors as I call it) depend upon the behavior being simulated. I really know nothing of how the SC:S game engine is going to accomplish what appear to be large groups of individual lives and choices.

DF is beginning to get to me - I can't figure out half of the basics for survival, despite my best efforts so far. I mean, how do you accomplish something as simple as gathering wood to build something, or, how do they gather food or items left lying around? I can get miners to mine, trappers to hunt and fisherdwarves to fish, but other than that, I can't even build a bed!

Despite all that, I'm still battling with it, to figure it out, but it's damned irritating, in a good way. :)

You can't build a bed without a carpentry workshop. Many other things require a workshop to make it in. Your best bet is to look at http://dwarf.lendemaindeveille.com/index.php/Starting_Builds which will talk you through initial setups. I wouldn't suggest you get too wrapped up in DF. After all, the original point was to view how DF deals with 'individual non-player characters', aka 'sims'.

Yeah, and that's a shame really - it seems like the 'what can we get away with' clause coming into effect. I think that's something which makes me wince - tight, efficient programming isn't something which I get the impression is endorsed or promoted in the industry. Compared to the modding scene, games seem to be bloatware when it comes to serious coding/programming. There just doesn't *appear* to be the impetus or motivation to create lean, mean code.

'Lean Mean Code' has an expense attached to it just like anything else.

1) In order to make it 'lean/ mean', you often have to take shortcuts that limit the functionality. You might have a section of code that handles making decimal percentages just fine, but then you have to go back at a later time to 'internationalize'. If you'd just use common available libraries that just have that flexilibility built in, you wouldn't have to do that.

2) A great deal of the 'bloat' is due, in fact, to 'bloated libraries'. If a company writes third-party game engines and whatnot, they want to make it generally useful so it will sell to more game creation markets. That means there's likely stuff in the library that doesn't fit to any given game, but it's in there just in case it's needed. And a game library that TM, for example, doesn't have to write saves time and budget.

3) 'Lean Mean' code usually implies code that is difficult to read and debug. Your program may have various tricks to it that might make it work faster by introducing hard to find bugs. That doesn't help anything.

4) When dealing with microcontrollers, such as a computer that's going to be in your toaster or a computer in your cheap pocket wireless router, at THAT point you have a very limited amount of memory space, and CPU budget. At THAT point it becomes well worth the effort to focus on 'lean-mean' code.

Everything has a cost associated with it. 'Lean Mean' code has its own way of being expensive, and there are trade-offs.

I like it, I really do, but it's finding the time to learn by mistakes, read up on it and squeeze in real life, unfortunately. I need the idiots guide, simple as that, to get myself started.

Actually, this is a classic example of how to promote and create games - if it's not easy enough to pick up in half hour, then what's the point in continuing? Repeated frustration and failure without any feedback (DF could do with an inbuilt tutorial, for example), devs need a good kicking for lacking the insight and omission of such a feature. Seriously.

I am seriously finding it hard to figure out what I was talking about, y'know? :D

DF is still in 'alpha', meaning 'beta, unless you don't like it, in which case it's alpha'. It is also 'free' as in 'free beer'. He's letting you enjoy it, kick it around, no charge. There's a cost associated with things like 'free beer'. That doesn't necessarily mean it isn't fermented bear urine (Bear-Wizz Beer).

So accept 'DF' in the same light as if you were to sit down and throw something together entirely of your own work on your own spare time and release it to the world for them to like it, or not. Because that's essentially what the Dev (only one person involved) did. And, having done so, Tarn has contributed far more than many of those who post here on TM screaming about SC:S. He 'did something'. Something I at least think it pretty cool as a concept.

--Romaq