Print

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

[FantasticDorf]

@Whatsifsowhatsit:
1.0 is a decade or two away.

That's a particularly pessimistic but reasonable estimation. I agree very slightly with that number.

Also to take note ten - twenty years is a long time, especially given that both Tarn (38) and Zach (40) aren't particularly young. None of us can properly read the future how their personal lives will unfold, but by the latter end of that estimate, they'll be in their 50's and 60's, though in a noble fashion we'd be honoured if they continued on with it, it'd be understandable if they wanted to step back from DF and mentor the new development hosts (if any are ever any 'successors') so that they can focus on themselves and have a well earned break and enjoy a legacy.

Technology as well, in 10-20 year may be better than we imagined. To pull a early 90's Jeff Goldblum film role monologue, 'life will find a way' and the least probable result in terms of scientific progression always remains within the realm of possibility for as long as the possibility exists that can be achieved by obscure but methodical happenings to be exploited (a butterfly's wings in Brazil causes a hurricane in Arkansas, or a creature revived and copied via genetic mutation is able to divide its sole gendered caste by exploiting a oversight of its creators in the content of its creation, its a unlikely and usually unforseen methodical process.)

Such as the eventual move to scientific testing method in space not constrained to a space centre, such as a laboratory on the moon, or storing data (as microsoft are making breakthroughs now) inside DNA so that one day we may run computer hardware off a liquid container of hybrid silicon carbon organic material at super high bit capacity that creates little to no heat. Then of course there is the challenge of keeping up with this technology and keeping it relevant to DF's application (moving 64 bit to 126 bit/whatever is relevant).

Look at the computer's 30 years ago compared to a laptop today. I've mentioned this before here and there but computers are slowing down development as per every 2 years due to the development process being stifled. In some abstract kind of way, zach and tarn NEED a market breakthrough in order to achieve their goals in a 'reasonable time' else DF will have a longer road to the set 1.0 finish line.

That's beside anything else that might happen across the way.

[Max™]

You mentioned world-gen heroic items and holy relics (the temples will be so awesome!) and I am curious if this will be a thing where you get the option to name something after some point, or even having them acquire names like we do after 5 hist-fig kills?

On that note, I'm sure you intend to get around to it at some point, but would other things which would/could bestow a title besides the kills mechanism in place now be a sooner or later type of change, like something that comes in with a language structure/dictionary rewrite?

[Inarius]

sometimes you get an elf commemorating the slaying of a minotaur by naming their wooden low boot which presumably saved their life or something

Toady, I quoted the last dev blog about magical items. I'm not sure I have understood it perfectly.
Does it mean "heroic items" are the item counterpart of the "heroic name" given to people when they have fought enough ?
If this is the case, will magical items/artifact still have a name before they are involved in some fights ?

[PatrikLundell]

@FantasticDorf:
The time to produce the DF 1.0 version will likely change significantly only if there are significant changes to the production process such as the expansion of the production crew to a significantly larger team (but I absolutely share Tarns view of the horror of becoming an administrator rather than someone who does hands on production) or a radical change of production methods, such as e.g. where some kind of AI would do the coding and testing based on descriptions of what to achieve, and be smart enough to ask what to do about boundary cases (unlike the current manager wet dreams of automatic error free code from high level sloppy "specifications", without the need for pesky programmers). If Tarn leaves DF for whatever reason DF 1.0 will never happen, because the group taking over would not share his vision or patience, so DF would take off in a rather different direction, and the results left when that group gets bored is going to be something else (not necessarily all, or even mostly, bad, but quite different).

While it's an impressive stunt to manipulate DNA to encode info, it's just that, at least for any kind of general computing. It's a slow, noisy, and messy process, and reading the data is very slow and noisy as well. In general, biology is a very poor model for digital reliable processing at it excels at producing similar results from a noisy and unpredictable environment, while sort of recovering from errors, and at a biologically relevant speed. If you're looking for something near magical I'd rather look towards 3D holographic storage technology, but whether that will work is of course unknown.

[Max™]

Think they're the "Urist McAxeLord has bestowed a name upon a steel battle axe!" items in game now, blue in the artifacts list instead of yellow.

Where did you get 'Urist' from? Has it always been (oor-ist) like "urge" instead of (yer-ist) like "urine" as I assumed? I was dismayed to hear you using the "urge" (oor-ist) sound in a podcast recently.

[Random_Dragon]

I've always wondered how those two created the in-game languages, yeah.

[Sizik]

I've always wondered how those two created the in-game languages, yeah.

'

From DF Talk 17:

Capntastic:   EggFibre asks, 'How did you create the languages for all the civilizations? Did you base them off of something in real life or did they all come from your imagination? Also, how long did the process take?'

Toady:   We created all the languages. They were not based on anything in real life and the main reason for that is because we haven't added grammar yet. What we did is we had the list the words, which took a while to just type in all the words - the English versions of the words - and then we had a generator that just had certain rules about how consonants could be combine and what frequencies there were for the vowels, and that kind of thing. It had to generate a list of words and then we went through by hand and picked out words that were either existing words - especially words that are profanity, because it would generate plenty of those - and words that just sounded wrong; when you look at the word for candy or something and when you end up with some really harsh sounding word, and then you just roll the generator on that again until it gets something that matches. That's really all it was. With goblins we let more things go through regardless of their tone because we wanted the language to sound more alien, but it was pretty much just that. When we get to things like grammar it's going to be harder to completely ... we'll just put in different processes there, but I don't think it's going to be randomly generated because you either have a rule or you don't, or you have variants of a rule or whatever, but there's only four or five stock languages, so we're probably just going to pick and choose which rules we think are appropriate and place them in.

Then there's the larger question or what about a randomly generated language. The computer can spit one of those out really fast with random grammar and random words and all that, but then you will have the problem where some of the words are really not something you want to have. We haven't tried that yet, and it might not be that bad of a problem in retrospect so it could be that we do the random generation and then what we'd do it just ... depending on how we store our sentence trees, when we get to that kind of thing, because computers are good at that sort of thing, so you can just throw it in ... It's obviously going to be simplified from the giant thousands of pages of, you know, even entry-level textbooks you can get on this kind of thing, it's progressively more complicated theory that they've got floating around now, but we throw some simple stuff in there and I imagine it'll all work really nicely. The only thing I'm not sure about is when you get a lot of words that are recognizable or so on, if that breaks it too badly. I mean it would be interesting to have a conlang generator and just see how it turns out.

[Max™]

Oh yeah, I recall seeing that somewhere else in part, it's just funny how much Urist feels like it should have existed already, is that just me?

[Random_Dragon]

Oh right, I could've sworn I saw t hat before, seeing it again. And that sounds almost as derp as just doing my normal worldbuilding exercises for fictional words. :V

[King Mir]

I agree with Fleeting Frames, although King Mir is correct in that DF is very far from s a perfect candidate for parallelization. Since everything has to synchronize at the end of a tick, you can run things in parallel DURING the tick and consolidate the results at synchronization points during the tick. Different methods can be used for different sub problems.
One example is pathing, where you can have separate threads calculate paths for creatures. At a synchronization point you then validate the paths (to ensure they paths calculated are still valid) and recalculate the ones that are not. Validating a path is a reasonably cheap operation compared to calculating the path in the first place, so there's a potential for a gain here, but there's overhead in organizing the farming out of the path calculation, in addition to the rather substantial effort of ensuring the data used for path calculations is stored in a thread safe manner that doesn't slow down things more than you gain.

The trouble is, syncing every tic, by default every 10 ms, is expensive. An that has to happen for every shared bit of data, one way or another. That's going to eat away any gains you might have gotten from parallelism. Plus there's going to be a significant portion that you can't parallelize. Plus there's the challenge of breaking of breaking up work into suitably equal size chunks...

It all means there's just not going to be significant gains from multi-threading.

All of this is dependent on the CPU being the bottleneck and not memory bandwidth, though. If memory is the bottleneck, the additional administration of threading might actually result in a loss.

Yeah, that's another problem. DF is known to be significantly bound by memory latency.

[King Mir]

@Whatsifsowhatsit:
1.0 is a decade or two away, and I think DF will have slowed down too much by that time unless parallelism is used to some extent.

I don't think this is necessarily true. Adding new features does not have to mean making the game slower. AFAIK, Toady does not update his hardware often, and the game will continue to be playable on the hardware he uses.

[Putnam]

Consider how the game runs currently.

[Bumber]

@Whatsifsowhatsit:
1.0 is a decade or two away, and I think DF will have slowed down too much by that time unless parallelism is used to some extent.

I don't think this is necessarily true. Adding new features does not have to mean making the game slower. AFAIK, Toady does not update his hardware often, and the game will continue to be playable on the hardware he uses.

Playable to the extent that he plays. His forts don't last long enough for FPS death.

[Random_Dragon]

How many players even HAVE forts that last long enough to suffer FPS death? More often than not, death by boredom claims my forts, especially these days since kobolds and goblins never seem to visit. :V

[PatrikLundell]

How many players even HAVE forts that last long enough to suffer FPS death? More often than not, death by boredom claims my forts, especially these days since kobolds and goblins never seem to visit. :V

Most fortresses suffer FPS slowdown if played any length of time, but where the death threshold is individual. Enjoyment falls as the (real) time it takes to do something increases without any "valid" reason.  I certainly constrain my fortresses because of it (playing dead civs to keep the number of dorfs down, having minimal amount of trees on the surface [a rainfall of 3 seems to cause some trees to be present on embark, but no saplings to appear], constant house cleaning to get rid of old clothes, avoiding flowing fluids,..., but haven't yet shrunk the embark size below 3*3). Boredom is likely to set in unless you have some kind of goal for your fortress: random plodding along when you've already probed the secrets isn't likely to keep the interest up. Getting sieges seems to require more careful planning than it should, since you run a significant risk of having a crappy outpost as the closest goblin settlement (and it's not unusual for it to be set up during the 2 week embark period), causing the sieges to be drawn from a goblin pop of 50 or so, and thus be rapidly depleted.