Print

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

[sockless]

With something looking vaguely like mass and volume being implemented sometime soonish (by the looks of how DF is shaping out) I thought that it was a good time for this suggestion. Kohaku's suggestion here seems to vaguely touch on what is here.

Currently in DF, all our alloys are fixed, if you want to make bronze, you add one tin to one copper, which is also the formula for pewter IIRC. Most alloys in the game are similar to this. The main problem with this is that the ratios are unrealistic, in real life bronze is ~90% copper and ~10% tin by weight. This has been brought up many a time, so that's enough of that. However, there are many different ratios and compositions of alloys depending on what you want to make.

If instead of having several fixed compositions of alloys, we could design our own alloys with our own compositions, we would be able to make all the different alloys we want to. The different alloys would have different properties, since even the addition of 1% of a certain metal can have effects on a metals properties.

I guess this is in theory easier than in practice, since accurately modelling the properties of the alloys that you make would be difficult, especially since Toady is a mathematician and not a metallurgist.

For the average player that doesn't want to play around with the different alloys and just wants bronze, like they can currently, there won't be any change, since there would still be preset options for the existing alloys currently in the game.

[nanomage]

The biggest problem I can see here is that alloys' properties are not that easily derived from the properties of their components. Huge and complicated sciences have been built to explain what happens there. How would you propose to calculate the properties (melting point, moduli)of an alloy, which is, say .7 copper, .1tin, .1zinc and .1bismuth, so that the method is comprehensive enough for a non-specialist to understand?
The second problem is how are you going to display alloy composition in user interfaces. =(.985Fe .015C) battle axe= or something like that? 

[King Mir]

If Toady don't know metallurgy, the average player won't use it, what would be point of learning and adding it?

[Starver]

There's not just the matter that novel alloys/mixes give strangenesses such as hardness well beyond what either/any of the constituent materials exhibit (because, for example, the different sizes of the constituent atoms arrange themselves in a body centred cubic pattern, thus resisting any tendency towards ductability that each of the 'pure' substances tend to have), but there can also be large effects on end-material quality according to how the processing occurs.  Quenching of metals might freezes a materials phase to form a different quality of material from one left to cool naturally.  Work-hardening (and its inverse) exists where (for exampled) cold-forging gives different effects to pouring a much more freshly decanted metal into a mould.  As do several effects that are time-related (either through natural oxidation of the surface or to allow low-temperature rearrangement of the material's crystalline structure).  <= Simplifying a lot of the real physics, there...

I'd find it interesting if all of this could be replicated with the (being the avowed aim of the game) quantum-level simulation of the world, and all, but I think we're probably going to be stuck with, at best, an in-game-universe mock-physics (perhaps even randomly generated for each worldgen?) where the player, after getting his dwarfs to experiment, can establish that a mix with 1% unobtanium in a 66% dalekanium/33% vibranium matrix makes for a material with great strength and low weight, where an unadulterated Dk₂Vb mix actually flows at room temperature, but giving the mixture 2% infusion of Uo, instead, causes the density to rise drastically even while it still acts like plasticine.

So now you know you need your metalsmith to get the mix right, and maybe even quench it as part of the process, whereupon the material will now never break (and cannot be reformed), but act in all other respects as if it's rubber...  So it might be useful to drawing it into a fine wire, reheat then dunk it in a bucket of water, and then it becomes an indestructible thread useful for everything from pulleys to weaving into armour.

Or something like that.

[sockless]

I guess the more simplified version of this would be to have several recipes for alloys, where you can adjust the ratios to give differing alloys, so you can make 90:10 bronze, or 80:20. This would save Toady having to figure out what would happen if you were to make an Fe Sc Al Cu Sn Au Ag alloy.

And yes, strength effects is another ballpark, I was assuming that it would be up to the dwarves to heat treat it to purpose, you aren't going to make a sword out of annealed steel, they'd harden and temper it to the appropriate temper (although that does vary to individual preference, such as hardness vs sharpenibility in knives).

[Maklak]

What I would like to see is 1 bar having 150 units of metal and alloys having fixed ratios and smelter having fixed size. So to make 1 bar (150) of bronze, you'd need 135 copper and 15 tin. The ores would have 150 units metal each. The partial slabs would stack, and there would be a preference for putting the same kind of metal in the same bin. 'melt metal object' jobs would immediately produce partial slabs of metal. Metal items would require amounts of metal proportional to their mass.

[NW_Kohaku]

Keep in mind that the game does not record properties of materials per item: It records materials in definitions that each item simply points to. 

The game will either need to start generating material definitions for each bar of an alloyed metal you generate (which could be one new material definition per every iteration of possible combinations) or else take the time to algorithmically calculate the effects of each ratio of materials that you generate each time those material properties are tested... which is basically every temperature check.

The system I was suggesting would use the algorithmic method, but the algorithms in that system would be very simple addition of a small number of tokens modifying properties.  If you want to have complex, realistic metallurgy, it will likely take more complex algorithms, especially when you're throwing how the metallurgist treated the metal times the purity of a metal and what it is alloyed with.