Below I will be referring to both water and magma. If I don't want to make a distinction, I will just say 'liquid' or 'fluid'.
I know there are a couple threads on the subject floating around, but most of the ones I've seen are either horribly out of date or else self contradictory. I've done a few experiments myself, but I wanted to know what if anything people have figured out how to do, and whether or not I'm wrong about anything.
First, can liquids evaporate while falling?
It's easy to see that liquid, unlike walls, falls at a fairly slow rate, often taking several ticks to drop a z-level. I am fairly certain that liquid does not spread out while falling, and that all water in a tile will fall at the same rate (that is, you never see a 7 turning into a 3 and a 4 on the z-level below) and thus, if it lands in a pool of its opposite (magma into water or vice versa) it forms exactly one block of obsidian, replacing the magma and water which were there before.
Second, is it possible to use pumps to form blocks of obsidian?
I'm pretty sure the answer to this is no, but I want to see if other people have had different experience from me. I tried setting up a pump that took water and pumped it directly into a magma reservoir, in the hopes of forming a plug directly in front of the pump. This just resulted in a lot of steam and seemed to drain both the magma and the water.
Third, is there any way to find exactly how much liquid will be pumped in a given amount of time?
Specifically, assuming that you have a huge reservoir of liquid with a pump at the bottom of it filling directly into a smaller reservoir, probably several z-levels below, can you say that (provided you are filling the reservoir enough to avoid any evaporation) the reservoir will be exactly filled to the 4/7 mark, or something like that? Tied into this are two smaller questions, which are the first question I asked and 'is there any minimum amount of time which must take place before there is any chance of liquid evaporating?'
Also, there are a few things which I'm pretty sure are true, but about which I may be mistaken.
First, liquid cannot spread in midair. Speaks for itself.
Second, if water falls into magma (or vice versa) as a single block, then it will form a single block of obsidian, whereas if it falls off of a ledge it will fall 1/7 at a time and potentially form several blocks placed unpredictably around the first block, as water falls onto the newly formed obsidian and then pours off to the side.
Third, and this is the weirdest one. If water is dropped onto magma from any height of at least one z-level, while the magma is on the ground, then obsidian will form. If water is pushed directly into magma, from a pump or just plain from flowing under some circumstances which I haven't quite figured out, then there is steam and no obsidian, though the magma and water vanish. If water and magma are both falling, presumably from any heights reasonably close together, (tested for differences of up to five z-levels) there is steam but no obsidian, and some of the magma is used up while the rest spreads out.
Some of the above behavior is documented in the wiki and in other places, but I couldn't find any explanation for some of the other stuff I saw and I could find no clear answer to any of my questions. If there is some kind of liquid megathread that I passed over, please let me know and I apologize for not searching more thoroughly.
In case you're wondering, I'm trying to make a large scale precision automatic casting apparatus, and am having problems in getting magma/water to fall in a straight line without spilling and making extra obsidian. My most recent plan is to build a set of bridges and hatches to control where magma gets to fall into the water, but it's stupidly hard to prevent too much magma from spilling through the hatches after you open them. Any advice, related or unrelated to the fluid dynamics aspects of the project, would be appreciated.