You're right about everything you said... but how is any of it different than a regular soldering iron that can get up to around 900f with no precision control? I was always taught you heat up the piece you're soldering to, you don't put solder in contact with the rod itself.
It differs in several important ways, most importantly how heat is managed. Heat (as in calories) not temperature (as in degrees). In general, hotter irons are a bad idea. Optimum is probably 700F for common electrical/electronic soldering. Any hotter and you 'cook' the stuff on the iron and have even more trouble controlling the heat at the joint (keeping it from going too hot). Some has to do with a guy's timing, of course.
So, 'the NASA way' calls for a tinned and clean iron. You wipe the tip if it's full of old solder and more important sometimes, old burned flux (which of course contains the concentrated form of the very junk you're trying to keep out). Put a small dab on the iron tip to form a droplet to do the work. The flux will flash off almost instantly and form a crust at the edge of the droplet.
Use the droplet to mechanically contact the joint at it's most massive point. You need enough liquid to fill the voids for improved contact for rapid heat flow. The work is done there. When the work comes up to working temperature on the far side, feed the solder in from there for the connection. First to flow will be the flux by over 100F. The flux we use is of the so called activated type and is only active and able to attack the oxides and junk over a fairly narrow temperature range, but the time the solder melts it's spent (the other reason you don't 'carry the solder on the iron' to the work.....no flux action). After it's stripped the oxides and carried them away it will flash off (don't hit the vapor....it's the wrong kind.....). You want to keep the het on as this happens (smoking stops) as that clears the flux allowing the molten alloy to get atomic level contact with the metals.
Now, for a very critical part, the connection needs to be above melting point by 30 or more degrees, uniformly hot, then allowed to cool until well solid (say 30 or 40 below, 70 or more degrees from the start) without being disturbed mechanically. The situation is like some concentrates solder doesn't 'set up' promptly at a single temperature, but rather has a 'pasty range' where extremely small and not well linked crystals from rather than what we want. This is the so called 'cold solder joint', one where heat wasn't controlled such that it wasn't hot enough at a critical point. Marginal heat transfer makes this much more severe since it's nearly impossible to routinely set up the right conditions of very uniform connection heated uniformly through cooling normally (stress free) and not moving in the process.
Avoid 'reflow soldering' by hand, that's for rain gutters. You want to heat the joint, apply the solder, burn out the flux, give it a small bit more time to reach stable temperature then pull the iron holding everything else stable.
Cleanness, heat and flux. With modern materials, the first is less important than in the past, good gear and technique make control of heat and effective use of flux pretty routine if you know the rules. The common mistakes include not adding enough fresh flux, poor use of heat (generally too cold) and moving stuff too soon. It's interesting to note that not all defective solder connections can be found by inspection. You have to build the quality in at assembly, you can't inspect it in later on.
OF