The way we did 'em at Davidson (mid-'80s thru '90s) seems to me to be the best, especially for efficiency — we made them so quickly that they were cheaper to us than buying pre-made forks from Tange. We even sold finished forks to a couple other companies. Probably after our markup they were a bit more $ than Tange forks, but those wholesale customers were selling high-end bikes where the higher quality of one from 'big D' was worth the small price premium. But I just made 'em, I wasn't privy to the business details. One company, after buying a hundred forks from us, switched to having them made in Taiwan with knock-off crowns identical to ours except for that company's logo cast-in, by Long Shen or Everest or who knows. Ours were precision cast in Japan by Hitachi, noticeably finer if you have the eye. The Taiwan forks were a little less visually perfect but I'm sure they were much cheaper and most customers will never notice. And the cast-in logo is a flex.

I hope Bill won't mind me describing his method, since he's retiring and probably hasn't made a fork in some while.

• Braze crown to steerer and straight blades in one go, on a 4-fork carousel. Two oxy-propane rosebuds at a pre-warming station before it gets to the human brazer, then a big hand-held torch to finish. They heat the crown rapidly, but while also soaking the heat to the deepest areas. Brass preforms inside the blades, so they're brazed from the inside out. Close to zero post-braze cleanup needed, pretty much ever. The whole heating cycle took about 2 minutes. That's one minute at the pre-warming station, and one with the human brazer. Station #1 is a second person (apprentice) fluxing and loading, unloading finished forks after they cool at station #4, and locking/unlocking/rotating the carousel. Pneumatic cylinders to lock/unlock, and move the rosebuds out of the way when unlocked, to clear the fork, then back in on the next fork as it locks, all from one foot-pedal. (Bill singlehandedly designed and built the carousel, and it was flawless! Rotated on a Volvo wheel bearing from a junk yard.) We often completed a batch of 50 forks in under an hour, around a minute each, but with two people working that's two man-minutes each. One of whom could be a low-skilled worker. Keeping up that pace would be hellish if you had to do it all day, but it was more like once every-other week, and it was actually kinda fun. Me and the apprentice sort of racing each other, because if you got ahead, you got a little rest (maybe 5-sec.) between forks.
• Do a preliminary alignment check so they're equally spaced right to left and front to back. Often no change needed, so maybe a 1-minute job on average, counting the ones that did and didn't need any adjustment. If a blade needs to be moved, that's done with a heavy soft-rubber mallet, which takes around one second. Check - whap - check - maybe another whap and check, done. With practice you know exactly how hard to swing the mallet for a given amount of movement needed, so the first whap is usually all it needs.
• Curve both blades at once, obviously on a fixture with two curved forms. These float horizontally to align themselves with the blade splay angle, which varies depending on how wide the crown is. 30-second job
• Cut the small end to length on a fixture that indexes off the brake hole, so the tips are equidistant from the brake hole. Well, two fixtures really, the first one with an abrasive cutoff saw that removes the little (2 cm?) straight section from raking, and gets 'em close to the final length (30 seconds). The other fixture on the lathe, to make them perfect (another 30 seconds). The lathe cutter reams and faces the small end at once, exactly like a tiny headtube reamer/facer tool. Both these fixtures just have a 6 mm pin that goes through the brake hole, and the fork is dropped on and held lightly by the worker's free hand, zero time spent locking it down because it doesn't need it.
• Braze the tips, which were a cast part that plugged into the ID of the blade, and had a shoulder that mated up with the faced end. The blades being lightly reamed, fit the OD of the plug with a just-right brazing clearance and were guaranteed to have clean shiny steel for braze to bind to. The dropouts had their cylindrical plug surface lightly cleaned (cast surface barely removed) by an apprentice who could do a hundred in a sitting, probably 30 seconds per pair. Brazed with pre-placed springs of brass (purchased already wound up, from Silva), super clean and almost no post-braze cleanup. About 1 minute total to braze both sides, maybe 2 minutes for cosmetic cleanup after the flux soak.
• Final alignment. Wheel sits straight close to 100% of the time, and we were picky about alignment. Never filed a dropout, which is a substandard way to align a fork IMHO. The proper procedure takes longer the write up, and I probably have, check the archives? Basically you increase the curve on the longer blade, or decrease it on the shorter, then compensate up at the crown so the tips are back to the same place fore-aft. But regardless of how the fork was made, you should never file the dropout. It's visible evidence that you failed.

But what if you don't want those IC dropouts — you want to use classic slotted type? The fixture to hold the fork to the lathe, for facing the small end, could also be mounted at a right angle to the lathe bed, where a slitting saw of the correct width could make a slot for slotted dropouts. Run that up to a hard stop on the cross slide, so the slots are exactly the same distance from the brake hole.

Now, regular forged dropouts like Campy have a somewhat imprecise rounded shoulder, which could be a source of misalignment. Many FBs file that rounded shoulder to a sharp 90° one, but the filing isn't precisely repeatable, so each DO can be a little different length from shoulder to axle. We made that repeatble by machining the rounded shoulder to a sharp one, again on the lathe, so they were exactly the same length from the axle. Took a few minutes, but worth it to have forks come out straight, first try ever time. The arbor to hold dropouts on the lathe for repeatable shoulder-squaring is a bit hard to describe but if you just looked at it you'd grok the concept instantly. It was just a slug of round bar that you'd chuck in the 3-jaw, with a bit more than half milled off at the end, so it's a D-section. Then there's a tapped hole for a shoulder bolt that secures the dropout, shoulder at nominal 9 mm that fits the DO just as an axle would. The D-section is milled at a slight angle to the lathe axis so the part you're turning is on the axis, correcting for how dropouts are slightly bent for R & L.

All those special tools and fixtures cost money, so this method doesn't lend itself to one guy making one fork. It sure paid off over the thousands of forks we made though. Maybe there's some take-aways there for the one-man shop making one fork, dunno.

OK I ended up writing a dang book again, I must enjoy remembering those times. Apologies if it's tl;dr