mshilarious said:
it's easier to screw with a tube because there are more parameters. You can change their performance by modifying plate voltage or filament current. Changing bias can alter the sound in a major fashion.
True, but it's been a long time since I've seen a compressor or a mic pre with grid, plate, or load controls on them...and I'm not about to put my old Yaesu linear amp in my signal chain

.
The question was whether there was such a thing as having too many tube devices chained together. If it wasn't a problem for Ellington, Sinatra, (Arlo) Guthrie, Elvis (pre-Costello), or the Beatles, it's not a problem for us. No more than having all transistors in line is a "problem".
Like Fletcher is always fond of saying; what people seem to misunderstand these days, quality tube gear is designed with the attempted goal NOT to color the signal. If there was a major "tube sound", the design was thought inferior.
mshilarious said:
PS My Apple ][ experiments would have been c. 1983-4. By that time one didn't need to write a network protocol, because 300 baud modems were commercially available
What, a 110 baud acoustic coupler wasn't good enough for you?
Nah, we were just connecting across the room using homemade (what would later come to be called) "modem cables" made out of 4-conductor phone line. No need to convert to analog and back again, just flip the leads for pins 2 and 3 (TX and RX) and throttle the signal via either pin 5 or 7 (I forget which one now). If we wanted to connect over the phone there were acoustic couplers available even in '79-'81. But we never bothered with that until the 300 baud modems became available. But when the Hayes-Stack 1200 came to town, that was heaven, Baby!
G.