Sorry, I'm a little behind in reading this thread. (Maybe I should give up while I'm behind?)
Jan Heine wrote (snipped):
"It is interesting to compare the individual "improvements" and what they mean for performance. I limit myself here to racing bikes of the last 40 years... by 1965, alloy components, 10 speeds, geometry and everything else that makes a modern racing bike had evolved.
"Shifting The fact remains that shifting has no influence on racing performance. Tour de Frances were won with friction shifting long after STI became available. Look at the videos, you don't see Indurain (Campy C-Record) frantically trying to catch up with Rominger (with Shimano Dura Ace) after each shift. (Of course, the marketing people will say the NEW and improved 2004 model is so much better than... and if Rominger only had had the latest version, Indurain wouldn't have stood a chance.)"
I agree with you here (as with most of your points). I think that STI is much more important to an Abdoujaparov than to a Rominger or Indurain. But I believe that the example you used is incorrect. When Rominger gave Indurain his biggest challenge in the Tour, in 1993, he was still using downtube shifters on his Colnago, while Indurain appeared to be using Record Ergo on his Pinarello. Rominger raced with CLAS-Cajastur in '92 and '93, using (I think) Mavic components. In '94 his team became Mapei-CLAS, using Shimano Dura-Ace. Indurain last used C-Record in the Tour in '92. So if anything you might say that Indurain, not Rominger, had the advantage of modern technology in 1993.
Please correct me if I'm wrong here.
"Number of gears A few more gears or less don't make a difference. Triples were tried in racing on and off, and racers found they didn't need the extra gears."
Rather, they didn't want the complication of front shifting with a third ring. Racers want tight gears and also low gears for the steep climbs. With a 10-speed cluster, you don't need to sacrifice the low gears. You can use a 12 and 23, and still have small steps in between.
Of course, if you aren't racing, you can get both tight gears and low gears with five cogs by using half-step plus granny.
"Weight Only a few pounds at best. 4 lbs. on a 170 lb. combination of machine/rider/clothing/computer/radio. Racers used to carry less water to make up for it - a full modern bottle is 2 lbs!"
A 4 lbs. difference can mean a few seconds, or more, on a climb. It's also psychological: a person feels faster on a lighter bike, even if his times are the same. If frame weight were so unimportant, why would racers use carbon TVT bikes, that supposedly made descents so exciting, just to save a pound or two?
"Frame stiffness Rominger rode Alans (I believe, or which other make has dropouts that fit _over_ the fork blades?) rebadged as Colnagos. If an Alan is stiff enough for him, any bike is stiff enough. (LeMond's TVT carbon bikes were so flexible that testers found them disappointing, yet Greg won Tours on them.)"
Some riders sacrificed stiffness in order to have a lighter bike. But with today's stiff, featherweight frames, no sacrifice is needed.