My failure for forgetting this is epic XD.
The 25% comes from the "ideal plan" (this you could argue, I'll accept that), that you'd build HE in amsterdam, as you were going to settle the GG there. This only goes the premise that you get literature and hence HE before civil service and hence bureaucracy.
At that point, you have a capital with +100% military unit production. The extra 50% from bureau makes it an effective increase of only 25%
About binary, 1 out of 15 is a 6% increase
. (I'm not saying this is insignificant btw, every advantage is an advantage) compared to 25% from the libraries. Except that's only one "point" of binary research (since waiting for libraries/next +25% tech would mean binary research wouldn't be as popular), the thread also made mention of keeping research at multiple of 5 (this is only if you are alone) due predcessor techs known. There was a few more things but I simply can't recall atm
, probably something involved with opportunity or what not ^^". The way it is talked about actually has quite a bit of subtlety O_o which I can't convey ^^". I think I'll read it again!
http://www.civfanatics.com/civ4/strategy/micromanagement.php
There's the link for any interested. Definitely suggest for reading if you are not too sure on the whole business, because I know I made some wrong statements above
(edited now but still XD)
I also retract my statement about the 6% to rounding. This is due that most of your research will certainly be in the techs between say, writing, education, astronomy etc.
. So the % might remain the same, the "extra" from binary on just the building bonus techs is alot less than 5% i imagine ^^" (possible exception if you are 0% from time you are in education till you oxford and you've build your empire around a SSC).
I guess I have to emphasise binary research more now O_o