Hubble is 2.4m - the existence of facilities to make space optics close to that size factored into the cost of the project. (It is widely rumored that there are now orbiting systems close to 4m in aperture, not even counting ESA's upcoming Herschel far-IR mission). The guidance and instrument packages would have been quite different - reconnaissance needs precise pointing, but assuredly not the ability to maintain pointing at the milliarcsecond level for hours (before launch, many of us thought this would be the major hurdle). Also, it turns out (according to some folks who claim to know, anyway) that inteligence satellites deliberately spend a a lot of time in "safe mode" anyway between taskings, which is what HST operations try very hard to avoid.