Lyxen
Great Old One
For arcane casters, spell selection and loadout were (and still are) quite optimizable.
I think we need to distinguish between optimisation of your character build, and the much more simple choices that you make when playing it. At the time, I don't recall anyone spending hours in his basement dreaming about builds. You came to the table, you rolled your stats, and you started playing (compared to other games for example like Chivalry and Sorcery or Traveller, where there was actual building of the character and potential optimisation).
Spell selection was not a difficult choice and daily loadout was for all casters, and usually fairly standard unless you had a good idea what you were going to encounter. I would hardly call that optimisation.
For martials, weapon proficiency selection could be optimized as well, but only to a point.
Again, very simple choices with few consequences.
The other major point where optimization occurred was rearrangement of stats, which IME nearly every 1e DM allowed. This one ain't a big deal to me, particularly in a system where having a good stat in your prime requisite is so important.
I would hardly call putting your best score in the one important score for your class "optimisation".
For the masses, yes. Some early birds trained up during the late 2e splatbook era.
Yes, I thought I would mention these as they were really the proto-3e, but I am not sure how many people actually played with them to any serious degree.
Mild optimizing at the who-cares level - e.g. putting a high roll/number into your class' prime stat(s) and choosing feats and skills that suit your class, stats, and character idea - is fairly strongly hinted at in the PH.
Again, I would not call this even mild optimisation, I suppose it's a question of vocabulary, for me it's simply not gimping your character on purpose.
What's not hinted at or suggested is the sort of all-out power-building that (IMO rightly) gives optimizing a bad rap.
Indeed.