All of which were spells. I rest my case.
The fact is the only thing the Psychic Warrior got was a bunch of bonus feats and ALMOST ONE THIRD the psi points of the full Psion. All of the awesomeness was because the f****ng Fighter could use Psionics. About half the powers known and a third of the points available. It was a freaking 5e Eldridge Knight with Psionics instead of magic.
*drops mic
*hands mc back to [MENTION=6855204]tglassy[/MENTION]. "You're not done with this, yet."
At a core, there are really only three classes: Brute, Sneak and Spellcaster. Everything we have could be approximated in those three classes (or a multi-class between them). So, noting that there is a similarity between prior Psionic builds and the 5E Eldritch Knight really doesn't matter too much - The Ranger is like the Paladin. The Cleric is like the Druid. The Monk is like the Ranger.
So why don't we just have 3 classes?
Because we want variety. We want to specialize. We want ... character for our characters.
To that end, we're best served by asking the question of what creates the most character for our characters when it comes to class. What is the best way to create a home for a fantasy character design?
We have to two primary options: 1.) Make a class that embodies the core of a fantasy icon, or 2.) Add a subclass to an existing class to 'fill that class out' to embody the concept.
We're discussing two character concepts that are not well served by the existing classes. One is the psion, a psychic combatant that uses the power of the mind to achieve effects similar to, but distinct from, magic. The other is the jedi-esque warrior that merges the power of the mind and the blade.
So why not just use wizards and fighter/magic users for these concepts?
Because they are not spellcasters. The core of what they are is alternative to magic. It is psionics. It is the power of the mind to control the world.
To that end, if we're going to get these concepts in our game, I want to see them done with respect. I want to see them treated as true conceptual builds where we don't just their elements on to existing classes. Sure, there is room to tack a little psionics onto existing classes for subclass builds, but for these core concepts of the psionics in D&D we need to have classes built that serve those concepts specifically and intentionally, rather than just as a tweak on another design.
If I were writing a psionics book, I'd definitely have two core classes (psion and psychic warrior). I'd have a power system for these classes that was *not* a direct parallel of spellcasting. Their powers would generally not be written as spells. Instead, they'd be closer to Warlock invocations. Power Points are a sacred artifact of the Psionic World of D&D, but I'd modify how they're used - PCs would have fewer of them (like Ki points for a monk) and they'd add to the core abilities of the class, but they would not be required to be used to feel like you're playing some form of Mentalist or Jedi.
I've thought that they might be the classes that could lean most on the 4E style of character design. They'd have abilities that do more than just deal damage, and do it on every strike - and don't require slots, charges or points to do it.
In terms of access to powers they could gain keywords as they advance and psionic powers might require multiple keywords to be used - and might have a basic use, plus uses that happen when you augment it with power points. This allows players to create tapestries of related abilities rather than pick and choose select optimal spells. Heck, I'm still a fan of characters not being able to learn certain spells to provoke variety - you could do something similar here and give them a chance to learn keywords, and if they fail they have to learn to look at other designs for their class.
Basically - When I think of playing a psion or psychic warrior, I don't want to feel like I'm playing a different type of wizard or a different type of fighter... I wasn't to feel like I'm playing a super-hero with mental powers or a jedi knight. If you constrain the design of these concepts to the architecture of existing classes, REGARDLESS OF HOW HARD YOU TRY, you're going to miss out on the opportunities you have to service the designs by making them their own classes... and both ideas have enough variety behind them to support several subclasses.