Psionics has regularly been an unbalanced feature of D&D since the first edition 20-odd years ago. Originally, it was simply an overpowered "gimme" that you'd get when rolling up a character. If you got a psionic power set, you were superior to all other players and monsters.
Second edition improved that a bit, turning the psionicist into its own class, so that one was no longer just a standard class with imbalancing psionic powers. However, the levels at which the powers came in were generally far too soon. A teleporter could easily have clairvoyance to see into an unseen room, teleport in, steal stuff, and teleport out to get away cleanly, all at 6th level. I am not kidding. Now, this could not be done often each day, but it was doable. A wizard would have to be 9th level to gain access to spells that would simulate this.
3rd edition cleaned up a lot of this. Now powers came in at levels commensurate with similar wizard spells. Unfortunately, however, this turned the psion into the party's utility box. Need a teleport? Need an invis? Need a stoneskin equivalent? Need instant mental communication with the rest of the party? The psion has plenty of power points to supply all of these needs. And absolutely no worthwhile combat power outside of mind control. 3rd edition also introduced a new class, the Psychic Warrior, which was probably its best contribution, combining nifty fighter abilities with a good power list of psionic buff-ups. The prestige classes in this edition were poorly designed, and no one would trade the poor power-point progessions in these for the next level in psion.
Now what about the book in question, the expanded rules for 3.5? Well, I discuss the prior editions to introduce perspective, because the whole history of psionics rulebooks in D&D has been a sequence of drastic alterations of rules because the prior rules were horribly imbalanced.
Not anymore.
Yes, this ruleset rehashes much of what was in 3.0, but what -has- changed is significant. Psions are no longer toolboxes: invisibility is gone, as is the free "stoneskin". Fly and Teleport require that the psion specialize in psychoportation OR spend a feat to gain the power. The basic powers have a good general rule: each power point does 1 die of damage, and you can spend up to one power point per level. So it doesn't matter what level you get the power at, so much as how many points you spend. This has an interesting implication: to get the equivalent of, say, the wizard's lightning bolt cast at 10th level, the psion spends 10 power points -- that is, effectively spends 2 3rd-level spell equivalents. At 20th level, sure, the psion can do 20d6 damage, no problem, but has just spent 4 3rd-level spells to do so, and any psion casting regularly at maximum power will run out of power points after 10-15 uses.
The result is amazingly balanced. More raw, up-front power than the sorcerer, and more flexibility than the sorcerer, BUT the sorcerer can go for much longer, and can do much more damage in the long run. For example, a psion might be able to do 10 10d6 "lightning bolts" in a row, but the 10th-level sorcerer can do 17 of these, 4 of which can be Empowered to the equivalent of 15d6 damage. Don't be too quick to say that the psion surpasses the sorcerer in this respect -- I'd favor a sorcerer against a psion in a psionic-magic duel anytime. The psion's advantage is a better tactical versatility than the sorcerer (which is saying a lot, since til now, the sorcerer was the most tactically versatile class), at the cost of running out of power -very- quickly should the fit hit the shan.
As for the rest, the prestige classes are much better than before, BUT the metamind needs serious rethinking. Even at 15th level, with all 10 levels of metamind, a psion has more power points (by almost 100), and the metamind has a single minute of infinite power-point spending (i.e., once a day, he can use powers without deducting points for 10 rounds, all sequential). The psychic warrior, well-designed last edition, is even better in this edition, removing things that should be fighter-unique, and adding more capability for psionic fun.
There remain some imbalances, but it is unclear whether they are typos or poor design. For example, Energy Missiles get +1 DC and +1d6 damage for each extra point (so you get 5 20d6 electro-missiles, for example, spending 20 points at 20th level). This should probably be per 2 points, not one point, since the DC becomes horrific in this case. As a general rule of thumb, if the result is too good to be true, it's either a typo or a broken rule, and in either case should be remedied or disallowed.
There's more to it, of course, but my space is limited. Suffice it to say that you need to keep an eye out for the occasional overpowered rule, but otherwise, this is the best designed and balanced official ruleset for D&D psionics to have been published to date.
The author, Bruce Cordell, is in my estimation among the best d20 rules designers in the business. I find that he and Monte Cook write the best source material for 3.0 and 3.5 that is out there, with an eye to optimizing fun without destroying the balance of the game, so keep an eye out for future work by either of these two game designers. This rulebook is no exception.