Sleeping is treated as being unconscious. Thus if an elf maintains consciousness throughout their trance, they are fully aware.

The big difference here is that the elf maintains consciousness. We do not at this time, have rules for what perception is like when a PC is sleeping (considering the use of unconscious, I'd guess it's automatic failure), so we can't necessarily determine what effect the semi-conscious meditation has on perception checks.

I'd say it's safe to assume that an elf can make perception checks (though generally for this kind of thing, passive is used), as normal with no penalty while meditating.

It boils down to what kind of style you like. Most systems have an inherent style to the way the world works and people tend to pick the game that fits their favorite style.

The style in D&D is medieval (or early renaissance) sword and sorcery, where themes of powerful wizards wielding magic, gods being real and bestowing powers upon their clerics and supernatural beasts ranging from demons to dragons are all part of the storytelling conventions enough that players would expect and readily accept these as part of the game. Note that in all of these cases magic is a regular (one might even say natural) part of the world. Even sorcerers only affect the power of the "weave" of magic that is omnipresent.

Psionics on the other hand are a set of powers that are not tied to the world itself, but rather to the characters and creatures themselves. Some very classic D&D monsters wield psionic powers, such as Illithids, Aboleths and the Gith. However, since these are considered strange, outlandish or downright freaky in the context of the world, they are easily accepted by people liking the sword and sorcery style.

Having characters, which are an integral part of any RPG story, wield powers that go against the assumed ground rules of the fiction's genre is often where the willing suspense of disbelief breaks down. Remember, in sword and sorcery magic is a (somewhat) natural part of the world and mages, etc. just know how to affect this otherwise hidden power.

In different settings more separated from these assumptions, often in a world closer to our modern world rather than a medieval setting, psionics will often be considered as acceptable as (or even more acceptable than) "conventional magic" by most people. The reason is similar to what I described above: in such a modern setting magic that permeates the world seems out of place, while powers bound to individuals and ancient items are imaginable.

In the end, it boils down to what assumptions everyone brings to the table. A genre trope of sword and sorcery is that magic is part of the world and characters only shape what is already there. This is why psionics get so much hate in the standard D&D settings.

The fact that game designers often want to separate psionics from magic in terms of rules only adds another layer to which people can refer for their criticisms. As you wrote, there usually are plenty of badly balanced parts in D&D anyway. In my experience, nitpicking about the rules for psionics often is a surrogate discussion. Maybe one is not comfortable with defining the boundaries of the campaign's genre, possibly afraid of being considered close-minded or adhering to clichés. Maybe one can not even put into words what is the actual issue and then starts to argue about rules just from a bad gut feeling.

I myself have belonged to the latter group of people for a long time, trying to argue why psionics is bad rule-wise, without realizing that it just is not a regular part of medieval fantasy worlds that I can or want to imagine.

In the end, if you find yourself in such a discussion, it might help redirecting the focus to what themes are considered part of the canon of the game. The people arguing against psionics would most probably just as much hate a gnome tinkerer "inventing" modern weaponry.

As with most matters of opinions, there are no actually wrong ways of thinking about it. Still, it helps the game a lot if everyone at the table has the same (or at least compatible) angles on the subject.