I think melee SD is a really cool idea and has potential if they made reasonable buffs to characters like they did for G&W, Link, and Falcon but what they did for characters like Pichu and Bowser is just way too ridiculous and feels way out of place.
Falcon doesn't seem to mind having a ridiculous and out-of-place design in the Melee environment.
What even defines what is in place and not? I'm sure there are plenty of players who would say something like Jigglypuff is ridiculous and out of place in Melee.
The matter isn't if something that is powerful "fits" into a defined notion of character design, but instead whether that character design fits into a notion of what is powerful. It's fine to have a varied cast, as long as they are all given equal opportunity and parallels in authentic character strengths. Balance by way of "Human error" or "low health" or "bad recovery" don't really address actual character problems in interacting with their opponent. Don't get me wrong, having a strong recovery and living long means you can afford to make mistakes and can minimalise character flaws, but those flaws are still going to exist no matter how long they live.
I mention the notion a lot, but there is an important difference between a Weakness and what I coin a "Weakness multiplier." Having bad priority is a weakness. Having low damage output is a weakness multiplier. Having slow start up is a weakness, this causing a lack of a combo breaker is a weakness multiplier. Being unsafe on block is a weakness, being gimped any time you get thrown off stage when you get shield grabbed is a weakness multiplier.
The point being, if your character, or, in a more meta sense, if your collection of hitboxes attributed to a collection of hurtboxes alongside that hurtbox's ability to navigate the stage and interact with your opponent, has no inherit weaknesses, they have nothing to multiply.
If your character has 100% safe options in every situation, no amount of low health or bad recovery or low stun meter or light weight or susceptibility to combos will matter. The character has
no weaknesses, no matter how you paint it, because the character has no flaws in their interaction with the opponent.
Now, whether having blatant and abusable weaknesses in a character's design is a good thing or not is a matter of opinion, but I can say for myself that I'm not a fan of it. Not to say that every move needs to be 100% unstoppable, that would be even dumber; just that there should not be one single weakness nearly universal to a character with nothing that exists as a contrast to that weakness.
For instance, if a hypothetical character "P" had a universal weakness of "bad hurtbox coverage on his upper body," and only one situational move that serves in contrast to this trait, that character has a blatantly exploitable weakness. Anything the character does can be beaten just by hitting high to either trade or outright beat everything they do. Their single outlying move can easily be dismissed in the meta, because it just means you have to look out for two easily telegraphed attacking options. Even though P was given a tool to combat his weakness, it doesn't really do much for the character as a whole because that tool can be circumvented without issue.
Here is where I feel good design vs bad design comes about. How do we solve the problems P has? Well he is easily punished for his low hitting moves and is weak to 50/50 mixups, so we could make him have really high health- that way he can afford to be hit more and hopefully put in work to make up the deficit. We can also make P hit like a truck, so each hit takes out a healthy fifth of the opponent's health bar. I see this as an example of bad balance and bad design choice. This does not address any issue P has as a character, and only serves to make his strong matchups more polarized. He still does just as bad against characters that can abuse his design, and now just does that much better against those that can't.
A better balance for this character design to me would be to keep him at that standard health, hell, even lower it if you want, because we're going to instead just make this character good overall. P still has a strong number of sweeps and they do in fact compose about 60% of his moveset. His upper hurtbox is still exposed on these moves, but there's a big difference in the character's design. Now P has, in addition to his anti-air elbow or whatever, a fast starting flash kick that, while lacking invincibility frames, quickly covers his entire hurtbox, that ends close to the ground. Additionally, his sweeps generally have a disjoint to them, and a combination of faster startup, low cooldown, or a little of both. Also, P has been granted the marvelous ability to MOVE! No longer is P restricted to pacing out throwing low sweeps and anti-air elbows waiting for the opponent to make a dumb mistake coming in, and he can now move without dedicating himself to a slide kick that gets beaten by any high or mid attack. He fits along with the rest of the cast even though his lows vastly outnumber those of the other characters and his design can easily be seen as "gimmicky" or "out of place."
Luckily, Project M has very few problematic design choices, and none are as extreme as P here. There are quite a few odd choices that are indicative of the beta nature of this project though, and hopefully all these can be rectified through overall better character designs without any "universal" weaknesses and instead with a conglomeration of moves that all have their own individual strengths and weaknesses looked at in likewise individual case-by-case bases.