1) In the Mishna at the bottom of 20b, the simple difference is between the objection to intruders, that is, foot traffic making noise and making the courtyard into a less private area, and noise intrinsic to the work. The children studying (or playing) make noise, but this is not the same as making the courtyard into a public thoroughfare.
The various braytot saying one may not establish a school obviously are contradictory to our Mishna and Rava, and claiming that it refers to gentile schoolchildren is a forced answer. It seems more like a matter of Tannaitic dispute. By reinterpreting those other sources, one establishes like the law in the Mishna and like Rava. Meanwhile, the easy answer was not offered by the gemara, that Rava (or by the setama degamara immediately after Rava) already made the necessary distinction, namely ומתקנת יהושע בן גמלא ואילך. Those other braytot were talking about before this takkana, so there would be more intruders, from far off. Or the takkana placed greater importance on this teaching, such that it would override the objections.
2) In terms of punishing a student:
Artscroll’s footnote #10 is based (at least in part) on the next Rashi, or at least that is what they link to. I wonder if this is deliberate.
3) The story with Yoav and reading zecher as referring to males. Artscroll writes it as זְכַר, zechar, with a sheva under the zayin and a patach under the chaf:
This makes some grammatical sense. The idea is that zechar is the construct form (males-of) and would be vocalized in this way, rather than the absolute form (males), which would be vocalized as zachar.
Rabbi Dr. Shlomo Goldberg told me of an interesting explanation from Rav Meshulam Roth (see here) that the difference is between zecher (segol segol) and zeicher (tzeirei segol). With the improper segol segol, it is parallel to eshen hakivshan, a construct form, where the absolute form is ashan. So with segol segol, Yoav believes that it was the construct form of zachar, males.
Shadal uses this as a proof that they didn’t have nikkud in the times of Yoav, because otherwise, “how did the teacher of Yoav not see any sefer which had the nikkud under ther word zecher in the verse timche et zecher Amalek, such that he taught Yoav z'char with a sheva under the zayin and a patach under the chaf?”
(Artscroll presents Tosafot’s read of the gemara, with the teacher knowing it to be zeicher, and answering the adult Yoav with zeicher, but not having corrected Yoav as a child.)
I would add that if Chazal had written nikkud, and / or names for the nikkud, this would be one of the places in which they would have utilized it, at the very least to disambiguate which reading the teacher had, instead of writing the difficult to disambiguate:
אמר ליה והא אנן זכר קרינן א"ל אנא זכר אקריון אזל שייליה לרביה אמר ליה היאך אקריתן אמר ליה זכר
This and other instances in which Chazal should have used nikkud or referred to nikkud by name is telling.