True, but then you have to look at Unearthed Arcana (1e), v3.5 and Essentials (4e) in the same light as the Tasha changes. There is a tradition of "mid-edition" changes as well.
I sorta do.
Note: some of this isn't necessarily directed at you; your comment simply prompted thought.
In my mind, Essentials -while compatible with 4E- wasn't really 4E anymore. It was different enough and had a different enough design mentality that it was essentially (no pun intended) a different edition of the game. The Monster Math for 4E was finally somewhat fixed right before Essentials*, but it didn't really matter because (what I perceive as) shifts in the mentality and principles behind how/why parts of the game were being designed were different enough to play out as a different game at the table.
(*And, I may be mistaken, but I believe it was at some point admitted that some of the Essentials stuff -much like Book of 9 Swords era 3E- were early tests of new-edition concepts.)
I can't speak on behalf of UA for 1E because I have no experience with the product.
I'm not quite sure where I would place 3.5. Much like Essentials, it was technically compatible with 3.0, and (I think) the mental thought behind the hows/whys of game design were (mostly) the same, but there were some significant changes to core components of how things like creature size and weapons worked which altered a lot of things to which those components were connected. Trying to use 3.0 supplements with 3.5 was trial-and-error when it came to figuring out what was fine and what produced extremely broken results. I want to say 3.5 was effectively a new edition, but I think there was less of a difference between 3.0* and (early) 3.5 than there was between 3.5 and (later) 3.5 products, such as Book of 9 Swords (which were early tests of 4E concepts).
(*honestly, my memory is fuzzy on 3.0 because the switch to 3.5 happened so fast. I bought 3.0 books, went on a military deployment, and came home to a group who played with 3.5)
Do I think systems should change and evolve? Yes, they should. It's natural that a design be improved and made better once faults are seen. Though, I think there is a way to do that while still maintaining a coherent design philosophy behind doing what you're doing. If the philosophy itself is flawed (and that does happen,) I believe there are ways to approach that, but I feel as though getting into that would start to go into a rather long tangent from what I'm attempting to currently say in a long-winded way. (FWIW, it is my opinion think some aspects of 5E design are partially flawed in that things pointed out during the playtest were "fixed" in ways that I don't feel accurately addressed the feedback given and because monster/encounter design seems to have kept some of the 4E problems without the benefits.)
Mechanics is a big part of it, but I also think "design mentality" (a phrased I've used often here) is part of it too. The "why" of doing something is often just as important as the "how" of doing something when designing a product. While the "why" isn't necessarily a mechanical change in many instances, I believe that a drastic change in the "why" can cause the methods involved in the "how" to be applied to a product in ways which produce other changes. I'm not even sure if that makes sense; it involves a lot of things like "feel" and stuff that I don't yet have adequate language to express to others, but I think there's a valid point somewhere in there.
Upthread, somebody mentioned the difficulty in consistency when multiple writers and designers are involved in something. That's true. It is difficult. That's why television shows and movies which are part of an ongoing series typically have a "bible" to help maintain consistency. The Marvel movies are not by any means perfect, but many of them are good examples of making sure that established details "make sense" or "feel real" in the context of that "in-game" world. Designing a rpg (both mechanics and fluff) is something I feel could benefit from doing the same thing.
(I assume this does happen behind the scenes of WoTC, but I don't know. If not, it should.)
Personally, what I perceive as violations of established parts of the franchise are part of why I do not feel the Captain Marvel movie was very good. IMO, It would have been perfectly fine as a stand-alone and unconnected movie, but I did not like it as a movie which was intended to be part of the larger story and franchise. I think the character, as presented, would also be an example of the "mechanics" of handling a fantasy situation -like those presented in a rpg- in a way which is inconsistent with a lot of what came before, and the presentation makes several drastic changes to both the "why" of things and the "how" of things. To be completely fair, other movies in the collection are guilty of the same, but not quite to the extent of that particular product. I think it serves as an analogy to D&D because it is a portrayal of fantasy stories, and each movie could be viewed somewhat as splat books which add new things to the overall "game" (both narratively and mechanically). Differences to the "why" and "how" in that movie's portrayal connect to other products and alter the collective whole.
Highlander 2, with the weird alien planet story line might be a better example. I think most people can more-easily agree that was awful and a radical change from what was established in the first movie.