I think the DMG suffers from the Jack of All Trades issue - it's supposed be a DM's primer, a toolbox for adventure creation and optional game components and a host of reference charts & information (such as magic items). It's either going to have to drop something or be a (much) bigger book.
Yeah, that's what I was trying to say when I called it a "dysfunctional book design".
Honestly I don't expect much to change because the whole D&D community is too married to the "core 3" tradition, but that structure doesn't allow much room for improvement.
At least I would remove everything too basics from all books, and put it into the "Basic Box" product. Think anything that you generally need to be told only once, and most people buying core books already know from other sources. I really don't think the DMG should ever be thought of a beginner's book, possibly a "beginner DM"'s book but not an absolute beginner. Even the beginner DM reads the PHB first. Someone wanting to become a sports referee doesn't need to be told in the referee's course that the game is about throwing a ball into something and that the team with a higher score wins... if you don't even know that already, go watch some match first.
Moving topics between the 3 books is always possible but never easy. I think they tried to move magic items to the PHB in 4e? If you do that, it significantly steers the whole game one step towards players' entitlement, which is not good for everyone's preferences. Also, usually the PHB is already packed with important stuff, not easy to find room there at all. If you move magic items to the MM (I would personally like the idea) instead you're going to end up either with a much larger book, or have to cut 1/3 of the monsters, or dimish the information or the artwork about each monster.
It's just an idea, but if I could change the DMG layout, I would also put "running the game" upfront, expanding this part to at least 1/3 of the book (not 1/10 as it is now), organize it a bit more like the 3.0 DMG which had sections on how to handle each specific tricky part of the game like invisibility, magic resistances, getting lost etc., and these were already visible in the
table of content on the first page, not buried in a confusing index at the end of the book. The best thing, would be to present each topic first from a RAW point of view (if there is any RAW, which doesn't have to be the case since this is not 3e anymore) and then present the DM with different ways to manage the situation in practice. I would write this whole "running the game" part of the DMG without assumptions on whether the DM uses publishes adventures, creates her own, or improvises everything at the table.
Only after the "running the game" part, I would have chapters on designing your own stuff or using variant mechanics, because only some DMs actually need these, while I would argue that every DM pretty much needs to run the game! But the 5e DMG starts on the wrong foot with its top-down worldbuilding guide, possibly giving people the impression that they
have to know how to do it, which is not true. In fact, we usually acknowledge around here, that worldbuilding is mostly a futile activity that we do because we like doing it, but is almost never going to matter to our players. Designing
adventures and encounters instead, that is generally more useful if you ask me, even if you use published adventures you might have to come up with something quickly when your PCs step out of the adventure's expectations; but here the main problem is that adventure design is a very wide topic that might require a whole course rather than a chapter in the book, which runs the risk of being so generic and thin as to be useless. I'd put NPC design under the same category, as it's more about the characters than their stats, but
monsters design is probably even less common to do by a DM, partly because there's always a lot of monsters to choose from already, and partly because it carries some stigma of being too mathematical and prone to errors (in reality, it depends on the rules of the edition, it could be even quite easy in some games, but it matters a lot whether who writes about it in the DMG takes the approach of a rigid rules-based procedure versus staying rules-light).
Optional rules modules are certainly fun to have in a DMG, here the difficulty is choosing which ones to include, and which one to save for another book. I have my own preferences of course, but the only thing that bothers me about the 5e DMG is that it marks some stuff as "optional" in a way that makes you think that the rest is
not optional, but there are even gaming groups who do not have the DMG, plus as the thread says many DMs have the DMG but didn't read it... it is a flaw of WotC game designers not realizing that many rules are optional by nature, such as for example those rules under Exploration and Social Interaction, because not everybody actually needs to have hard rules for these pillars of the game. The best advice the DMG can give, on the first page, is to treat an edition's core books and especially the DMG itself as a toolkit to make the game your own, but it's in vain if the rest of the DMG still talks to the DM as certain things are "must do".
