Interesting idea, but not classically timed chess. I’d like to see two or three championship matches with 18 games, champion keeps the title in case of a tie.
Indeterminate length matches would interfere with the world chess calendar. I suspect that the length of the most recent match was set in part taking into account the schedule of other events that Carlsen was already committed to.
A 24 game match would take up a month or more. I doubt we’ll ever see another world championship match that long.
Fischer-Spassky 1972 had the first game on July 11 and Spassky resigned game 21 on September 1.
Spassky-Petrosian 1969 ran April 14 to June 16 with 23 games.
Petrosian-Spassky 1966 ran April 11 to June 3 with 22 games.
Botvinnik-Petrosian 1963 ran March 23 to May 20 with 22 games.
So, for a 24 game match, 2 months seems like a better estimate.
Note that the hall would likely have to be paid for, even if the match completes early.
GM Short is leading the study group considering format changes. Our Ruth Haring was to serve on that group. 
I talked with Nigel at some length on this. I expect we will see more games, fewer rest days, no rest days on the weekends, and a similar tie break structure. I don’t think you will see the champ with draw odds - Nigel talked at some length about the large advantage that would provide. But he won’t be the only voice on this.
I doubt you will see a return to the championship setting all of the conditions.
I believe all of the matches in the 60’s and the Fischer-Spassky matches had days built in for adjournments.
I’ve discussed the issue a number of times with Nigel. We’ve been friends for close to ten years. A certain woman GM who shall not be named brought us together initially. He’s heard from many people. While he was leaning toward 16 games, I think he could be convinced to support 18 games. A similar tie break structure will lead to the same problem we presently have, to wit, classical championships with a real potential for being decided with a non-classical time control. As Short wrote in his NIC piece, “…if the Challenger cannot, say, score 8 1/2 points over 16 games, why should he deserve to win?” IMO, the challenger would not deserve to win at a different time control.
I’d be curious to know if the various web sites covering the event in real time (or very close to real time) experienced a drop off in site visits once Carlsen/Caruana went to rapid games. Several of us who were meeting at our local coffee house for some of the classically timed games stopped watching in real time once the games went to rapid timing. Everyone I’ve spoken with who followed at least the classical portion of the match, was against using rapid and/or blitz games to determine the championship. The small, and admittedly unscientific, survey I posted on Survey Monkey had 38.46% favoring G-90 tiebreaks, 38.46% favoring no tiebreaks with the champion retaining the title and 23.8% favoring G-25+10.
If adjournments existed, yes; but the days existed and could be used for rest or preparation if an adjournment didn’t exist. It’s not clear to me that some days wouldn’t still need to be built in, particularly in a longer match, although you are correct in assuming that some could be cut.
Depends on the rule. There was a recent game between Nakamura and someone else (I can’t remember the other player, or what event it was) in which they repeated the position three times very early in the game, both players claiming that they would be at a disadvantage if they did anything else. This was an event with a rule against quick draws, and I believe they discussed this game for an hour or two before deciding to allow the draw. It was a draw “by rule”, but still very controversial.
And, the point is, the game was a draw.