healing from gods should be made less effective (but still possible)
How does this fix anything? The incentive here, remember, is to open up DD-of-not-Makhleb, in a way that's at least not more abusable than what you can do with current DD-of-Makhleb. (If you don't care about opening DD to other gods, then current DD is no worse than heal-on-kills DD or other suggestions in this thread; as I've said the design problems are pretty much unsalvageable so it's just a choice between different bad things in the end.) Unless you make god healing so bad that it's basically worthless (in which case--why not just go all the way and make it not happen at all?) DD of Makhleb will still have a really huge advantage if it gets healing so you're still strongly pushing DD into Makhleb.
--
The main reason I'd suggest DD getting heal on kills instead of other options is that, frankly, it's pretty much identical to the current situation, which is apparently acceptable to the devs since it's existed for close to ten versions now (and personally I really hate special-casing things, so I value consistent behaviour--nothing but innate heal-on-kills works--over other concerns like "but there's already a no-healing-consumables (except not really because potions of magic etc. exist)" race (yes it bothers me that VS can use !magic)). Other things like allowing DD to heal
only from heal wounds are things I've considered in the past but I suspect they don't work very well in practice (especially if you want DD to appeal to anything resembling a variety of player skill levels), even if they're less abusable from a game design standpoint than heal-on-kills.
(I guess if you want to adjust item generation specifically for DD then you could actually do the heal-only-from-hw thing, by generating a lot more heal wounds potions, but then you're special-casing something and this is, I assume, a lot more work than just giving DD innate heal-on-kills.)
--
Removing maxhp entirely (and figuring out what to do with torment/flay) doesn't have any design problems that I can see but I suspect that in practice it will not prove very fun, since it's actually even more different from normal crawl than current DD is. It would be tremendously difficult to implement in a way that proves enjoyable and not abusable (you need to gain hp more rapidly for exploring in later areas than in earlier ones, but at what rate? how do you do this adjustment? (if it depends on XL or skills you're still potentially creating problems) how do you deal with fighting normally giving maxhp? is it even realistic to make this lead to games that are both challenging and winnable reliably? etc.). So the idea is good but I have pretty serious doubts that it's realistically doable in crawl. This sounds much more like something you would adopt as an idea when creating a game (it's pretty similar to desktop dungeons, as has been mentioned, and was clearly adopted as a design decision early on) instead of something you would attempt to shove into an existing game.