Disclaimer: It was going to be a review-like thing but a lot of the things have been said so well so many times here and elsewhere that i cut most of it away. Let's just say this game deserves the praise and high scores it has got. What's left is a few recurring thoughts and ideas i've had while playing, (mostly in group, soloing a little) which i couldn't stop pondering due to my work as UX designer. I haven't read this elsewhere so i kept it and decided to post:

Potential with digital cards:
A thing that becomes clear when playing text heavy games like AH and MoM is that quite sooner than later, you are bound to get narrative repeats. Back when this was card based, there were three good reasons for it: The card pile must be of managable size, not a towering monster that threatens to fall over the board, or divided and hogging more and more of the table. The other is manufacturing cost. The third is subtle, but the quality of the experience depends on it: a larger deck can lead to unfair streaks, in either direction. In the digital realm, the first two prohibitive reasons are nonexistent, which means the main obstacle for writing more 'cards' is basically the manpower required for copywriting and quality control. Hence, it is more viable to write a pile of stuff that eventually reaches beyond the treshold of clear memory for the player. The third obstacle, being of importance, still needs a system to be handled correctly. How this would work and still be manageable is that you'd have nested layers of randomization. The first layer is purely technical: at scenario start, randomize/shuffle a virtual deck (list) of effects. (for example: Test lore, 2 must succeed, then monster damage x+item, else player damage 2). Each post in this array points to its own array of entries which holds several narrative versions eventually leading up to the same game-mechanical effect. The ideal program would select one of these in revolving order (the variable is autostored, and doesn't reset on new sessions) and present it to the players through the UI. This way, things are put neat and game designers can manage narrative expansions without altering the in-game balance.

Had this been with real cards, the only sound way to manage would have been to create more separate monster decks (new 'types').
I've read comments from people that monster battling should have been left out of the app. Either way works for me, but i can see how something like the above incentive (to silently be able to add narrative card content, post-release or as part of an expansion) would help justify the choice for those who felt it was unnecessary. However; the real benefit is helping enhancing the replayability and long-term popularity of the game.

More potential - On computer randomization and passive chance:
Another effect the app has had on gameplay is the item decks etc. You can't (the app doesn't) instruct the player to pick an item at random from the decks theyre likely (as advised) to be sorted alphabetically. The game can mimmick this by randomizing an instruction on which card to pick from the deck (and keep track of items in the background, which it does, which is important), but experiencewise, it's very different. I believe that something important gets lost here, and that could be brought back in a future expansion. What if the app instructed you to draw a card from a face down chance/treasure/mystery deck where you won't know what it could be? It can be useful items, clue rewards, one-shot spells/scrolls/artifacts, tests, traps/curses/cursed items or any combination or all at once in one or two decks. Keeping tests mixed in would certainly add to the suspense. The app wouldn't need to keep track of these externally randomized things as long as their content is within reason (ie. no cards that instructs you to pick another card from the decks that the app does keep track of).

Memory markers:
When you save and quit, and return to a game one or two weeks later, it's not so sure you'll remember positions of players and monsters, what was barricaded etc. A workaround would be to make notes, of course. Though, the user/s could be able to put virtual markers on the playfield in the app, preferably in different colours.

Dissapearing barricades:
Moreover, barricades should be permanently visible on the playfield so you get help remembering they're there from session to session. As it is now, they disappear right after you are instructed to place them. This point is less of a pondering and more of a 'this needs to be fixed' kind of thing.