erzboesewicht

Well we could create a language model limited to texts that make sense in the Doom universe, more precisely: the universe that can be created by the Doom engine. For example, train a neural network with the DWmegawad club, Cacowards, other WAD reviews, technical descriptions of the Doom engine, and so on. A very interesting text format to train would be probably demos transcribed to text, with descriptions of the movements of player and monsters, stats like health/armor and the environments. These texts would be quite large if they are precise, and they could use the "jargon" used by Doom players and reviewers, so there would be enough data for the AI to learn. That would be actually a fun experiment :)

Exactly. On a more abstract level: Doom is essentially a mix of a "counting game" and a "running game". On UV-Max particularly you have two counters: monsters and secrets. You have to bring both counters to zero, killing monsters and discovering secrets. (Of course, your key pressing and mouse moving skills are what determine your ability to "count", not your mathematics skills ;-) ) And for many players (including me) "exploring" and "killing", i.e. lowering these counters, is what's part of the fun. Nightmare turns it basically in a almost pure running game. The exploring and counting aspect may remain on a very basic level (in NM100S you have to lower the secret counter) but in general the maps are not balanced in a way you really have time to enjoy your journey.

E3M2 ("Slough of Despair"). When I started with Doom I considered it confusing and boring. It was where my first playthrough stopped, for months. But now I see it as an interesting sandbox for combat experiments, and one of the more engaging maps in Doom 1 due to the (light but significant) space restriction. On the contrary, I could cite MAP27 of Alien Vendetta as a map I loved first (for the blue archvile section, for example, and its general variety), but now I consider it a almost completely unconnected, ultra-linear and slow affair.

Sphagne. Is known mostly for some big but interesting maps in the Community Chest series. Seconding Bob Evans, B.P.R.D., Sandy Petersen, Sverre Kvernmo (who it seems is still active in the forum and every 5 years or so releases something), Use3D (idem), and the Casalis. I'm not so much a fan of Erik Alm (save some of the later Scythe2 maps) but it would of course be interesting to see what he'd do in a modern context :)

Erkattäññe is very (recreated) oldschool and has a purposefully simple visual design, so don't expect much tech "blingbling", but quite good above all if you like good techbase secret hunts :) Only 9 maps so you won't get overwhelmed. Adonis: Escape from Urania is not all techbase (about half of the maps are mostly techbase approximately, the others have city and void themes with some hell elements) but it has a very "techy/sci-fi" atmosphere with some new buttons and in general very playful color combinations. It has also a lot of moving parts and geometry-based puzzles. It's a midsized megawad with 16 maps.

Maybe a controversial opinion, but anyway: Don't make MAP01/E1M1 too easy. I like it when it's slightly below the average difficulty of the set, just like in Plutonia, and when it really shows what you have to expect in later levels. It's a showcase and should not necessarily be beginner friendly if the rest of your mapset is not. I have often quit mapsets because MAP01 and MAP02 were boring, only to retake them later and realizing they weren't that bad. For example, I loved Ancient Aliens' cyberdemon in the first map, even if you don't have to fight it. Kama Sutra, Three's a Crowd and Adonis also have very good MAP01's. You can introduce breather maps later, then they're more necessary as otherwise players will become exhausted, above all in wads with more than 20 levels. For me it's also perfectly ok if you introduce breathers in the last stretch of the mapset, although then it's a good idea to offer something special, like a special kind of gameplay, puzzles, or a great atmosphere. That's just my 2 cents as a player who does map sometimes but isn't experienced at all.

For me, for example Urania did surpass it, because of its combination of difficulty with big maps and puzzles. Eternal Doom's maps are big but not that huge, most of them are not bigger than Doom2's Refueling Base, for example, and most (except some of the final stretch from 20 on) are also quite easy combat-wise. It has more difficult puzzles though which will "stretch" playtime of course, above all in maps like 20, 26, 31 and of course 30. Everything depends a bit on what the OP is searching for: playtime or simply size. :) To the Community Chest suggestions I would add the original Community Chest, above all for MAP29 (one of the biggest maps I ever played) but also for example for Use3D's and Sphagne's contributions.

Darkwave0000's maps in Speed of Doom. They have lot of stuff (up to almost 3000 monsters) but are quite spacious, so if you don't mind moving around a bit they're not difficult. Joshy's maps in the same mapset are another story and much meaner, although they're also not super-hard.

Interesting project. Could fail miserably, but could also go into interesting directions. Let's see :) (saying this as a Plutonia fanboi) If there is lack with MIDIs till the end of the year I can probably contribute some (perhaps up to 3-4). (For vanilla maps I'm too inexperienced and also probably too busy in the next months, have to finish my beta myfirstmap first ;) ). In general though I don't know if it's the best strategy to let people claim specific MIDI slots before they know the maps (with the exceptions of mappers composing the MIDI for their map). Perhaps it's best to simply distribute 32 slots and then select the MIDIs to match the atmosphere of the maps. (I don't know which are the best practices in such quite open CPs however).

Eviternity is from late 2018, so the 2019 "slot" is also still "free" ;-) I would nominate Lost Civilization for 2019. For 2020, Three's a Crowd. Both were Cacoward winners and exceptional wads with unique ideas. Heartland got much attention in 2021, I've still not played it but it would be an obvious candidate. Idem MyHouse 2023. A "slot" I personally can't give any specific nomination still would be 2022. Edited: 2022's nomination could go to NERF.

I very much agree with what Xaser wrote, old maps should not be generally evaluated on a "different scale" than modern content. There are definitely examples of good maps from the 90s/early 00s which still would be considered good today, which had fun, varied gameplay, or also good detailing. I've still not played STRAIN but in Requiem, Capellans "Rats in the walls" comes into my mind (which also was liked by MtPain27), or "The Devil's Coterie" from MM2 from the same author. Or "Resistance is Futile" from HR, AV's "Hillside Siege" or "Monster Mansion" from Eternal Doom (the latter only if you like puzzles). Only that filling a whole megawad with such good maps was a challenge in these years I think, partly to the still not-so-well-developed tool set. In my case also nostalgia only plays a minor role, because my Doom journey really started only about 2015 and the wad I'm most nostalgic about is probably Sunlust ... Little PS: For people interested in 90s stuff I would recommend not to start with Memento Mori (1). It's simply a wad specifically made for coop play, and while almost all maps are maxable in single player, you sometimes have to do weird stuff for that (e.g. not using a certain teleporter or cross a certain door until you clear another section ...). Memento Mori 2, Hell Revealed, Icarus and even Requiem are imo better candidates, in addition to Eternal Doom (imo the best 90s megawad) but only in the case you like adventure/puzzle gameplay.

Just for fun I played "Den of the Skull" (the map MtPain27 considers the worst) again, and yeah, it's quite a bad map. Some of the gimmicks are interesting (I somewhat liked the idea of the fight with the AV from the water pool, but at that moment you already probably have a BFG so it's very short) but most are simply "what?" material :). And it's an extreme example of the "room after room" approach where the different areas have almost nothing in common and aren't interconnected at all. In general I agree somewhat with MtPain27's review, Requiem is a very uneven wad, and I remember when I first played it I was a bit disappointed because despite of so many technical tricks it falls short of the better stuff from that era like MM2, Eternal Doom and even HR. But just two cents about the question "why people like 90s wads": For me it's the wild experimentalism many of these wads (let's say until the mid-2000s) show. Some of the Requiem maps definitely deliver in that point, as did the first Community Chests and the Memento Moris. Of course many of these experiments fail, see Den of the Skull or several Memento Mori (1) maps. But in general many of these maps are unpredictable and have also a wide variety of combat scenarios, even if some of them (shotgunning barons haha) don't match exactly the definition of "good combat".

I like it when the fights make me use features of the terrain (cover, barriers to impede you to simply circlestrafe, height differences, space restrictions) to win it and give you some room to explore how you can beat it. An example is the first big fight of Plutonia MAP10, when you cross the first barrier and monsters attack you from all sides (including the starting area) with different attacks. It is also easy enough to tackle it aggressively without too high skills but not totally trivial. Or also MSCP MAP06, or Adonis MAP01/Map05. I also like in general crowd management fights, but these are mostly part of slaughter-y gameplay. Another concept I very much like is high-density, chaotic incidental combat, where enemies are steadily appearing "randomly" from nowhere (UDINO E2M5, Urania MAP06).

Try UV Max. Current record is ~7 minutes (I think). So perhaps you missed a couple of shotgunners in a bathroom or so, they can be easy to overlook ... :)

Thanks. I'd have liked to experience also a bit of the combat, but this could be an idea to try, to play it as a "pure puzzle/horror map". It's strange though what Devalous wrote, as I experienced exactly the contrary, the map got progressively slower. Perhaps I did something I should't have done and a script "hangs" now? (Never mapped for (G)ZDoom so I've no idea if this is possible...) Anyway nevermind, I'll at least start again completely fresh, as with most puzzles solved it shouldn't take too long ...