Splitting larger maps

[Spectre01]

At which point is it better to make one map into several smaller maps while making a wad? It's easy to get carried away and just keep making sectors and adding dudes to fight, but when is a good time to just call it and end the level? Is there like a general principle of good mapping when you should stop and continue from a new map?

[scifista42]

rileymartin said:

Is there like a general principle of good mapping when you should stop and continue from a new map?

In my opinion, no, there isn't any qualitative principle, as even really long maps have their appeal. Only technical principles: When you run out of all types of keys and locked doors to use, when you come close to exceeding limits of the map format or of the target engine, etc.

[baja blast rd.]

rileymartin said:

At which point is it better to make one map into several smaller maps while making a wad? It's easy to get carried away and just keep making sectors and adding dudes to fight, but when is a good time to just call it and end the level? Is there like a general principle of good mapping when you should stop and continue from a new map?

I'd nearly always prefer one quality large map to a similar amount of gameplay parceled over multiple small or medium-sized maps. Exceptions are when small maps are sort of the point of the project (e.g. Scythe clones) or with difficult maps. Ribbiks's JQ2 is an example of the first, and, uh, Ribbiks's, uh, most things are an example of the second.

Good larger maps tend to have some sort of narrative arc to them. This happens even when one isn't intended, just due to the pacing and travel from one location to another and stuff. For example: start in a rocky cavern complex -> reach a techbase and probe into its depths -> board a spaceship to visit an alien planet -> battle through the alien planet -> have a climactic fight with flying cyberdemons at the heart of the planet. Typical stuff. Anyway, if you find yourself extending the narrative in a coherent way, maybe you can keep going until you reach the map's limits. Why not.

But maybe you might find that there is no logical way to continue that narrative (meaning you'll just be designing additional fights and areas with the sole purpose of extending the map's length)? Maybe the narrative when continued would become too disjointed, one distinct chapter after another distinct chapter after another distinct chapter, without too much spatial interaction. Then that could be a good time to split the map!

[40oz]

Mapping is pretty open-ended so it helps to set goals for yourself. The question is pretty open ended as well because a map that is physically huge (as in uses giant open areas) can take 40 seconds to beat if the monsters are easily avoidable and there aren't many things to interact with like riding elevators, doors, mazes to navigate, switches to press or keys to collect.

Unlike rdwpa I personally prefer shorter maps to larger ones. When I play doom I record demos of me playing a map from start to finish so dying on a large map is a real bummer and a tremendous waste of progress. I like shorter maps because the appeal of trying to record a better demo is more apparent where a large map is a pretty tremendous timesink some times and staying alert, precise, and consistent throughout the course of a half hour or more is draining for me.

My basic formula for mapping is usually

Start area
Objective 1
Objective 2
Final Battle
Exit

The objectives are usually something pretty simple like getting a keycard, or pressing a switch that opens a door, raises a pit, etc. The final battle is usually something pretty climactic and fun that is difficult relative to the rest of the map.

If you have a large map that you think you are going to need to seperate, don't be coy about deleting areas to make new ones in order to wrap the map(s) up nicely. You'll spend a lot more time and stress trying to make something "work" than to delete a part of it and start over with a different goal in mind, and in many cases it results in a map that is much better than you expected.

[Fonze]

scifista42 said:

when you come close to exceeding limits of the map format or of the target engine, etc.

<3

Or when you put 50 revs in a small, dark, crowded room with trees everywhere and 3 keys scattered around.

[TimeOfDeath]

When my objective is to make a large map, I map until doom builder just takes too long to build nodes when going into 3d mode or saving on my 15 year old pc. That's why my maps don't go over 25k sidedefs and 10k monsters.

[Quineotio]

I think there's a certain pacing and flow to a level that dictates how long it should be. If you pay attention when designing the map you can create an interesting story arc for a level of pretty much any length. What are you trying to make the player feel?

[Dragonfly]

scifista42 said:

When you come close to exceeding limits of the map format or of the target engine, etc.

On this topic alone it depends on your desired audience size - if you're trying to appeal to the masses you want to avoid a map large enough to make older computer systems lag.

I personally have a preference for smaller maps in a megawad, gives me something that feel like distinct 'save the game and go take a break' markers; a long map I'd naturally try to play from start to finish.

[scifista42]

Dragonfly said:

if you're trying to appeal to the masses you want to avoid a map large enough to make older computer systems lag...

...or to make Eternal Doom haters quit.

[Spectre01]

Isn't performance more related to the geometric complexity (number of polygons) in the visible area, rather than general map volume? I haven't tested it, but it seems like it's possible to make a medium sized room run like crap by spamming very smooth circular pillars.

[scifista42]

rileymartin said:

Isn't performance more related to the geometric complexity (number of polygons) in the visible area, rather than general map volume?

This is true for rendering performance. But overall performance is EVEN MORE related to the number of active monsters (including those in teleport closets) and other so-called "thinkers" (including idling monsters, flying projectiles, animated decorations, moving sectors, blinking lights...) within the whole map. Having too many sectors within one map can also slow down performance of certain linedef actions. Overall performance can also be affected by the sheer amount of RAM that the engine uses during playing the map, and of course, bigger maps require more RAM.