Prefab Modular Speed Building First thoughts

[Fidcal]

I just picked up on what nbohr1more posted here about building from prefabs. I was just wondering how easy could we make it for novices to plug together modules to make new simple FMs. OK, likely they would lack originality and have lots of weaknesses but they might draw in more mappers who would not normally map and not all maps need to be highly advanced 'super maps.'

 

I can't work on this currently but we could spend a few months kicking ideas around so maybe one day...

 

Many of my first thoughts likely won't work. Let's see what if any will.

 

Performance:

• 
   
  
• Keep it simple so performance is less of an issue.
  
• Do not optimise. You only normally gain tiny fractions anyway except in complex situations and the prefabs rooms can be kept small and simple anyway. Just cover up ragged edges and junk waste. Keep focus entirely on speed building.
   
  

 

Start Map:

• 
   
  
• Needs a new simple one just for this probably just extended objectives, tools, ammo etc. Needs some thought.
  

 

Prefabs:

• 
   
  
• So I'm thinking first of customisable 'rooms'. A 'room' is defined as one closed volume with 'doorways' whether it be an alley, a mineshaft, whatever.
  
• Each room built to a standard wall thickness and grid.
  
• Each prefab labelled, eg, 'plug in compatible'.
  
• Standard texture types of same scale and alignment so can easily change.
  
• Each prefab has distinctily named layers for ease of re-selection and re-texturing.
  
• Each room has doorways at set points, eg, always say at positions 0.25, 0.5 075 along any wall. So to connect any two rooms you can align them easily.
  
• Doorways designed so they can be blocked off easily. For instance, if all doorways have a frame that protrudes 0.5 units then a 'blocker' prefab can be dropped over it.. This might for instance just be a a big vertical brush. Possibly leave the doorway inside for quickness but disable any working door.
  
• Stair room connectors and stair units.
  

AI and pathing:

• 
   
  
• A set of prefabs each with one AI type and say 8 paths + waits linked in a loop. Waits set at say min 10, max 30. Easy to change.
  
• Mapper changes AI skin & head for variety.
  
• Paths dragged to where needed. You can't really have too many; a one corridor guard can have 4 at each end. Paths are just discarded if not wanted by removing the AI's first target. Don't waste time deleting - you might change your mind later.
  
• Sitter prefabs with seat
  
• Sleeper prefabs with bed
  

 

Problems:

• 
   
  
• Dark Radiant's multiple item rotation bug is a big problem. Rotating a complete room prefab is hopeless. 4 sets of every prefab is not practical.
  

 

So anyway, the mapper just

• 
   
  
• Starts with the start map
  
• Inserts prefab rooms so each connects with another at doorways.
  
• Forms a basic structure of rooms.
  
• Retextures if needed.
  
• Adds or moves or changes any furniture or other models (or maybe does this later
  
• Inserts AI and adjusts
  
• Inserts loot
  
• Inserts custom objects
  
• Inserts and modifies just a few ultra simple pre-made generic readables.
  
• Modifies simple objectives.
  

[Sotha]

Sorry for being negative, but I don't see much point.

 

We already have the excellent tutorial, and I'd say learning the basics of mapping requires only the ability the read and the persistency to follow the tutorial through.

 

If this prefab building is intended for newbies, they will probably achieve a leak and things will not work.

 

It will be a lot of work to make the standard prefab rooms, but creating a room can be as simple as draw a brush and 'make room.'

 

If mappers only use prefab rooms, they will not learn to use the clipper or other standard mapping techniques.

 

Creating AI pathing nodes is so easy, that I do not see any idea in having prefab AI's with a bunch of path nodes: I believe it will be more confusing to import such a prefab, deselect it and then try to find the path nodes. All the mapper needs to do is to insert path_corner and a path_wait, clone them where needed and adjusting the spawnargs.

 

The end result would be unimaginative and 'standardized.' The mapper would not learn the important things in mapping. I just fail to see it worth the trouble of creating the prefab set.

 

I'd say we do not even want crappy speed-built maps, but rather well done carefully planned unique work. What if new tdm would-be-contributors play such missions as their first games and get a poor impression?

[Melan]

Heh, a bit like geomorph-based dungeon design. I am with Sotha, however; I'd rather people built simple maps, but their own maps. No problem with using prefabs, but it should not be exclusive. A map like Thief's Den or The Outpost is completely okay! I will play them and I will like them. Not everything has to be on the scale of St. Albans' Cathedral or Heart.

 

What could help, though, is more prefabs, especially if Serpentine completes his repository system. I have used others'prefabs before (Siyah Parsomen's and Jesps' houses, although I have modified them to fit the mission better), and have prefabs of mine I'll be happy to contribute -- things like fancier windows and arches, a big liquid tank that's rather neat, portcullis variants, towers, a transformer... people could accomplish a lot if there was a large library of them at their disposal.

[Sotha]

What could help, though, is more prefabs, especially if Serpentine completes his repository system. I have used others'prefabs before (Siyah Parsomen's and Jesps' houses, although I have modified them to fit the mission better), and have prefabs of mine I'll be happy to contribute -- things like fancier windows and arches, a big liquid tank that's rather neat, portcullis variants, towers, a transformer... people could accomplish a lot if there was a large library of them at their disposal.

 

After reading this I must add to my above comments that yes, more high quality prefabs would be better. All sorts of containers and stuff, like the safe in RTTC, for example.

[Tels]

Heh, a bit like geomorph-based dungeon design.

 

You laugh, but I already experimented with a javascript app that could export the map as .map file

 

Maybe I should pick that work up again. But then, SEED would be more important first

[Bikerdude]

I agree more with Fid's, new mappers could use these to get inspiration and use the contents of said prefabs. I have already done this very thing with my sewer tunnels from my cathedral map!

[Fidcal]

Perhaps I should have posted in the FM forum. It's not about what we mappers want but what players want. Do they want more small lower quality FMs from newcomers (instead of abandoned over-ambitious projects or even no FMs at all from the daunted.) It won't have much if any effect on existing experienced mappers. They either won't use the prefabs or more likely use them partly and/or heavily modify them. In addition, some of those novices will learn to expand and improve their efforts.

[Melan]

I think the issue is that it is not at all hard to create small, simple FMs. Thieves, Thief's Den, Special Delivery, The Parcel or The Outpost are all missions a newbie could create with a week or two of work. (And I am not knocking these missions. I love them for their simplicity and they are among my most replayed ones.) At least not substantially more difficult than stringing together some prefabs. DR also offers premade architecture - models (windows, railings, balconies and buildings) as well as prefabs (houses, stairs, railings, platforms and what have you), which is more than Dromed did. These building blocks alone could spice up even a simple map. What I am saying is that there is no need to go even further - rather, help should be given on a lower unit level (of individual building blocks instead of groups of building blocks).

 

[edit]I am actually thinking of taking a short break from my current large project and building a small mission like that just to make my point. [/edit]

Edited April 8, 2011 by Melan

[nbohr1more]

Couldn't you "plant" prefab buildings or rooms with SEED?

 

Then mapping would be as easy as drawing the grayscale image maps to tell SEED where to place buildings...?

 

All the placement and connective rules could be handled "programmatically"

 

(not to add more work for Tels... just food for thought for the future...)

 

Edited April 8, 2011 by nbohr1more