Dungeon Design Methodology

Des · Post by **Des** » Thu Dec 23, 2004 10:19 pm

I notice a number of people are currently working on new dungeons, and of course quite a few have actually completed one or more (respect! respect!). So far I've fannied around a bit in CSBuild and made a couple of little tiny test dungeons, played with portraits, and I came up with a few quite promising plot ideas (on the bog

) a while ago.

So the question is, how does one go about creating a full-size adventure? Being a geek I've been thinking about the two main computer system design methodologies - "Waterfall" (formal design) and "Iterative" (e.g. eXtreme Programming).

So I guess in a Waterfall-designed dungeon you'd start by writing a storyboard or narrative. Then plan the dungeon out on a paper "specification". Then decide upon characters, puzzles and monsters, and add them to your spec. Then build the whole thing and test it.

With an XP-designed dungeon you'd open up the editor program straight away, create the characters and the first couple of levels, test and tweak them, add another level, retest, etc. etc. The "plot" and the overall structure and even goal(s) of the dungeon would be constantly open to change according to whatever you think best at the time.

Perhaps a compromise approach is best - do a fairly rough design of the whole thing, then wing it thereafter?

beowuuf · Post by **beowuuf** » Thu Dec 23, 2004 11:00 pm

I prefer to act with an idea of the whole first, and find that there is much creativity in then fillign in the blanks, shaping and subtly alterign and augmentign the idea while still retaining the overall structure

This is in general, but it also applies to dungeons, not that I have got any further than most who have unfinished ones lying around. The trouble is that having an idea of the whole means many uncorrelated ideas are developed, whereas at least the iterative approach formalises thinking and structure so you are always allowing somehtign to grow normally, logically and connected. Not having ideas expanding and exanding without any limits nor connections.

For example - temple dungeon. Idea was to have a furnished dungeon, only a few levels, with a story of a truely good temple that was won from evil, and that therefore had darker areas blocked awaiting the wrong people to dig them up. The basic story came very quickly to mne, so
I had the idea of the rough areas it would possess, and opened up the editor fairly quickly. The dungoen design then flowed as i created these areas, and realised the logic of having other features, or just decided to place items and then weave them into the story afterwards.
It worked out quite well, just never managed to stick with it long enough. But i have five levels, of varying complexity, with a common thread and logic, even though each has expanded in conception from the original idea. For example, and abandoned evil acolyte area - was to simply be an abandoned area with atmosphere and 'find the object'...but a twisted idea allowed me to show why the evil temple had to be stopped when it did, and also provide a whole new lease of life on the level and make it more active.

sucinum · Post by **sucinum** » Fri Dec 24, 2004 1:44 am

1. get an idea what the dungeon should be about. you need a "theme" or some kind of "red thread".
2. a rough sketch of what the dungeon should consist of
3. balancing with items and monster types
4. some traps and a general layout (still on paper here)
5. start the editor and begin to create it, fill gaps with some stuff
6. playtesting
7. bugfixing
8. adding story elements and hints
9. prerelease to get comments
10. bugfixing and maybe rebalancing
11. done if you are lucky

Post by **PaulH** » Fri Dec 24, 2004 3:20 pm

I just made ToC up as I went along!

My new dungeon follows sucinums list above roughly, though I got the editor out early on. New ideas are constantly being added.

Post by **Zyx** » Fri Dec 24, 2004 6:51 pm

1. Find a great idea. Let it boil down a week or two.
2. Then try to remember it and take some notes this time.
3. Some days later start designing your dungeon according to the few notes you manage to decipher.
4. Realise it won't work this way and betray your ideas.
5. Playtest.
6. Fix bugs, balance difficulty and other tweaks.
7. Just when you're about to release the dungeon, Paul comes up with new features.
8. If you get inspired with new ideas, go to 4. Otherwise go to 9.
9. Try to use the DSA (Designer, Surrender it All) for some months.
10. Renounce, release, avoiding 7 with luck.
11. Optionnaly, find a name for the dungeon to save the apparences and give coherence where there was none.

Des · Post by **Des** » Fri Dec 24, 2004 11:52 pm

Thanks guys, your replies have helped to focus my thoughts - though I'm not focussing too well at the moment due to Christmas Eve drinkies down the pub (Descartes wasn't there).

Sucinum makes the dungeon creation process sound easy, and Zyx makes it sound really hard - guess it depends to some extent on how much customisation you go for. DSAs look quite tricky but very powerful. I've had a look at rain's graphics.dat tools (DMExtract and CSBEdit) and whilst they can do a lot, his new one ADGE looks to be the one to go for (when completed of course).

Whilst I doubt if I have the time or the patience for something as mega as Conflux 2, I think some customisation is a must. People will still play "vanilla" dungeons if they are good enough, but customisations pertinent to the theme of the adventure add an extra dimension.

sucinum · Post by **sucinum** » Mon Dec 27, 2004 4:31 pm

Des wrote:Sucinum makes the dungeon creation process sound easy, and Zyx makes it sound really hard

compare conflux II to imprisoned again and you know why - but even a vanilla dungeon like mine is a lot of work and it took me weeks of planning and it was built with dmute, which is less comfortable than csbuild.

beowuuf · Post by **beowuuf** » Mon Dec 27, 2004 6:15 pm

as always it's only 10% inspiration, 90% persperation...
Which is why any idea of mine are never finished!

Post by **PaulH** » Mon Dec 27, 2004 9:04 pm

Yes, I agree. It is playtesting that takes most of the time up. Just one little actuator out of place and thats it, the dungeon might not work. And I do this regularly, load up my fiendish puzzle and get to the first triple locked door only to find I have left a switch on it. Or left a playtest item in. The editor, while very good, is easy to make silly mistakes with.

Take today. After designing a nasty puzzle, playtesting realised I had forgotten to 'unswitch' the very first door. Annoying. Edit, reload and then find I forgot to set a pressure pad to 'clear'. Bugger, edit, reload. Then I realised the pad didn't target the correct face of a wall. Edit, reload. Now the teleporter doesn't teleport monsters. Or make a noise. Edit, scream, reload. Aha, getting there. Then find I forgot to put giggler in actuator room, and the key you gain doesn't fit the lock, the hard monster only has one hitpoint and the fireball shooter is firing zo spells... Turn computer off, cry.

Des · Post by **Des** » Tue Dec 28, 2004 12:09 am

I actually find all the fiddling about in CSBuild strangely addictive, which is a good sign, but I can imagine it does wear off and turn to frustration after a while.

I've started doing some rough designs on paper (reminiscent of when I used to play AD&D in the 80s) and it currently looks like *gulp* 16 levels, so I should be finished sometime in 2007. Unless of course I can do something about annoying distractions such as going to work and sleeping.

Alternatively, a smaller first attempt might be a better approach. I guess it all depends on whether or not dungeon creation is like making a rock album or writing a novel. If an album, a group traditionally produces a rough and spunky first record, slowing down when they get to that "difficult third dungeon". If a book, they do say everyone has one good dungeon in them....

Finally, a wee question - is there a good reason why most custom dungeons are done with CSBWin rather than RTC?

beowuuf · Post by **beowuuf** » Tue Dec 28, 2004 12:12 am

more people still like to play as close to the original engine as possible, plus at the moment RTC's format is unstable, whereas CSBwin will never change in a way that negates old dungeons

Post by **PaulH** » Tue Dec 28, 2004 12:18 am

Work can't be avoided. Sleep, however, can.

16 levels, 1st dungeon? I bow to thee, my Lord! The CSBWin editor is remarkably easy to use but remarkably easy to make a total balls of (see above). Somehow, DMute made you concentrate more on what you were doing, but it was a slow start into the art of dungeon building. RTC is great. It just lacks 'something'. It has better graphical support, a large fan base, 'modernisation', but Paul Stevens' conversion has captured EXACTLY that DM feeling.

My current creation is a CSBuild created effort, as the last three were mostly Dmute. I hope this will add a different dimension as it opens up the databases.

Keep on building, take time of work, I'll play it!

Des · Post by **Des** » Sat Jan 08, 2005 1:24 am

I've now done quite a bit of work on my dungeon, and it does seem that the initial design is the easy bit. I wrote a precis with story notes and key texts, plus a sketched layout, for the important parts of each level. The original version required 16 levels, but when I started actually building the thing I wimped out and amended the design to 10 levels, otherwise I'd never finish the bugger within the 1 year I've allowed myself (I'd really like to finish it in 3 or 4 months if possible to avoid it going "stale"). I do think there is no point going further than this at the start as whilst you are fleshing things out you will get new ideas which you can incorporate.

Testing is *definitely* the main bugbear, along with setting the right difficulty level. I'll have to deal with those in earnest later, for now I'll enjoy the initial shiny happy creative phase.

Des · Post by **Des** » Sat Mar 05, 2005 10:31 pm

Playtesting of Angel's Egg is sort of going well and sort of isn't. The problem seems to be how long do you do it yourself before you release the beta version? The dungeon is non-linear so there are lots of permutations. I also have quite a few "peril avoidance" puzzles where if you do it wrong the party ends up in the doo-doos. So I have to try to forget that I know what to do and deliberately make "likely errors".

So far I am about two thirds of the way through with my first party, all resurrected: Arnold Schwarzenegger, Margaret Thatcher, Bill Gates and Gandhi. I haven't had to restart once though there have been several trips to Vi Altars. In one battle, the party were completely surrounded by Vexirks (avoidable, but I chose not to) and in the end only Arnie was left standing with about 20 health.

By having custom characters I seem to have made a rod for my own back, as I really should try them all out. Of the first party, Gandhi is probably a bit too good. His high wisdom and mana were supposed to be counteracted by low health, strength, dex etc. but after beating up a few screamers he's OK on those and as long as he is well fed can cast priest spells as needed with alacrity. On the other hand the dunge does pit you against nasty monsters quite early on so maybe a roochy priest is essential.

So I reckon about another 2-3 weeks of testing before the beta. Any testing tips from more experienced designers welcome.

beowuuf · Post by **beowuuf** » Sat Mar 05, 2005 11:00 pm

tip one would definitely have been to take the weakest starting charatcers you could to test your own dungeon with rather than custom characters...but otherwise you sound like you have it all in hand!

really, players are the bets testers, they will tell you when somethign is too hard - they can also poitn out an unobvious puzzle you thoguht was really simple : )