A Puzzling Adventure.
(Or, Myst-like Puzzle Tips)
For games like Myst, the earlier difference I pointed out between a puzzle and a mechanic, melts into bad terminology - because puzzles ARE the mechanics. So here, I'm going to start off with a look at the various things that make up a puzzle in a Myst-like, and what we can do with them.
The Anatomy Of A Puzzle
Geography - Actually being in the right place in your world to find the puzzle. You can hide it in a cupboard somewhere, under a table, or on the scales of a fish. 0bfu5cat1ng*/hiding can work with any puzzle, but it is not particularly rewarding unless the player is given clues that a puzzle actually exists.
Discovery - Working out that what you're faced with IS a puzzle. You might be stood next to the lift buttons, but you may not realise there's a puzzle to them, and just think they're broken. Harder to hide from the player, and leads to quite a lot of confusion - but can be rewarding. This can be used as an alternative form of 0bfu5c4t1on - hiding in plain sight.
Goal Setting - So, it's a puzzle, and we've found it. Now - what's the goal? This should be, in most puzzles, quite obvious. Usually you want to let the player know "If you win, you get this:" whether that is access to a new room, the puzzle piece they want or the battery to save their robot's life. Puzzling without a goal is not usually fun.
Rule Learning - Ok, we know what we have to do - what tools do we have at our disposal? Working out the rules of the puzzle (and they all have rules) can often be one of the more fun parts to a puzzle. Finding out that using blue dust makes the plants grow bigger, but that green dust has random effects, is where your classic "Explorer" type gamers will thank you. Make it consistant and they'll be happy. This may involve the 'meat' of your puzzle time.
Solving - You've found the puzzle, you know what you want to do - and you know how you can do it. This last part is the actual solving and completion of the puzzle - putting everything else into place. This will probably be quite a brief stage, but it should be satisfying. Give a good payoff - win noises, graphical effects or compliments from annoying alien sidekicks are good here.
As you can see, there are ways to string out a relatively simple puzzle - either by hiding it or making the goal less obvious. I'd advise that unless you've decided that the style of your game is to have 'Hiding Puzzles' as a mechanic, then try to avoid arbitrarily putting wallsafes behind washingmachines. Which brings me to puzzle wrappers in context.
Puzzle Wrapping - Making Context Work
Wrapping is taking the basic mechanics of your puzzle, and placing it in the world in such a way that it makes sense in the world. Now, if you are the game designer, you can make this easier on yourself. Setting a game in a dreamworld, or a fantastically surreal environment will let you 'get away' with lots of crazy puzzles - but at the expense of the players suspension of disbelief. Here's some examples of a simple puzzle, and the contexts it could be put in:
Fantasy Odd One Out: (easy - there's magic to excuse stuff)
A mighty pedestal rotates before you in the great and mysterious tomb - you are faced with a collection of brightly coloured scarab pendants and the hiroplyphics read "Find the Infiltrator, it will be your friend. Break up a familiy, and it will be your end!" One of the scarabs is the odd one out (with colours and patterns - think Mensa test).
Sci-Fi Odd One Out: (mysterious science can be as handy as magic)
The door is locked and the alien script is unintelligable to you - every time you press one of the buttons, they all change. Slowly you realise that out of the 10 there, one is always the same - pressing it, and the door opens to the engine chamber...
Modern Odd One Out: (hard to think of)
The man eyes you suspiciously, shows you the 'quality merchandise' that he's obviously stolen. Your informant told you there'd be one piece that was real - and the rest are just fakes. Better pick the right one if you're to 'prove' your worth as a fence.
As you can see - setting can be very important for your Myst-like adventure games. Myst-likes are traditionally a bit 'gamey' - and players often forgive and almost expect that situations that would just seem a bit odd arise in their games. They do, however, expect certain rules to be adhered to, and if you break those rules, they will be annoyed.
General Vs Topical
Puzzles are a test, and just as if you were suddenly tested on Rocket Science at your French Speaking Test, you'd be very annoyed - so too players have an expectation on what you will 'test' them on. For example:
Bad Puzzle: Odd One Out - 4, 8, 12, 15, 16, 23, 42
Answer - 12, the rest are the LOST numbers.
Good Puzzle: Odd One Out - 4, 8, 15, 64, 128, 512
Answer - 15, the rest are powers of two.
Here, the player understands that basic mathematics is something that they may be tested on - not topical knowledge of a TV program. Of course, if your game is about LOST, or TV trivia - then the above puzzles should be reversed.
Some of the things that your average puzzle player can expect:
Basic mathematics: Adding, subtraction, multiplication, division, number sequences, etc.
Word puzzles: Code cracking, mixed up sentences, basic reading, etc.
Simple logic: Odd ones out, matching, AND/OR logic (sufficiently hidden), group logic, building, combinations etc.
You may choose to add to this depending on your setting and/or gametype, like in the LOST example above.
Variable Difficulties
You can make later playtesting and balancing much easier on yourself at the puzzle choosing stage by picking puzzles that can be balanced easily. For example - a timed 'wobbly wire' task can be made more or less wobbly, and the timer changed, and 'lives' given. Whereas something like an odd one out puzzle is either hard or easy, and balancing it later is going to be a pain to do. Try to give yourself several different balancing variables that you can tweak later.
Tested on Animals
Finally though, please remember to test your puzzles on the uninitiated, because it's all too easy to forget what it's like to see it 'fresh'.
As an example, on a game I was working on once, there was a puzzle with four switches in a row. If you pressed a switch, it and it's two neighbours would toggle their lit state and the aim was to get them all ON. Now, that sounds like a very simple puzzle - and indeed it was set up so that you could complete it in three presses - just three. There were people who would look at it, blank out and start randomly pressing buttons in the hope that they'd chance apon the solution.
Blindingly simple - if you know what to do.
I hope that helps. Thanks again for the nice comments about the other tips so far - you've actually got me considering writing a short book/let on various game/level design aspects. We'll see though - I like posting in this thread too
*For some reason, the word "ob-fu-sc-at-e" causes any post with it in to be unable to be posted, this confuses me greatly.