Since most people seemed to ignore my rather lengthy reply, I thought I'd cut to the more interesting half of my post, and make it a discussion unto itself.
There are three major problems with graphic adventure games.
1. Uneven puzzle design. Even the best of them can get tricky. Grim Fandango threw some real curve-balls at me, as did The Longest Journey, Syberia, Day of the Tentacle.... they were all plagued by certain puzzles that were just bad ideas.
It's been said that a writer should assume nothing about what their audience knows, but to assume the audience has some intelligence. The same principle applies to graphic adventure games. Do not assume your player immediately understands what something does, the context of a situation or that they know what's happening. This doesn't mean you have to tell them outright; if you want, feel free to puzzle it out. However, you better be sure that every single person, with some thought, will be able to understand what an item does or what a person wants.
My single biggest pet peeve is to attempt to solve a puzzle multiple ways, fail, then check a walk-through only to find out that I had attempted something similar. Grim Fandango had this puzzle where you had to pop a balloon in front of these birds, but if you tried simply using the balloon, you failed. This ruled out, in my mind, the possibility of the balloon being at the heart of the solution, even though it was. It's possible to allow this sort of thing to happen, as long as it's implied that you're on the right track, and should be doing something similar.
Getting stuck will happen even if you design intelligently. Some people just won't get the puzzle. For them, provide an out. Make at least one more puzzle solution that will get them to the same place. One of the things I like about graphic adventure games, actually, is that there is only one solution. That said, I don't like it so much when I'm stuck, and I have no idea what to do. Have an alternative puzzle or two that will show up when the player gets stuck.
2. Graphic adventure games are so alive... until you get stuck. How many times have I been punished for not knowing what to do next? New areas stop popping up. Dialogue options are exhausted. Every item in your inventory seems to be entirely useless. This has to stop. You can do all you want to help the player get through the puzzles, but at some point, they're going to get stuck anyway. Might as well give them something to do in the meantime.
When the game stops moving forward, get it moving sideways. Open up new dialogue options, create little cut-scenes, allow the player to go on a vacation of sorts from the real game, and visit some sort of eccentric and fascinating location totally aside from the story. Needless to say, none of this should be important to the main game, but it will at least keep the main game a game, rather than a puzzle come to an absolute standstill.
Misery loves company. If I'm frustrated, I want my main character to be frustrated too. It makes sense; after all, he's the one who can't move forward until he figures things out. If anything, the main character should be more frustrated than the player. There should be dialogue options opening up whenever the player is stuck that reflect that. In fact, this is a good place to integrate hints for a solution (though that won't work for every game).
Integrating character into solution is a good idea. I don't get the sense that graphic adventure designers ever ask themselves, â€œwhen this character reaches an impasse, what would he/she do?â€. It's a good way to integrate the alternate solutions I suggested earlier. If a character is a hitman, then why wouldn't he try force? If the character is a diplomat, wouldn't she try to talk her way out of her problems? It's a good way to ease the obtuse, abstract nature of some of the puzzles without ever breaking out of the reality of the circumstances.
3. For a genre called graphic adventure, there doesn't seem to be much adventure. This one drives me crazy. There is a large element of exploration, but it's not properly handled. There seems to be a real lack of proper pacing in this area.
Most adventure games show their hand too early, giving you access to most areas quickly, then forcing you to retread them repeatedly. Come on, the pacing has to be better than that. Open up a new area every couple puzzle solutions, consistently reward players with new areas. Not only will this better distribute the novelty of exploration, but it will better motivate players to solve puzzles. Just be careful that you don't give the player access to too many different areas. If you do, you exponentially increase the number of variables, and therefore possible puzzle solutions, making the design overwhelming. Open up new areas gradually, close off old areas gradually.
It'd also be nice to be rewarded more often for going off the beaten path. If the player tries to go someplace out of the way, or do something rather odd, try to give them some kind of reward. Some kind of gag, a pretty animation, many a new little area, just a little perk of some kind. Just don't let these things become possible puzzle solutions, or you'll unnecessarily complicate things.
As a personal aside, I'd also like to see more graphic adventure games explore time as a puzzle element. In some games, it adds an intriguing layer of depth and urgency, having to consider both what to do, and when to do it.
The graphic adventure that has true longevity is the one with terrific writing and storytelling depth. Accomplish that, and most will overlook the faults.