I've considered a few more things worth tacking on to my previous post about calling for rolls. Hey, if I'm lucky, this could turn into an entire series of posts.
The Knowledge Dump
Many games have knowledge skills; for example, in D&D 4th edition, you have History, Religion, Arcana, Dungeoneering, and Streetwise. It can be tempting to put game information behind these skill rolls in order to get some use out of them. I'm inclined to think these days, however, that that is a bad idea.
Information about the game world touches on a lot of elements. It gives depth to the setting, showing off the skills of the GM (or whoever wrote the material being used). It provides players tools for interacting with the setting and understanding the nuances of how things work. It can even be critical for short-term goals, such as deciphering the instructions to a puzzle in a dungeon, or discovering the location of a mission-critical goal.
So why would putting any of that behind a skill roll be a good idea, given the chance of player failure? If it's mission-critical stuff, then we're back to the problems in my first post on this topic. If it isn't mission-critical, then risking the players not rolling high enough does not provide anything. Your world will seem flatter and less interesting for the lack of information.
However, if only ancillary information is put behind knowledge skill rolls, some players may be less likely to take them, since they don't provide anything "necessary."
Success, but . . .
The Dresden Files RPG describes this strategy as, "Success, but . . ." and I really like it. That scenario from before, with the locked door? Instead of interpreting failure as the players not getting through the door, the players get through the door but make so much noise that they attract a guard's attention. I like this, as it makes failure much more interesting without letting the game grind to a stop while your players make the same roll over and over until the dice-gods relent.
Here's an example to put it together: Let's say your players are looking for the location of a mages' tower long lost to common knowledge. They rightly suspect that the historical archives in the capital city will have the information they seek. The player with the best score in History would want to go to the archives to study and find this information.
In the past, I might have had the player roll, with a bonus to the roll for using all the reference material of the library. This still invites the possibility of failure, unless I make the bonus so outlandish that the information was coming one way or another. I don't think this is a very interesting way to do things.
There needs to be a way to give the information on a failed roll, but offer up a minor penalty to the players anyhow. You could say that the research takes longer, depending on how badly the roll failed, but that can be game dependent. In some games, the passage of extra time might not mean much to the players, nor to the GM. Instead, let's say that failure results in needing the staff of the archives to help the players find the information they seek. Of course, the staff of the archives doesn't work for free, so it'll cost the players some amount of gold to pay for the services. Even better, let's say that instead of paying in gold, the players could offer to run "errands" for the archivists, retrieving specific historical texts from the locations they visit, including that mages' tower.
Failure has been made more interesting, but what about success? If the players are getting the information either way, success needs to be more than just, "You don't get punished." Perhaps, in finding the location of the mages' tower, the players also find some other interesting bit of information. A treasure depot, perhaps? Or maybe they learn about some of the defenses the mages' tower used to have, giving them a heads up on what to expect when they get there. Maybe there are monsters in the tower now, but the excellent research gives them an advantage in combat when they fight the monsters. In a completely different route, you could say that the player's research leads him to translate a section the archivists had been unable to interpret previously, so they pay the players for their help.
It seems like it would be difficult to make every skill roll this dynamic and interesting; not every player action is quite so complicated, after all. Still, as far as skill rolls are concerned, I think a little extra planning and the right frame of mind can do a lot to add to a game.