Elder-Basilisk said:I've done this too. It works OK as long as there are only a few personal agendas and they aren't too different from other members' of the party and aren't too earth-shattering. If one player wants to conquer the kingdom of Cimmeria, another player wants his character to join the assassin's guild and avenge his father's murder, another player wants to dethrone Orcus and take his place, and another player wants to become a great hero so that the bards sing tales of his praise and his beloved's father will consider him worthy to court his beloved, the campaign is destined to be incoherent and short-lived. If all (or even a significant number of) the PCs have different personal goals that are primary for them then the group won't share the common perspective necessary to act in concert. Rather than create one story, you will create four or six separate stories and give the DM a headache in the process. Players who have individual secondary goals for their characters are an asset to the campaign. More than a couple characters with individual primary goals for their characters, however, will derail a campaign. In that sense, their functionality in most games is contingent upon having a majority of players who just want to "play the adventure."
That's why we never start without a basic premise to the game anymore. It can be really hard for a DM to tie all the PCs together without railroading if there isn't some group cohesion to begin. So instead of everybody making whatever characters they want and the DM trying to find a way to fit them together. Character creation is for us now a group effort based on the camaign premise, such as: you're all members of the post-war baby boom in a small border town, or you're all members of an oddly blessed generation of a desert tribe, or you're all members of the Wolf clan or its demihuman neighbors.
Sometimes that's still not enough to get compatible goals for an adventuring group. So we've instituted the rule that you must have ties to at least two other party members. These can be friendship, blood relationship, rivalry, mentor-student, betrothal, whatever.
If you start out with a random assortment of PCs with incompatible goals you probably need such "railroading" tools as the prophecy or the guy hiring adventurers, to get a coherent game. (Is this really railroading?)
For D&D anyway. For our pulp superheroes game, everybody is pretty much on the same page of, the GM presents an adventure and the heroes go do it, because they're heroes. Which makes me consider that maybe railroading is built into some genres like pulp and classic superheroes. You can have side plots and stuff (our heroes certainly do), but it's kind of genre convention that the main plot hooks are really obvious and the heroes don't have much choice in whether to pursue them or not.
As for the bad guy escaping on a ship scenario, I don't consider that railroading, and not just because the players agreed to go to Khavayin beforehand. The DM had been specific in having the Marshals of the country tell us to keep our mission quiet, so as not to alert the villains to how close we actually are to foiling their plans and to keep the general populace from knowing that it's anything more than a localized bandit problem so there won't be a general panic as rumors of war and the resurrection of long dead evil tyrants spread.
In our relatively low-magic campign world a giant winged magical beast (that's never been seen in the area before) and a person riding a flying arrow, both carrying people with them flying out to a ship and forcing it to return to port, is not exactly inconspicuous and would certainly alert our enemies to the fact that one of their ploys has failed. Plus we had no real proof that the bad guy on the ship had done anything wrong. So all the Marshals had to do to keep the PCs from doing the stupid thing was to refuse to give them an arrest warrant. I don't think that having government officials decide not to help the PCs in a hairbrained scheme is railroading, it's realistic.