While I understand your definition, it is NOT the typical definition used by most people.
Show me your market data. I think my definition is very good for describing most of the things people describe as railoarding, and very few things people describe as not-railroading. And I've been at this a while, so "people" involves dozens of acquaintances, hundreds of people on the Internet, and any number of writers of article. While it's possible that your experience is very different than mine has bene, my inclination is to suggest you are arguing from assertion and trying to make your facts more different than mine than they probably. However, if you can find some good evidence that railroading has a well-codified, generally accepted meaning, and I am using it in a different way, please share links.
Scenarios, particularly published ones, are regularly referred to as railroads, analyzed for railroad-i-ness, and generally discussed as if scenario design were a means by which the players can be placed on rails.
So what? Simon Cowell says people's performances are "horrible," but I've never seen anyone die from listening to them.
Which makes sense, to me. If you've got DM A who makes up explanations why you can't do what you want, and DM B who anticipates you trying to do something and makes up explanations in advance for why you can't, which he then writes into the scenario, well, chances are that the player's experience in both situations will be nearly identical. The only difference is that GM B might come up with better explanations, or might accidentally miss something he intended to railroad.
I am not quite what you are saying here, but it sounds like you are saying that planning a scenario to take into player choice is equivalent to simply pulling stuff out of your portable hole. If so, I disagree entirely and completely.
Your opinion seems to come down to an acceptance of railroading if the presence of the rails is believably explained in-game (you also seem to add a presumption that the only way to believably explain their presence in game is to do so in advance during setting design, this may or may not be the case for a good impromptu DM, but the overall issue, believability, is core in what you've written so far).
If you have gotten that impression, I am sorry I did not cleary communicate my opinion to you, because that is not at all what I inteded to convey. I don't care if something is believable, although that would be nice. What I care about is that it things are logical and follow from a meanginful way from the GM's design and the player's actual choices. Whether the overall result is "believable" or not is something else. You could have a game session that consisted entirely of the GM narrating Caesar's war against the Gauls, which would be very believable but absolutely impervious to player choice.
Which gets back to my pet issue- railroading is really just restriction on player freedom, plus a negative connotation.
Railroading: restrictions on player freedom that the speaker doesn't feel are reasonable.
That is not how I am defining it. My definition:
Railroading: GM actions that negate meaningful choices simply because of what was chosen.