The more restrictive definition is better. Anybody can run around creating any sort of crazy explanation for anything, from crop circles to the Bermuda Triangle. But mere statement doesn't come with any credibility. Who's to know what's credible? Do we teach crystal therapy in med school because some new-age people think it works? A scientific process is in place in order to separate the wheat from the chaff.
Even long-standing theories have had to be changed over time or partially ignored due to a constant attack from within the scientific community. We've punched holes in Newton's gravity theory, but we did it by offering a testable, falsifiable, predictive scientific theory that better explained the same phenomena.
Thus you can be sure than ideas that went through the scientific vetting process unscathed are at least pretty damned good explanations for what we see around us. The longer the theory's been up, the more credit it has.
The rest, ID included, is chaff.
Certainly there is something to be said for agreement in numbers when it comes to science. Doubtless there are cases where observers have been deluded and come up with highly improbable theories. But it seems that when consensus is given as a reason to accept or reject an arugment one might just as soon declare that objective reality is not dependant upon the number of people who accept or reject its claims.
ID and it opposite are two legitimate ways of doing science. Both begin with assumptions about the universe that are beyond proof, but neither is wholly unreasonable. I would posit that 99% of the world's population holds to one or the other of these assumptions, or a combination of the two.
If you can find an individual who seriously espouses spaghetti monster theory, please let me know.