Scientific hypothesis versus speculation

I have a different definition of "speculation" from what I often see. Scientists think about known phenomena and known data that puzzles them and try to think of ways to fit theory into the observations. They then come up with ideas, and make predictions of what else we should see if these ideas are correct. They then do experiments to see if the predictions of the idea are true, and if so they incorporate the idea into existing theory. This is usually called a "hypothesis."

"Speculation," on the other hand, are ideas for which no experimentally testable predictions can be thought of -- either because we don't have the needed technology or because it is too unpredictable ("woo") to be reliably experimented on.

String theory, for example, is often labeled speculation for the reason that most of its predictions are effectively untestable -- to test them would require technology way beyond anything we can even imagine having. Hence string theorists are constantly looking for things string theory predicts that might actually be testable. An example is, if gravity waves exist, and if we can measure their polarization, this could test the reality of string theory.

The thing that impresses me about string theory, as compared to a lot of speculation you see, is it is so consistent mathematically -- the string theorists call it "beauty." I know that is what keeps them working on it in spite of the criticism. Maybe that is enough to convert it from speculation to hypothesis -- the boundary between the two is fuzzy.