"Testable/Untestable": I think all traps should be testable, although the players might not know how to test the trap, or the characters might not have the tools or materials necessary to test.
"Clued/Clueless": I likewise think that all traps should be clued. Again, the characters might not have the ability to recognize the clue (warning in an unknown language.) The clue might not even be in the location of the trap. The clue might not be obvious: the corridor ahead might narrow for 20 to 30 feet, then open up again, and that might be the only clue that something's different here and you should proceed with caution. But, if there are non-magical flame shooters in niches in the walls of the narrow section, they probably have "pilot lights", which infravision (if you're using it) should detect, and if there's a magical trap, Detect Magic should detect it.
Granted, those are tests, but they're tests that are prompted by the clue of the narrow section of corridor, and they are tests that in turn reveal further clues. Which kind of points to "testable/clued" as being a single merged criterion, with perhaps different divisions of tests (easily tested with natural abilities, testable only by expert, testable only with special tool) and also different levels of clues as the result, ranging from generic warning through exact nature of the danger.
The other two criteria Zak proposes are useful, but I think he's leaving out some functional criteria I hinted at in an older post I did on tricks and traps. These days, I'd tweak a couple of the definitions I gave earlier:
- Hazards harm or limit options for those who interact with them; they are paired with Resources, which provide additional options instead.
- Traps are triggers that produce Hazards as a result.
- Puzzles are triggers that produce Resources as a result.
- Tricks are triggers linked to a lure/bait promising a Resource, but delivering a Hazard as a result instead.
I might try breaking out some trap, puzzle, and trick types based on some of these criteria in the future. But right now, I just realized I haven't worked on tomorrow's geomorph yet.