I'm okay with traps if they're good traps.
My first thought too.
Traps have their place, but the classic way of using them is just a mess. Most commonly, they are either a skill or health tax to pass, used in their crudest form, they provide role protection for the trapsmith classes. They are placed more or less everywhere, and old school DMs seem to delight in using them as a surprise screw you whenever paranoia isn't quite paranoid enough.
First, the trap's primary contribution to the game. Does it make things more interesting? Traps should add tension or augment plot, thus placing them haphazardly(like in every other corridor) should be avoided, when all it does is add time while they disarm, overload or otherwise bypass it to advance.
Second, consider the nature of the trap. Where the plot or challenge requires a trap, a trap
needs not be artificial. A half collapsed building is a natural trap, where a bad trigger causes more of it to crumble. Avoidable environmental hazards are just alternate traps, using different skills to detect and disarm. A factory or alchemy lab alike are natural traps, only awaiting misuse of their equipment to go off. There is basically no difference between a natural methane pocket's reaction to the party's torch and a fireball trap. Exploit this and make them natural to the environment.
Third, consider placement. Any intentional trap would have arming and disarming features for the rightful users. They wouldn't be placed in high traffic areas because some doofus can and WILL set it off by accident without a safety or a nonlethal outcome. Such traps work well for limited access areas, tombs generally aren't meant to be opened again, making them heavily trapped has no drawbacks, but it is more likely to experience traps planted in personal chambers and containers, or tactically set up to provide an advantage in a conflict(naturally, not automatically armed in these cases).
Fourth, cost to the party. While the attrition model can be useful, recovery is usually cheap enough, and time plentiful enough that all it does is slow the group down while they heal up and head on. Traps which nickel and dime their way don't work, use those only if the party is already under time pressure. I prefer traps which alter the nature of the scenario ahead, if you're going to damage or debuff the PCs, set an alarm off so they actually have to deal with the consequences. Alternatively, adding more obstacles or re-designing the dungeon can be valid uses of 'quiet' traps. More long lasting damages can also go a good distance, ability damage recovers slowly, as do negative levels and disease, so they might represent too big a delay to actually fix without interfering with exploration.
Fifth, bypassing. If it's straight out impassable without the trapsmith, that is bad design, since there are many ways to be deprived. One of the most popular is of course, automatic reset traps, but while they interfere with just discharging onto a hapless minion, they just pose an insoluble obstacle. So think on it, doors have keys, traps have their bypasses. Make use of those, provided the party doesn't insist on headbutting the problem until the end of the day.
Sixth, plot, plot, plot. The whole point of anything is to advance the plot or generate inherently interesting reactions. Traps make a good way to limit exploration areas, particularly backtracking, due to the punishing costs for crossing them over and over. Traps can represent creatures and entities that are improbable to contest directly. Mechanical or magical, think beyond what is written on the stock devices and allow for innovative ways to attack the problem.
They have their uses, but their misuses are just so much more.
