Is it poor form to switch accepted answers?

Question

I had a question of the form "does something like X exist?"
Someone wrote a response to say that, No, X does not exist, but I could write it if I liked.
I had suspected as much, so I marked his answer as the accepted answer. Today someone actually posted the code to do X for me, and I feel that the new poster deserves the "Accepted Answer" checkmark.

Is it poor form for me to pick a new Accepted answer? Does the original answerer get penalized in terms of reputation?

While the original responder's answer was accepted at the time, I feel this new answer is really more helpful.

Edit:
Which of these responses should I accept?
What does the notion of a single accepted answer mean with a question that is subjective that could have multiple "right" answers?

Answer 1

Is it poor form for me to pick a new Accepted answer?

Absolutely not!

If it's switching to a more correct answer, it is the right thing to do.

The original user will lose the 15 rep bonus immediately.

Answer 2

Is it poor form for me to pick a new Accepted answer?

Not in my opinion, though I don't do so gratuitously.

Does the original answerer get penalized in terms of reputation?

Well, they lose the 15 points (if not community wiki) and that answer might cease to qualify for a badge (the badge doesn't go away, but they have to "make it up" before they can get a new one of that type).

Answer 3

I see no problem with changing the accepted answer if a more comprehensive or factually correct answer is added at a later time.

It isn't mean or spiteful. It is just you doing the responsible thing with the questions you own and it benefits everyone who comes to that page later.

Answer 4

As everyone else said, there's certainly no problem with it. If it seems to be happening to you a lot, you might be accepting an answer too fast -- you should wait a couple days to give people a chance to answer

Answer 5

If StackOverflow is going to be a self-perfecting resource, then optimized answers should ALWAYS be permitted/encouraged regardless of their timing. When readers/researchers come to SO for answers/methods, they do so with confidence that they are going to find the most refined answer with the green tick (or with a disproportionate number of upvotes) with minimal wasted time.

Computer sciences must not allow emotions to inhibit creativity and excellence. Askers have the sole responsibility to award the green tick to the best performing (and most educational) answer for the greater good of SO and its ever-growing audience.

Not shifting the green tick (out of pity or whatever) to the best answer does a disservice to SO and its users.

As a knock-on effect, by rewarding later/better answers and stripping points from earlier/worse answers, a culture is established where the "Wham-bam thanks for the rep points for my half-considered method" posters are not rewarded and the posters that make fully considered answers are highlighted/rewarded. This is the only way to encourage the sloppy high-volume posters to slow down and write their best possible work.

Answer 6

Think that the idea about these sites is to get the best and most complete bookwork there is about the subject. So it is always the best to accept the best answer suitable. In time that could change.

Your question could slighty change too in time due to new knowledge gained and you have refrased the question, so another answer could be netter suited.

Don't feel bad for dismissing some point as it ain't a race to win. Not in that way.