I'm pretty extreme (converts are always at the extreme end), yet even I would allow that there are good soft questions. None particularly spring to mind, but I'm sure that there are some. But as there are so many completely banal ones, a soft question really has to shine out to be acceptable. If the best that one can hope from a soft question is for people to say, "Oh, that's cute" then it has no reason to exist. Snuff it out.

I thought that my question about Lang's famous exercise in Algebra was a good soft question, but then again, it also didn't lend itself to much discussion.

[...]I'd probably only use the nForum and nLab in that case as those are the only ones I know of with a higher density of mathematics than MO (several orders of magnitude higher, in my very biased opinion).[...]

Since you've broached the topic, I must take a moment to note that the nLab is surprisingly sparse regarding commutative algebra or algebraic geometry. Since MO has so many algebraic geometers, I think that mentioning this in a popular meta thread might convince some of them to write some of the basics of algebraic geometry up in the nLab...

If you're ambivalent, pretend I convinced you.

It is like the question about using the blackboard,

Assuming that you refer to this question, it's worth stating for the record that this question has been closed. Also, it was the asking-and-answering of that question that largely lead to my conversion from "Let's use MO for whatever nefarious purposes I can get away with" to "Let's focus on the math, the whole math, and nothing but the math."[1]. The answers that I got were deeply frustrating: I wanted to come back at almost every one and point out how ludicrous their answer was! But the MO software doesn't let me do that, not really. There's no real possibility of having a true discussion/debate that can lead to a reasoned argument that might actually change someone's mind. I would be very surprised if anyone, upon reading the question and those answers, has gone away and really thought about why they prefer one medium over another. I'd be even more surprised (I may have to resort to astonishment) if anyone had actually changed their behaviour (or at the least done an experimental change). If I'm right, what was the point of my question? Yes it's interesting, but ultimately is it going to change anything? No? Then there's no point in asking it.

I'm pretty extreme (converts are always at the extreme end), yet even I would allow that there are good soft questions. None particularly spring to mind, but I'm sure that there are some. But as there are so many completely banal ones, a soft question really has to shine out to be acceptable. If the best that one can hope from a soft question is for people to say, "Oh, that's cute" then it has no reason to exist. Snuff it out.

As for "good answers make good questions". No, No, No. If you've got a good answer to a rubbish question, ask the question that allows you to pose that answer. If there isn't one, then it isn't an answer, it's a statement. MO isn't for statements.

Deane's comment:

The traffic on MO of mathematicians at all different stages of their careers here is completely unique and it would be nice to allow the discussions to go beyond what is currently the officially defined narrow focus of MathOverflow.

is very interesting. That's what I used to think. I don't any more. The key for me is what makes it have such traffic. I believe that if we allowed further discussion then the nature of the traffic would change considerably. I've now almost completely cut out reading maths blogs in favour of MO. If Deane's picture of MO came in, I would cut out MO in favour of something else where the maths is kept at such a high level - I'd probably only use the nForum and nLab in that case as those are the only ones I know of with a higher density of mathematics than MO (several orders of magnitude higher, in my very biased opinion). Another thing worth pointing out in Deane's comment is the word "discussion". MO is not for discussions. I like that. Keep things focussed on the maths! For one thing, not being a great place for discussions means that I can simply ignore it when people call me names or write incendiary comments on my posts (I know I ought to just ignore such things anyway, but hasn't anyone else noticed how it's much harder to ignore an inflammatory discussion here on meta than on MO?).

To Bill:

Can't MO have a second forum about the profession to take care of questions like this one?

This has been suggested many times. Both Scott and I have said that we would be happy to set one up. However, neither of us (I think I'm not misrepresenting him) want to actually run such a place (as in moderate, we're happy to do the technical side). Indeed, I now have the software that runs the nForum in such a place that setting up such a forum would take about 30 seconds. So if anyone wants to get a team of moderators together to run it, vær så godt!

[1] No, I haven't gone over to the other side. It didn't scan so well with the correct word used.

I find it interesting that though the question has 16 upvotes, it never received any answers. (Admittedly it spent some time being closed, but I think it must have been open for several hours at least.) Does this mean that the people who upvoted it didn't have their own stories to tell but wanted to hear others' stories? Perhaps for inspirational purposes??

I remember Ben making a good point about this kind a question back in the prehistory of MO: if you want to ask questions about this sort of subject, you have to work out the question which admits a definitive answer. I can't think of one right here, but the general idea was that instead of asking "What works for you in the classroom?", one could ask "What classroom methods have been studied and formally evaluated?"

An optimist would say that these "definitive" versions will very likely be more useful to later readers.

Of course it's interesting, getting a glimpse of what motivates other people is interesting, but that's not what MO is for.

I'd like to respond to fedja's comment:

MO definitely became more of a "socializing site" than had been planned originally

If the social aspect of MO ever becomes more than a happy byproduct, that will be the end of MO. I do enjoy the fact that I'm getting to know other mathematicians through MO (even one or two in my own department!), but that isn't why I visit MO. I'm getting to know other mathematicians through their mathematics. The difference between this and, say, reading papers on the arxiv is that I get to see smaller chunks of what people are thinking about, and get to interact. But it needs to stay focussed on the mathematics, otherwise it becomes just another Reddit or facebook-type site, and all its actual value goes down the plug-hole. People don't "come for the socialising, stay for the mathematics", they "come for the mathematics, stay for the mathematics". The social aspect comes in simply because, as human beings, we can't focus that strongly on one thing for that long.

It's like working in a library. It's good to take a break, step outside, have a chat with anyone else who happens to have stepped outside, but if we stay outside then we may as well have been in the tea shop, not the library.

There may be an acceptable question in there - something along the lines of

'How do you persuade someone that mathematics isn't just "some pretty dull multiplication-related stuff designed for people who didn't have enough creativity to tell a good joke"?'

but the current question is a long way short of that.

More seriously: while the question is now (re)opened, I agree with fedja that the question needs some editing.