For these reasons, I wonder if the page http://mathoverflow.net/questions?sort=votes ought to have a disclaimer on it, for the benefit of newcomers that might read more into vote counts than warranted. (If feasible, the disclaimer could be hidden from users with sufficient rep.)

Here's a typical example: I remember one post discussing whether: the dump should contain deleted questions? It seems reasonable (and some people thought "required by law") to give people ability to delete content they posted under some circumstances and remove it as much as possible from the public eye. One may argue either way, but for now, obviously, For Creek erred on the side of caution (a reasonable choice for me, as long as their backups are really OK) and doesn't put deleted questions in the dumps.

So, if I were them, I would indeed provide a one-click link which would contain exactly the data visible to public. And I could do it for everyone, subject to available bandwith.

I can see lots of reasons why that would be extremely useful to have, if possible.

(You'd need to ensure that the dump didn't contain any private information, though.)

This will make the list of favorited questions (a simple select statement which I tested on SO dump), allow "Want to help? Consider..." to be changed from in-page javascript to a true list of questions, allow to have different representations of most voted questions (e.g. display where votes >= 25 but order by date) and give power to make funny queries like the ones listed above.

While I'll post it to SE meta,I actually suspect we're out of luck for a while. Perhaps we could do a public dump with sql interface?

As for use case questions, while it's a great idea to insert them into FAQ, we could also go on to add about 1-2 exemplary questions per week, which is kinf of too long for a single page. I like the way this is done in Wikipedia.

About other queries:

• I don't know of any way to see most favorited questions (even once you get access to the moderation tools at 10k).
• You can highlight the questions without arXiv tags by clicking the "Want to help? Consider retagging questions with no arXiv tag" link in the sidebar (only appears if you have 500+ rep). I feel like highlighting non-arXiv questions is the right thing to do, but I'm open to suggestions.
• You can already sort the questions by votes. The top 15 are the ones with more than 25 upvotes.

Updated with more query fun:

• What types of questions get closed most often?
• Users who are fastest in answering .NET questions
• more at this meta post

On a related note, I know there will never be direct query access to the database but I wonder if admins could manually add pages with results of some queries other than standard ones:

• questions ordered by the number of favorites (or maybe "?sort=something" would do the trick?)
• questions without arxiv tags
• questions with more than 25 upvotes

I'd be much more in favor of a list of examples of good questions that illustrate different use cases. That way the person reading the examples might actually be inspired to ask a good question rather than avoid asking a bad question. For example, Harry was unsure about a certain type of question would fly. It happens that it was exactly one of the use cases MO was meant for, but there aren't too many examples of that sort of question on the site. If there were a list of use cases somewhere, he wouldn't have hesitated. Another thing I like about this idea is that we could walk people through the process of making their question better. I think I'm pretty good about polishing my questions so that they're appealing to others, but it would be nice to have a checklist. It would also be nice to be able to point people to that sort of documentation.

Okay, I've convinced myself that a list of use cases is a good idea. I'll post a URL here once I have a draft.

Anyway, I maybe the unfunny person here, but I urge you to think about people completely without sense of humor, people who don't speak English well, people who just lost their partner, people who don't know that the next line below the question is not the only answer posted and everyone else who might come by. It's probably impossible to write a joke that won't offend a single person, but I think it's reasonable to try a bit harder than just go posting all possible jokes :)

Didn't mean to offend you of course, sorry if it looked like that :)

The administration of this building reserves the right to escort out any person from these premises at any time.

You don't see many of these signs on supermarkets, libraries, or homes (maybe unless you're a Hollywood star...) yet it in no way diminishes the legal right of any person to throw out unwanted guests from the place s/he owns.

But, how about the banners similar to "Welcome to the place of learning", "Please take off your shoes", "Smile, you're on camera". The last one is especially true: on the Internet posts usually don't disappear.

Yeah, we could set up a Ten Least Wanted List, and maybe create a Vilification Committee Of Safety.

There certainly could be different opinions about this topic. Let me state mine: I would prefer to have any soft question (even the ones I didn't like listed in my above post) to an inflammatory comment like the one you posted.

Now this was 1.5 months ago, so I guess I won't say much more, but hope I fulfill my duty to discuss with the person who posted an offensive to me comment :)

Do not ask this sort of question on mathoverflow.net. Mathoverflow is reserved for research mathematics and the like, your question would be closed even faster on mathoverflow. If you need help with someone simple like this please refer to mathforum.org/dr/math, artofproblemsolving.com/Forum/index.php/…, nrich.maths.org/discus/messages/… . Math Overflow is a horrible place to ask this sort of thing, they will close the question and be annoyed. Furthermore the answer to your question is: is in.

This strikes me as an efficient and essentially correct description of the policy about asking questions. I wonder if we should borrow some of that for the FAQ?

The reason I open this topic is because I have a feeling that "is this of interest to mathematicians?" test is really way too open to interpretation. The questions like

• How does one handle two-body job searches?
• How to write math well?
• Using TikZ in papers

trivially pass. On the other hand, homework questions would appear to pass (since many students, especially at well-staffed universities, are used to TAs and professors politely answering all questions) but they don't under current policy. Moreover, I haven't looked at the list of closed questions but I'm sure for at least half of them the case could be made that they should pass.

Here are some hypothetical questions that I have hard time differentiating from the ones listed above and not closed:

• (Assuming the next ICM is next week in Peru) What's the weather in Peru and what cities should I visit?
• (After an announcement of a new Apple tablet) Will my TeX run on an iTablet?
• If an LHC finds a Higgs boson, will this affect the amount of NSF support for postdocs who work on physics-related things?

Now there's some intuition that I have that says these example questions aren't the ones that should be encouraged, but I can't find any reason why they won't pass the test of "being of interest to mathematicians" with flying colors. Did I miss something?