My question Open Source Math Library for F#? was closed for being not constructive. My guess is that it could have been worded better (though it was fine at the time). But here is my question:

Shouldn't voting or closing a question need a bit of awareness about the field itself, rather than a syntactical check?

In this case, there are not likely to be arguments or extended discussions because there are almost no open source f# libraries, so given the question there was (and is) only one answer at the time ( The fact that most answerers supplied commercial non-open source answers would tell that.

In effect, the question is one about existence, and domain experts would know that. Shouldn't that count?

share|improve this question
While I tend agree with this argument in principle, your question doesn't really appear to ask that question. –  TCPMAN.EXE Jun 20 '12 at 20:20
Good point @mootinator. I can't entirely agree, but to be more constructive, how would you rephrase the question in this case? Would "does an open source F# math library exist and if so, what is it?" look better for you? –  Muhammad Alkarouri Jun 21 '12 at 19:53
add comment

2 Answers

In effect, the question is one about existence

If that was your intended question, I don't think many people interpreted that way.

Even yet, it is still a recommendation to an extent. If we answered the question as you intended it to ask, the answer would have been:


That answer isn't entirely helpful. Knowing something exists doesn't solve you problem since you need a library to use.

Surely that would solicit feedback, such as "Can you name an example or two?" which digresses into "Well which one is better?"

And now the question isn't constructive.

share|improve this answer
I would hope the answer "Yes" with no elaboration would be considered trolling. The difficulty is in proving definitively only n of something exist. –  TCPMAN.EXE Jun 20 '12 at 20:25
As a matter of fact, almost all answerers did interpret it that way. The question wouldn't have an answer "yes", but would have answer "yes, the library X exists" for suitable values of X. And that is reflected in the answers. And no, presuming that soliciting feedback about which is better is presuming another question whihc I never actually asked. –  Muhammad Alkarouri Jun 21 '12 at 11:54
In other words, if you need a math library for a certain development environment how would you go about asking that question? Or is that invalid in SO? –  Muhammad Alkarouri Jun 21 '12 at 11:55
add comment

I think the main problem is that your question is not answerable:

  • There's no single criteria that determines if any answer is helpful or not. @vcsjones has provided with a great example of totally correct although totally useless answer;
  • The question assumes subjective answer, kind of "What's your favorite _ _ _", which is specifically discouraged in the FAQ;
  • The variety of answers would not be that different to a result provided by querying a search engine. Notice you are not asking for a rare thing that might be down within the list of search results;
  • Also, think of SO as a storage for questions and answers. If your question stays on the site, being in search results, and other people visit it — what do they get? Would it be useful for them?

Choosing a tool for your specific needs may be a difficult thing. Since you didn't specify any requirements, any recommendation may be wrong.

If you tried a certain tool and you feel it is what you need, except something specific, you would come up with more specific question like "I'm trying to accomplish X with a tool Y, and it does not work". And this question certainly will be constructive and more likely to be answered.

share|improve this answer
I beg to differ. @vcjones has not provided any answer if you actually read the question "Open Source Math Library for F#?" or at least the title. –  Muhammad Alkarouri Jun 21 '12 at 11:57
"Choosing a tool for your specific needs may be a difficult thing." That was exactly what I was doing. I needed a library with the following requirements: (1) open source (2) works with F# (3) supports statistical distributions". I have showed that the only two libraries I found that satisfy the criteria are alpha and deprecated. And the question is actually answered satisfactorily, more than a year ago. How more constructive do you want it? –  Muhammad Alkarouri Jun 21 '12 at 12:00
add comment

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .