4

I've seen a few cases where someone asks a question about doing some rather specific task A with some rather specific tool T and then an answer is provided which maybe doesn't quite do A (but probably also solves the problem of the OP) or doesn't use the tool T (but solves the problem anyway!). The answer, anyway, could be very useful and, in some cases, perhaps people having a similar question will also think that it's better to use this answer that to pursue "trying to do A with T".

To give a specific example, in a question on Graphing sine function in MetaPost an answer was given using TikZ, which I think many people would find just as useful.

Perhaps this isn't the best example, but specifically for TeX/LaTeX, very commonly I've found people asking me how to do some rather specific hack H. When I ask why do they want to do this, I figure that they don't really need to do H, but there is another simple solution (usually a package) that solves their problem without the need of any hacks.

How should we deal on this site with these out-of-the-box answers?

8

Questions that ask "How can I do X with Y?", when in fact you only care about doing X (or possibly even Z) and not necessarily with Y, are extremely common when asking questions on the internet. (Sometimes called the XY problem, examples here and here.)

It remains a matter of judgment whether it's better to answer to the letter of the question or the spirit (answering the question you think they "intended to ask"), and whether doing the latter is being helpful or presumptuous.

Personally, I think answers that "think out of the box" and answer the (presumed) "real question" should be welcomed — the alternative way may not have occurred to the questioner — unless the questioner explicitly makes it clear that s/he is aware of the other sort of solution and specifically wants a solution using the mentioned tool.

  • 3
    Indeed, I think this is the key: "unless the questioner explicitly makes it clear that s/he is aware of the other sort of solution and specifically wants a solution using the mentioned tool." – Juan A. Navarro Jul 28 '10 at 17:15
  • @Juan: I agree with your agreement. :-) Those who can think of out-of-the-box solutions they don't want can simply forbid them in the question, and for everyone else, in case such an answer does not help them, they can simply leave a comment on the answer saying it's not what they want. (In many cases the out-of-the-box solutions may even be exactly what was wanted, e.g. this question.) – ShreevatsaR Jul 29 '10 at 4:32
3

As the author of the cited answer, I can say that uppermost in my mind was that my answer might be useful to someone asked an almost identical question but without the meta-post restriction. I should also say that I was completely aware of the fact that this could be viewed as borderline for what is a clear and useful answer and that by posting it I was intending on people wondering about its suitability.

I hope that in this case, I was on the right side of the line. As anyone from MO (mathoverflow) will know, my views are pretty hardline on making sure that answers answer the specific question. But I think that this is one situation where the lower barrier-to-entry here compared to MO makes me more inclined to be relaxed; I hope that people will try to make their questions here focussed so that "I want to do A with T" is typical of the type of question we get. But I suspect that many people will not be aware that "S" is possible and so although their question specifically says "T", the answer with "S" will be fine. So that people who post answers aren't scared of posting such modifications, I would like to think that such answers are viewed as acceptable.

However, there clearly must be a line. It would be inappropriate to say "You asked about LaTeX, here's how to do it in ConTeXt".

If people agree with me that slightly out-of-the-box answers are acceptable, then for the actual line we can let the community decide. However, I would say that it's our responsibility (as early users) to set the tone that these, in moderation, are acceptable.

For anyone answering a question with an out-of-the-box solution, I would say that honesty is the best policy. Acknowledge that it may not answer the direct question, but that you hope it might be useful to others looking at that question.

As a general policy, I dislike doing things whose sole motivation is for "future questioners"; but I'm quite happy to make minor modifications in my behaviour to help these future ignoramuses (of whom I will probably be one!) if the gain is sufficiently large. My opinion was that this was one of those cases, but others may, of course, disagree!

  • I mostly agree with your view, the thing is how should we come up with a way on where to draw the line? – Juan A. Navarro Jul 28 '10 at 16:08
  • @Juan A. Navarro: I think that first it's important to agree that there is a line, which is why I'm glad you brought this up. I think that the community may be able to decide on where the line is, once it is aware that it exists. – Loop Space Jul 28 '10 at 17:24
0

Question about how to generate font metrics for TTF fonts to use with pdf(la)tex get "shot down" with "don't waste time use XeLaTeX".

I'm somehow sceptical that there will ever be a step-by-step guide / elaborate answer of what needs to be done to make tex, latex, pdf(la)tex recognise TTF fonts, weights, and styles. The best reference so far deal with raster fonts on MikTeX without giving a clue how to do on TeXLive.

On the one hand "it is easier and quicker to do it with xelatex" but on the other hand "out of the box" asnwers can "water-down" the expert quality of the site.

  • but there’s extensive documentation on how to do that with TeXLive (any recent distribution, as far as I know) – someone just needs to post a link to it. – Konrad Rudolph Jul 28 '10 at 15:18

You must log in to answer this question.

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