Is it a bad practice to make assumptions in order to answer a question?

Question

There are always going to be those low quality questions that have a few gaps in the question and are quite often down-voted for this reason.

A few times I have tried to understand these poor (often down-voted) questions and thought I may know what they are trying to say, but potentially English as a 2nd or 3rd language has been a barrier to a good explanation of the question.

A few times now I have made a few assumptions and explained these assumptions in an attempt to give info in an answer I think might be helpful.

So my question is: Is it bad practice to make assumptions about missing information in order to give an answer I think could help the OP?

In a few cases now I've been down-voted for my answer and told that it's not what the OP is asking for.

Update after being marked duplicate

I have been told my question is a duplicate question for "excessive comments trying to get the OP to display more information" and another question about waiting for OP edits before posting an answer.

My question is more about posting an answer with assumptions whether or not there are comments or other answers.

I'm happy to accept that "Making assumptions on poorly written questions is indeed bad practice" is the most likely answer to my question, but I'm keen to keep it open for others to comment on their thoughts.

Answer 1

There are SE sites, such as Drupal.SE, where answers may depend on the release of a software in the question, whereas that release is indicated in the question using either the title, the body or some release tag.

Where relevant, I used to post comments to questions to ask for clarifications about which release the question was about (to ensure my answer I was planning would fit with what the question was about). But only part of such comments would get clarified (or I'd not get notified by it). And if it'd get clarified I might have forgotten how I'd want to answer the question.

However, these days I've changed my approach. Where appropriate I include an assumption about the release that my answers are about, for these reasons:

1. If the release is not clear from the question I don't have to wait for OPer anymore.
2. If later on OPer somehow indicates (eg via an extra comment below my answer) that my assumption is wrong, then that's OPer's problem (not mine). Because based on my assumption, one can not say my answer is invalid.
3. In case existing release info ever gets removed from the question later on ... my answer does not get invalidated.

Example

Have a look at my answer to "Rule to loop over products with one specific SKU in Drupal Commerce cart", and note these things about it:

• The question does not explicitly indicate which version of Drupal it is (via title, tag or question content), though via deduction/experience you can tell it's either for version 7 (= D7) or 8 (= D8).
• With a little extra experience, it's obvious it must be about D7, and not about D8 yet.
• I was pretty sure about what the actual answer to the question was, so without posting any comments (or asking for extra clarifications), I posted a D7 related answer.
• Note the "PS" I added to my answer, i.e "I'm assuming this question is about D7." (which I often do by simply using my -d7 keyboard shortcut, as explained here).
• If in (say) a year from now anybody hits this question again, it's obvious that my question was related to D7. And maybe another answer got posted by then with a similar answer, combined with a "PS" like "I'm assuming this question is about D8 ...".
• Better still, when the time is right, I might revisit all my similar answers (user:39516 is:a assuming about d7) for which I may want to post a D8 related answer.

Another viewpoint

Here is another way to look at it all: if a question requires an assumption about "X" (say with 3 possible options like 1, 2 and 3), then the question has the potential of 3 different answers (based on the assumed value of the option for X). The user who posted the question can pick what fits best for him/her (via accept and/or vote), while other users may find another (not accepted) answer more valuable for their case (and upvote that answer). Looking at it from an SE-perspective: a question with potentially 3 upvoted answers is great also, and pretty sure gets ranked higher in the hot/weekly/monthly list of questions, as compared to a question with a single answer.

Answer 2

I have seen some comments and other posts and I think I now know the answer to my own questions.

The overall answer is "yes", it is indeed bad to make assumptions to answer poorly constructed questions. Stack Exchange sites are all about providing the community with well constructed questions with equally well constructed answers with references to documentation and other reputable sources.

By making assumptions there is a far higher chance I'm going to provide an answer that is useless and/or does not assist the OP.

I'am better off closing the question to encourage the OP to improve their questions and/or remove poor content from the Stack Exchange world.

Answer 3

I'd prefer not to. I do it anyway sometimes - or even dig up small clues and actually fix up the OP's answer, so it depend.

If you're going to have to read the OP's mind, you're better off closing as unclear. If you can deduce minor facts, its probably cool.

If you can make reasonable assumptions or better yet, reasonably include some of them in an answer maybe. But a better idea is to ask the OP these things, rather than to answer and find you have a completely irrelevant answer.

Even if you've answered, assuming its still the same broad question you should probably ask for more information and improve your answer.