1
$\begingroup$

The mods had a discussion about whether a coding question was valid without any stated relevance to bioinformatics, a typical question is "here's a bit of text how do I parse it?". The issue is that such questions could be answered on lots of different sites, not just Bioinformatics SE.

Example A current example, is how to perform a database merge/join/union operation between two dataframes. The information sought for this given question is what is the biological reason, or biological objective, for the database operation. In other words, how does this new database help further biological insight into the data set. The OP might want to augment the database to include new information for example on disease incidence outside the original study so the new db applicable to a bigger geographic area. The OP might want to correlate two specific factors of disease incidence for further analysis. In this given example, the question was simply

its not worked, here's the error log, please sort it out.

$\endgroup$

2 Answers 2

1
$\begingroup$

Biology The proposal is that such coding questions are cool, but the biological and/or bioinformatic background needs to be stated. What is sought is to explain the rationale or the objective behind the need to perform a given coding operation.

If we generally refer to the background information sought as 'biological/bioinformatic context'.

Rationale IMO there three good reasons for requesting the 'biological/bioinformatic context':

  1. Miscommunication is common in many bioinformatic questions between the OP and those answering. This can result in providing a solution outside the biological objective the OP requires.
  2. The XY problem. This is where the OP needs a solution to problem 'X' but thinks the way to answer 'X' is by solving problem 'Y', when solving the original (unstated) question is a better strategy.
  3. The species, genus, Family, Order or even Class is not stated and a solution is provided for a completely different part of the Linnean classification system which isn't relevant. Haplotypes require a completely different set of solutions to eukaryotes, lower eukaryotes compared with higher eukaryotes, or Orders within a Class.

To expand point 1. in the database example a 'merge/join/union' operation is sought but the OP does not know the operation name and tries to describe it by creating an ad hoc technical language. If the biological reason is described the reader can then reinterpret what is sought as a technical operation and that is how many bioinformaticians think. If the OP wants to query the database for a specific transmission factor with a join operation, but those answering think it's a union operation ... that will result in the wrong answer.

In summary, the 'biological/bioinformatic context' reduces the risk of 1., 2. and 3. by increasing the clarity of the question. That can only be a good thing.

$\endgroup$
1
$\begingroup$

The primary reason for adding context is that it helps to find the right solution for the problem that a person is having. I don't think that context has to be a story, but a story is usually the easiest way to find out a person's underlying problem. I agree that in general there is an issue with people not providing enough context, leading to incorrect / unhelpful answers, but don't think this question is a good example of that.

I don't have an issue with problem duplication or cross-posting (especially if it is explicitly stated / linked). If someone comes here with a problem that is a pure maths problem, or pure programming problem, that's fine by me. While they'll probably get a better answer on another site (and I would possibly guide them to do that), it's possible they feel more comfortable asking on this site, or think their problem is applicable within the very broad realm of bioinformatics.

I don't think that context has to be a story, but a story is usually the easiest way to find out a person's underlying problem.

To clarify on specific points relating to the referred-to question:

  • The contents of the examples they provided indicated that they were working with biological data.
  • It's unclear whether the question is a good example of not providing enough context, because they eventually provided input files that were sufficient to reliably demonstrate the problem. The "not enough background" point may still be applicable in this case, even if their immediate stated problem is solved. One thing to think about is whether or not this individual problem would be helpful to other viewers of the site (I would suggest it is, because it's an obscure problem with an odd cause). If so, it would be reasonable to keep it even if it doesn't provide a solution to their ultimate problem.
  • This was not a general database problem; they encountered a specific error using the merge function in R. This is a real function in base R; there is a help page for it.
  • The person asking the question was a new user. It's especially important that we (i.e. the community) put more effort into helping new users, rather than setting off quick-fire close requests and scaring them away because they get hit by the rules hammers.
$\endgroup$
1
  • $\begingroup$ Thanks, upvoted. It's agreed there's a problem. The next step is to find scenarios that are clear examples of this. No need for a 'hammer' just a reference point that can be presented as a link. $\endgroup$
    – M__ Mod
    Nov 19, 2022 at 14:39

You must log in to answer this question.

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