In this answer I was asked to quote Wikipedia as the source for the source that I used in the answer.

I responded:

I did find one of the links via Wikipedia, but I'm quoting the source directly rather than quoting Wikipedia. I'd consider it analogous to finding something via google, and not quoting google... unless I have something to learn about quoting on StackExchange?

The reply was:

It's not analogous because some person did the work for finding that reference and adding it to Wikipedia. Not acknowledging that doesn't seem right. I'm sorry, but I'm downvoting this.

I can see that things could devolve into quoting a source quoting a source quoting a source, but at the same time, perhaps one level could give a greater context?

What is best practice in this regard?

2 Answers 2


The litmus test on SE is effectively how people vote on your post, which in your case would tend to indicate that most readers don't mind much! (There's probably some amount of drive-by upvoting involved, as pointed out by muru. Upvoting and downvoting also tend to be unbalanced since the latter has a cost: it usually takes more incorrectness to get an answer downvoted, than it takes correctness and usefulness to get it upvoted. One of my most-upvoted answers was incorrect at first, and got many of its upvotes before I corrected it...)

As far as best practice goes, I would consider three aspects:

  • effort
  • sources
  • investigation


Posting an answer to a question (and even posting a question) involves a variable amount of effort, depending on the question, the answer and the answerer. In my mind, any person who helps reduce the amount of effort involved in answering a question deserves some credit: thus if someone points something out to me in a comment on one of my answers, I'll generally credit that person in my update which takes that comment into account. In this instance, whoever wrote or updated the Wikipedia article which pointed you in the right direction probably deserves some credit by that measure. What's more, you can do this in a way which doesn't affect the readability of your answer much:

Wikipedia indicates that the "u" stands for "unlimited", based on the fact that the random_read_unlimited function name seems to indicate that the etymology of the letter u in urandom comes from unlimited.


I agree that pointing at definitive sources is better than pointing at intermediaries, so it's useful to have the link to the source of the function in question. It's even more useful to have the link to Ted Ts'o's message on sci.crypt which Stéphane Chazelas dug up.

Nevertheless I think the Wikipedia article is also a source, and not just an intermediary, since that's where the interpretation is given that 'u' stands for "unlimited".


Finally, in many of these types of answers (historical, or digging into a complex technical issue), the investigative process is as interesting as the answer — in some cases even more interesting! So it's always useful to include that as far as I'm concerned, at least as long as it doesn't affect the readability of the answer. One useful approach can be the journalistic pyramid, or tl;dr: you give the answer in brief as the first paragraph, then add more and more detail in later paragraphs.

You'll find loads of examples of different styles on the site. Some users who tend to document their answers extensively are Stéphane Chazelas, Gilles, JdeBP and Thomas Dickey; it's no accident that they have some very highly-upvoted answers!

  • The problem with voting in this case is that this post hit the HNQ, which severely skews the voting because of drive-by upvotes (user see interesting post, user has association bonus, user upvote). I agree with the rest of the answer, though.
    – muru
    Nov 17, 2016 at 12:39
  • Does unix.SE have different attribution practices to other SE sites? Assuming no (and happy to be corrected): there has been time compression due to the HNQ status, but I'm curious about how would affect the value of the votes cast?
    – Tom Hale
    Nov 18, 2016 at 15:48

I would not generally quote the source of a source, and consider muru's comment out of line. In this case, I would only quote Wikipedia if it provided distinct useful information — if it's information that I could find with a minute's search on LXR, I would not cite Wikipedia, even if I read the Wikipedia article before I looked up LXR.

The standards for citations vary between Stack Exchange sites. Unix & Linux is rather on the low-citation side because most answers concern verifiable statements — you just type the code and run it. If you found information in a third-party, you should cite it, but it isn't strictly required. (If you copy text, that's a different matter: you must cite your source.) There are several purposes to citing a source: it's an acknowledgement (credit goes to the one who found or produced the information in the first place), an authority (if you don't believe me, believe that guy), and a place to look for more explanations (either background or further explanations).

Generally, if a source only says “look at this other source”, then there's no point in citing it unless the other source is hard to find. If Google hit #1 links to Google hit #2 but is otherwise not useful, there's no point in citing Google hit #1. On the other hand, it would be good to cite the Wikipedia article if it provided a comprehensive overview of the topic gathered from multiple sources, or if the original source was hard to find.

You must log in to answer this question.

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