3
$\begingroup$

There is at least one user who has been editing posts to make links into footnotes containing the references as full text (including links).

Can we please officially discourage this practice?

This is not how hypertext works: links replace footnote references. The are technically superior. Printed texts use footnotes because they cannot have hypertext links. But text on the internet doesn’t have this restriction.

Footnote links are user-unfriendly because instead of just hovering over the link, or clicking on it to get to the linked information, the user has to perform the manual operation of looking up a numeric reference further down the page.

Apart from that, footnotes simply increase the text size at no (or little) gain.

I do think that having a “references” section may make sense in answers. But this isn’t generally true, and unless there’s a compelling reason, links should not be replaced by footnotes. In fact, I have already rolled back such an edit performed by somebody else on my own post. I won’t roll back edits on other posts without a consensus, which I’m seeking here.

$\endgroup$
11
  • 3
    $\begingroup$ That same user is also changing direct links to dx.doi.org links. I disagree with this as well but that’s a different topic. $\endgroup$ Commented Sep 13, 2012 at 21:48
  • $\begingroup$ I would however like to at least recognize @aland's efforts. $\endgroup$
    – bobthejoe
    Commented Sep 14, 2012 at 7:29
  • $\begingroup$ Ach I approved a fair few of those last night from mobile - I had assumed that the numbers were links to source too. My bad. Although rechecking now they do seem to be. $\endgroup$
    – Rory M Mod
    Commented Sep 14, 2012 at 14:03
  • $\begingroup$ @Rory They are – I didn’t notice that, and even if I had I wouldn’t have thought this made a difference until reading Mad Scientist’s answer below. $\endgroup$ Commented Sep 14, 2012 at 14:49
  • 1
    $\begingroup$ I only change references when paper is linked via journal link, which can rot in case journal decides to change its webengine or something (this has already happened with link in biology.stackexchange.com/a/544/836; I've managed to find the paper only because there was verbose citation of it with the same broken link in wikipedia). I use citation style proposed in meta.biology.stackexchange.com/a/285/836 $\endgroup$
    – aland
    Commented Sep 14, 2012 at 20:01
  • 2
    $\begingroup$ @aland: obviously the same reasoning could be applied to dx.doi.org in the (albeit remote) situation in which they decided to change the way they handle links. That said, I personally would prefer to see links to Pubmed rather than doi. This is much better because: a) you will have all the Pubmed tools (related papers, links to other DBs etc) and b) when visiting Pubmed through certain institutions you would automagically have links to the full text of pay-per-view papers. $\endgroup$
    – nico
    Commented Sep 15, 2012 at 6:48
  • 4
    $\begingroup$ @nico: Even if DOI changes the way it handles links, you still have a) DOI itself as part of the link (many catalogs, including Pubmed, can search by DOI); b) full name of the paper. I understand that Pubmed may offer sometimes more tools, but DOI is, IMHO, more established. There already are several threads on meta regarding citation styles, and no single conclusion was reached; I guess the discussion PMID vs DOI is more appropriate there. $\endgroup$
    – aland
    Commented Sep 15, 2012 at 16:51
  • 2
    $\begingroup$ I think the edits are fine. We've already discussed here why dx.doi.org is superior - you can quickly swap out all dx.doi.org links with any other doi processing server and all the links are restored. A simple search and replace. Whereas there is no way to automatically restore links if a publisher changes their link format. $\endgroup$ Commented Sep 16, 2012 at 14:06
  • 2
    $\begingroup$ DOI is a much more universal system than PMID. PMID is restricted only to life-science stuff indexed by PubMed, whereas any digital content can have a DOI. If PubMed goes down, all PMIDs are useless, whereas if dx.doi.org goes down, we can switch to a different doi server. $\endgroup$ Commented Sep 16, 2012 at 14:09
  • $\begingroup$ I'm pretty certain SE would help us replace every dx.doi.org link with a link to a different DOI server if that one ever went down. $\endgroup$
    – Mad Scientist Mod
    Commented Sep 23, 2012 at 9:58
  • 2
    $\begingroup$ Doi may be more universal but Pubmed is kore appropriate to a biology site isn't it? If pubmed goes down we're in deep trouble. And @nico is right using pubmed links will give most of us direct access to the article. $\endgroup$
    – terdon
    Commented Oct 27, 2012 at 14:50

1 Answer 1

13
$\begingroup$

I didn't check all cases, but in the ones I saw the original link was not removed, only the footnote reference added. Both, the inline mention of the article and the full citation were links. There is no functionality lost there, you can still click or hover over the original link.

To me, the footnote citations are useful to quickly check which papers are cited without having to click on every link. Hovering over the link usually tells me the journal name at most, often only the publisher. In many of my posts I use them myself, but always with an inline link as well. They don't replace hypertext links, they are complementary.

There were a few cases where the full citation was added inline, I do disagree with that usage. We should stick to standard nomenclature here (Rudolph et. al, 2012) or explicity mention the title in the text, not put full citations inline.

Changing links to DOIs is generally a good idea, not all scientific publishers seem to place a lot of value on keeping their links stable. If the link is to a different source of the paper (e.g. free full text on PMC or the authors homepage) that link should be preserved, not replaced.

$\endgroup$
1
  • $\begingroup$ Hmm. I have to think about this but I think I agree that this makes the edits seem much more acceptable. $\endgroup$ Commented Sep 14, 2012 at 8:57

You must log in to answer this question.

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