In our post about the recommended style of citing on chemistry.se we state:
The highest principle should always be: Give credit.
I like it how many answers substantiate their explanations with external references. Giving the citation correctly is especially important when (block) quoting from the original.
It is also very important when using images from external sites, that require a certain way of attribution. (Please always check that you have the permission to redistribute images.) In some cases the image is used and the source is indicated by including a link. For example like this:
(source) [In this case both links are the same, because they are from our own site.]
Sometimes this is sufficient, but more often this is not. When using images from stack exchange, everybody is expected (read: required) to follow some basic rules. In that spirit the above mentioned source should have been:
Any other use would be a violation of copyright and could give rise to removal. That is especially bad when the question/answer becomes void. It is only fair to follow our own rules. When using public domain images it is nice to indicate that they are. For example:
I do always assume for an image on our site, that the author of the post is also the creator of the image if it is not stated otherwise. I admit that I might have done that incorrectly in the past sometimes, but I tried to fix these instances.
Obviously that doesn't stop at images.
Whenever I come across a post that is linking to an external source and when I have the time I try to include the actual citation or amend the source to follow copyright guidelines. (If I don't have time to do that, I usually only leave a link). Except for one time it was either completely ignored or welcomed. The following should serve as short example, what I find insufficient:
Steinmetz and Grimme investigated the performance of density functionals for bond activations with catalysts of these metals. They find:
The PBE0 hybrid functional, together with our atom-pairwise dispersion correction (D3), shows the best performance for the complete set followed by PW6B95-D3, the corresponding double-hybrid functional PWPB95-D3, and B3LYP-D3.
While it gives the authors names and links back to the resource pdf (in this case the publisher pdf), if the journal decides to change that, or the journal changes publisher, or is sold/closed down, you might not be able to find the resource again. In general I would at least see this as a somewhat questionable scientific method. [In this special case I used an open access resource, so there will probably be an accessible copy elsewhere.]
This becomes more dramatically worse when it is not the link to the publisher resource, but to a different source (like an institute homepage, preprint, etc., it does not have to be illegal). This is quite neatly demonstrated by the related meta question: Policy on linking to papers? The question linked there suffers from link rot via the exact process described in it. This wouldn't matter much if the citation has been given, or at least a (more) stable link would have been inserted. In the above example I would have included the following, after replacing the link with a citation marker.
This uses the double fallback as you have the written out citation, which you can find via a search engine, and it is linking via the DOI, which we can assume to be around for quite some time.
All this leads me to basically to the title question:
Should we enforce including actual citations?
And with enforce I mean (actively) encourage to include citations, or insert them ourselves as edits from a (not-yet-dead) link to the sources?