Wikipedia talk:Shallow references

Not always the best idea
One situation I often come across is where an article has an abstract entry with links to the actual article. Frequently the DOI for the article will point to such a page. It often seems like using this page for the URL is better than using the actual file URL. There are several reasons for this. There's often a choice for which format to download the article, and this makes the article more accessible. The abstract page helps provide the user some orientation about what they are about to download to their computer. Using just the file URL may be more fragile (more likely to change) than just the abstract URL. And perhaps most importantly, in practice it seems (and I've noticed this from doing fixing thousands of article references) that the abstract URL's tend to be better for following up on a reference. I often come across the situation where there's some file URL like " www.somewebsite.com/arv/index/dec2012/av22k2/index/somefile.pdf " is dead and peeling off the filename and each directory in sequence all result in 404's until I'm down to just " www.somewebsite.com ", which may not give any clue how the original file is to be found again. On the other hand, websites seems to keep their abstract webpages working through redirects and such. So the URL should be as specific as possible but editors should avoid a simplistic "file link is always best" interpretation here. Jason Quinn (talk) 22:28, 9 March 2014 (UTC)
 * This is true, and I don't think the "rule" is supposed to be read that narrowly either. The point is to link as unambiguously as possible to the cited document. Perhaps the wording of the page could be improved somehow.--Anders Feder (talk) 03:07, 4 August 2015 (UTC)

Referencing geographic information systems
The typical GIS generates content on the fly and does not have a page. To get the relevant information one has to do various searches and clicks in order to display the relevant content. How does one utilize such a definitive source? This can be important when some information published in books, journals, newspapers, and websites is demonstrably wrong. Is there some web trickery that enable my search to be re-executed when a user clicks on the external link.

My experience with databases other than GISes suggests that many databases suffer from the same kind of problem, so that they cannot meet the no-shallow-reference standard. I am especially aware of sources in biological taxonomy, sometimes the best, that have this characteristic. The dead-link problem referenced above also bedevils references to sites that change their search interface, sometimes without any notice or even documentation after the fact.

If this can't be resolved, I will stop trying to improve articles that contain falsehoods that I know of, eg, articles covering localities near where I live or have lived or subjects I have studied closely. Why should I add content to a reference (WP) that I know brims with error? I can only assume that the credibility of my content, sourced or not, is diminished by such obvious error. DCDuring (talk) 19:39, 2 October 2017 (UTC)