Of course. I agree with the parent and grandparent that Wikipedia should cite the best sources possible. However, the parent's suggestion that we then have to cite closed access sources is based on a dubious assumption.
Here's the problem: the best possible sources are closed access, especially when we are talking about
things in medical research and life sciences.
I cannot judge for life sciences, but in my field of ocean modelling and climate research this is not true. Good sources used to be closed access sources, but more and more scientists prefer open access. Right now, in my field, open access journals (e.g. EGU's Geosci. Model Dev.) are at least as relevant as closed access (e.g. Elsevier's Deep-Sea Res. Pt II). This will only become more important in the future, hence Elsevier's attempt to stay relevant.
It is beneficial for active content creators to have access to these.
They will be able to create citations supporting articles on subjects that couldn't even be written otherwise.
Notability is a frequent issue on Wikipedia with articles on important subjects frequently getting deleted, because high quality citations have not been made to establish their notability ---- citations good enough to meet the criteria are only available through closed-access sources, such as professional journals.
It depends on both the field and the time when the science was done. I find it obvious that at least the original paper should be cited for any finding. Looking at the history of modern science, and that of open access, this is of course often closed source (well, some publishers make very old issues more and more available, but so far the point still stands). Anyone would agree that such a citation may be amended with another open access review paper, for instance. More recent research can be cited more and more from open access journals.
Finally... the purpose of Wikipedia is to be the encyclopedia anyone can edit,
Nobody ever said anything about the sources used by Wikipedia having to be the same
Indeed, nobody is saying that. We are talking about the accessibility of the sources, not whether they are libre or you can edit them. (That being said, there are even more advantages if the sources themselves are not only open access, but libre; this is already happening, e.g. many EGU journal papers are libre, and there is no copyright transfer either.)
It would hobble the encyclopedia and greatly limit its coverage, if only free citations can be used.
I love the idea of a free encyclopedia..... and I love the idea of open access journals, BUT let's not delude ourselves into thinking that the canonical work in the sciences are always the open access articles.
I fully agree. But often canonical work in science is in open access journals. I don't think what you say here is consistent with the first sentence about the best possible sources being closed access -- that is becoming less and less universal.
E.g. In article discussing relativity, I would much rather see the cite in the journal where Einstein actually published,
than some 4th order / quarternary source that someone preferred since it was an online magazine article available free of charge.
Yes, but there are many high-quality open access journals that may amend the primary citation.
I would also point out... open access today doesn't mean open access tomorrow.
Many times Online sources later go offline, or the publisher breaks the URL!
There are a lot of crap open access journals. Don't publish in those, don't cite from those. Within most fields it is clear what I am talking about. I think there is enough expertise between wikipedians to be able to judge which are good and which are bad. Papers should have a Digital Object Identifier (DOI). That does solve the URL breaking to some extend, in theory. At least I have not seen any counter examples in my field.
Now, what would be really cool is if Wikipedia could get a fair use "Excerpting" / "Automatic clipping" service,
where readers of an article could click on an "Excerpt" link by the citation and see an archived exceprt from the
article from online or scanned version, with the cited portion highlighted in yellow, and a bunch of context.
Then adopt a policy indicating that an excerptable source should be included for every referenced fact or assertion, when possible.
Yes, that is better than no content at all of the paper (or only the abstract).