Let me go on a bit of a tangent here.
One thing that's always bugged me about Worldbuilding Stack Exchange is that we don't have a large community of experts. Now, really, really early on in Stack Exchange's life, Jeff Atwood (co-founder of Stack Exchange) wrote
The idea that you have all these experts waiting in the wings to do stuff is an illusion in my experience. There's really just a bunch of amateurs muddling along trying to do things together.
This works on a lot of sites, and in fact, I would argue that virtually no Stack Exchange site - Math Overflow excluded, especially as it was originally from outside Stack Exchange - needs to be a community composed solely of, or even mostly of, experts. I mean, what's the point? You need people of all levels, with all different types of knowledge, with all different backgrounds and areas of interest.
At the same token, you do need people of all levels of experience, and on Worldbuiding, we don't have them all. Don't get me wrong, we have a lot of very smart people doing very smart things, but we don't really have a non-negligibly small set of biologists, or physicists, or geologists. We do have some people - and thank you, thank you, for your contributions; you know who you are - who work in scientific fields and are generous to spend some of their time here on Worldbuilding, sharing their knowledge, but they're something of an endangered species.
This is why I've been pushing and pushing the hard-science tag for so long. The first Artificial Intelligence site was shut down because it didn't have expert-level questions - or answers, which is my focus here. Now, we can afford to totally through away the tag, burn it, and never look back, because the majority of our questions don't need it. But some do. And if we don't have a large expert community on our site, where do we go? Those who actually do the stuff we're talking about.
Fortunately enough, we have something called The Internet at our disposal. In less than thirty seconds, assuming my Wi-Fi isn't on the fritz, I can sit down and open up half a dozen scientific papers on a certain subject (true fact, by the way). I can scroll through scientists' websites and blogs (cough cough Sean Raymond cough cough) and get their thoughts on issues they work on each day. I can read and think and calculate and learn.
See, the Internet is more than just Wikipedia. Knowledge - real, true, verified, honest-to-goodness knowledge, not this stuff - lies beyond its borders. Even if you want to simply start at a Wikipedia article, it has references and citations and links to additional reading. Information on Wikipedia comes from somewhere. The scientists don't just do an experiment and then type it into a Wikipedia article. They experiment and they write and they publish in a journal.
That is where you should be going first. If you find a Wikipedia article, it either has
- Good citations that you can follow through and read, or
- Crappy or nonexistent citations, in which case you should doubt the validity of the article
Both ways, you have better options at your disposal than just linking to the Wikipedia page. You can either read the primary sources cited there - the papers, the code behind the models, the diagrams of experiments, the data - or you can start elsewhere, or not answer at all.
Some, I'd guess, would argue that it can be tough to slog through a couple of ten page papers when you're going to be writing maybe a three or four paragraph answer, especially if they're dense, esoteric, and hard to read. I get that. I love to read, but there's only so much I can do in one sitting. But my response to that is that if you're not willing to do the research, or if you can't understand the papers themselves, you should not be writing an answer. If you cite a source and base your answer on it, you should be able to repackage some of that information into a more palatable form for the curious passerby. If you don't know enough to do that, you shouldn't write the answer.
Let me discuss an example. Last spring, a popular question hit the Hot Network Questions list. One answer (which I won't discuss the specifics of) quoted verbatim the abstract of a paper, linked to it, added a one sentence explanation, and promptly got over 60 upvotes, bringing it to the top of the page (note: this was not a hard-science question).
A few people questioned its validity in comments on it, myself included. A couple of us - literally, two - downvoted it, and most moved on. I did a bit of digging, because the answer was so absurd that all my intuition said it must be false. Two months of intermittent research, coding, computing - easily the most work I've ever done on a Stack Exchange post - showed me that the paper was from a vanity publishing firm, the authors copied part of it from Wikipedia without citation, they know nothing about the field, misapplied half their equations, and contradicted 70 years of scientific results.
It turns out that the writer of the answer hadn't read the paper he or she cited, had no idea how the authors reached their conclusions, did no background research on the subject and did not bother to check the claims therein. Yet the net score is over 60.
This is an example where someone didn't check the source of the information they were citing. The same has happened - though, in my opinion, less dramatically - to answers citing just Wikipedia. The authors didn't bother to check the references of conclusions therein - if there were any references - and wrote answers that were completely wrong.
This is why we need to not just cut Wikipedia from being a valid source for the hard-science tag, but to get people to reference primary sources more. This means that people will actually have to check the information themselves, learn (or show that they already know, preferable) the concepts and ideas involved, and ensure that their conclusions are valid. If not, we're going to get answers that are not even wrong, as Wolfgang Pauli would say.
Plus - perhaps even more importantly - as I said earlier, there's no excuse for citing Wikipedia. Either you read the primary sources and cite those, or you don't reference the article because it's of dubious validity. It's not a hard decision.
So I vote that we not allow Wikipedia to be used as a main source in hard-science answers. If you use Wikipedia, it just doesn't count. I also suggest that we change the tag wiki, which contains the section
The answers should be based on current, undisputed science. This means no subjective sciences. Ideally, answers should be backed up by equations, relevant theories, and citations where possible - arXiv can be quite good for citations, though Wikipedia is usually OK too.
I propose we change this part to
The answers should be based on current, undisputed science. This means no subjective scientific ideas. Ideally, answers should be backed up by equations, relevant theories, and citations where possible - arXiv can be quite good for citations of preprints, although they may not be peer-reviewed.
Wikipedia is not acceptable as a main source in a hard-science answer because it can be inaccurate or poorly referenced; answerers should go to the effort of finding the original papers on a topic, reading them, and citing them in an answer.
Perhaps we could also add something on popular science, which I wrote about in this answer.