Well, if you did find just one rabbit in the Cambrian, you would probably be better off assuming that the fossil somehow got forced into older rock, or that the dating is wrong, or something, than that the whole theory of evolution comes tumbling down. In fact I'd be surprised if there aren't at least a couple of fossils somewhere whose placement biologists can't currently explain.

But I'm a little more familiar with physics, so I can give what I think is a better example there. Modern physics is faced with some big problems; inconsistency between the standard model and general relativity, dark matter, dark energy... but physicists don't say "oh no, the basics of our theories are refuted, back to square one". They say "we need to refine our model, in some unknown way, and in the meanwhile we'll assume without any theoretical justification that the standard model applies on the atomic scale, general relativity applies on the galactic scale (with the addition of some 'dark' fudge factors), some ad-hoc mix applies to things like black holes, and we don't know what applies on the Planck scale".

I'm not saying that falsifiability isn't, at least, a good thing to have, as a rule-of-thumb. Popper noticed something important. Other people have proposed other demarcations; Lakatos, for example, argued for judging 'research programs' rather than static theories, and that a 'progressive' research program is one that continues to produce novel predictions. Personally, I lean towards the view that there is no way to delimit science and non-science except sociologically; you just have to ask whether something is true or not.

Regardless, the point is, it's a lot more complicated than "no single piece of evidence would cause people to abandon this theory" => "this theory's proponents are not scientists".

Look, I'm not a climate scientist. I can't cite, off the top of my head, the number of degrees of warming expected after x number of years with y ppm of CO2.

My point was neither to get into a scientific debate about climate models, nor to argue that "AGW gets a free pass". My point was to call bullshit on "it's just simply not science" and the accompanying pop-Popperism (which is really prevalent on the Internet). You can argue with the conclusions all you want, but to say that it's "not science" because it doesn't fit some idealized picture reducible to a single principle is silly.

"arguing that peer review makes something patently true is absurd"

... which is why I didn't. Honestly on this point I have no idea where you're coming from.

I really wish Slashdotters would stop making arguments premised on a mish-mash of different "definitions" of science half-remembered from one source or another. Defining the scientific method in general terms is actually a really hard problem, which philosophers of science, and practicing scientists with an interest in philosophy, have struggled with for a century without coming to any sort of consensus. (It's known as the "demarcation problem", meaning the demarcation of science and pseudo-science.) There are no easy applications.

The parent - like many people - refers to Popper's "falsifiability" criterion, but nobody who specializes in the subject accepts this criterion anymore in any simple fashion (Popper himself was more complex than most of his internet would-be followers). These are the problems: Every real scientific theory, from physics to biology, has to deal with one or more falsifiers throughout its existence; there are always unresolved problems, apparent pieces of counter-evidence, inexplicable observations, mathematical inconsistencies or unwarranted assumptions. But on the other hand, for any given falsifier, someone can always come up with some sort of explanation which preserves the original theory, in the worst case by either dismissing the evidence as necessarily instrument error, or by modifying the theory in an ad-hoc, one-off way. And more careful, less idealized studies of actual scientific practice have shown all kinds of complications; for example, for a century after Copernicus the Ptolemaic model fit observations better.

So the assertion that climate modeling is "not science", because, given the unsupported assertion that climate modelers don't look for counter-evidence, it doesn't fit some abstract idea of what science should be, is worth pretty much nothing. In general, anyone who writes that they can dismiss some field of study practiced in research universities and published in peer-reviewed journals as "not really science" on the basis of a one-paragraph description of what science really is, is talking out of their ass.

On the specific question of anthropogenic global warming. As anyone who pays any attention to what climate researchers actually write knows, neither "warm winter" nor "cold winter" is a claimed prediction of the models. The predictions take the form of an average global temperature rise over a period of years, or a set of possible average temperatures given various possible levels of carbon dioxide emissions. And James Hansen's models from the 1980s are looking pretty good today.

It's irritatingly hard to find, but there is a way to report it. http://mail.google.com/support/bin/answer.py?hl=en&answer=50270 My significant other's account got hijacked yesterday and Google did react less than half an hour after we filled out the form.