Yes, that's why we abandoned the pseudo-science of medicine ages ago. Oh, wait...

Given the little data we have, psychology is 'average'. We won't know if they're doing exceptionally well, or exceptionally poorly, until more studies are done not only on reproducibility in psychology, but in other fields as well.

Reproducibility problems aren't often investigated, and very few fields are actively studying the issue. I suspect that we'll find serious problems in virtually all branches of science as these studies continue. Nature has already taken action. I expect this crisis to hit even physics which is certainly not immune to controversy.

There's also the question of fraud, to which no branch of science is immune. It would be difficult to determine, but very helpful, if reproducibility problems could be divided between methodological problems and fraudulent or falsified results. It's difficult enough to stop computer generated articles from slipping through. How much more difficult would it be to find "real" papers with falsified data?

If nothing else, this should stress the importance of replication in all fields. Scientists are humans, after all, not the purely objective machines you imagine them to be. It's a dangerous belief, often held by non-scientist "science fans", which ultimately undermines the whole enterprise in the minds of the public.

Like epistemology. Oh, wait. That's absolutely essential to modern science. My bad. I didn't mean to disturb your delusion.

What amazes me is that someone with no scientific background can make such absurd statements about a field, of which they have absolutely no understanding, with such conviction.

Nonsense is nonsense, regardless of the intent. There's little more dangerous to the public understanding of science than bad arguments and nonsense offered in defense of science.

Here, I say that if it's settled, then it isn't science as science can not be 'settled'. Science wouldn't work if such a thing were possible.

Equally, science is indeed frequently overturned. It is designed to be overturned! It could not progress otherwise. Further promulgating the myth of successive refinement (like the person to which I replied) is deeply harmful as it implies that science leads ultimately to truth. (Which, as you know, is impossible.) It sounds like a nice thing for people to believe, sure, but such a fundamental misunderstanding is deeply harmful to the public understanding of science. What would they think, then, when science is necessarily overturned as it advances?

I don't see why you think it's overturned infrequently. Even foundational aspects of science, in nearly every branch, have undergone significant revisions even over just the last century. That's a good thing. It means science is working. If more people understood science, they'd know that it's a positive thing as well. Hiding that fact, which you seem to find uncomfortable, just to get more people to "believe in" or "trust" science isn't helpful. They just end up believing in nonsense they mistakenly call science.

Yes, I'm a terrible person. Of course, my apps don't have serious performance issues. Deal with the devil and all that.

That's not even a little bit true.

It's a popular operating system developed by Microsoft, a large software company headquartered in Redmond, WA.

As this is the first time you've heard about that company and their products, you have no reason to care.

Then it's not science.

You're not familiar with science, are you?

Science cheerleaders like you have done more harm to the public understanding of science in the past decade than every creationist in history.

The options are: 1) Use a library that harms both performance and readability 2) Don't use a library and enjoy better performance and readability.

With jQuery, you save maybe a few minutes of development time, assuming the developer is already familiar with the library, but it costs you a lot in application performance and maintenance costs. The few extra minutes it takes to do things right in the first place quickly pays for itself.

It's all a matter of philosophy. There's been a bit of churning the past few decades, but I prefer this simple differentiation: A theory is a predictive model, an hypothesis is a testable prediction. The utility of a theory is then determined by the success of the hypotheses it generates.

Don't use just one bloated library, use two!

To your examples
1. If you're really that slow, write a single, very short (If performance is an issue,

Performance usually doesn't become an issue until you start using jQuery. Nothing kills your UI faster than that nightmare.

Learn a bit more about the topic and you'll understand.

I'd rather do things right. Sacrificing performance and, in the case of jQuery, readability to save a few minutes seems ridiculous.

It's not 2008 anymore. Get with the times.

Let's take FireFoxOS as an example. I have a ZTE Open, the lowest of the low-end, running FXOS 1.2 -- an older, slower, version of the OS. The only advantage is its excellent support for web standards.

There are some awful examples of HTML5 apps on the platform, notably the popular solitaire game offered through the marketplace. There are also exceptional apps, that you'd think were native if they were running on any other platform, such as fast-paced 3d games and physics-based games that run without slow-downs or stuttering at a high frame-rate. Seeing what's possible, there's no excuse for far less demanding apps to perform so poorly.

Now, you do take a performance hit when you use tools like PhoneGap that purport to abstract away differences between platforms -- doubly so when you add-on ridiculous libraries like jQueryUI, which are known to slow-down your UI dramatically. Don't be stupid, don't use awful third-party libraries, and avoid unnecessary layers of abstraction and you can have a fast and responsive HTML5 app.