thanks. what development tools do you use to develop JS applications?

Oh just the typical stuff. I'm old school so...

Your favourite text editor(I use vim) or you can be a bit more fancy and go with something like Netbeans or Eclipse(although I always go back to an editor). And firebug for debugging.

You can add a bunch of stuff to the mix, but that's all I need really.

(Oh, you might also want to throw in a mini-fier for when you're done.)

ok yes I know about those libraries...I guess what Im asking is why use GWT when such libraries exist?

If you have a piece of algorithmic code that really needs to exist on both the server or the client, and maintaining two versions of it will be too much of a pain. That's why Wave uses it (the Wave team would probably give you other reasons too, I've heard they like it, but that's the only technically defensible one IMNSHO).

If you want to have the same code run on both the server and client, your only options (that I know of) are GWT, server-side JS like Rhino or V8 or Node.js, and Pyjamas. GWT's the most mature of those.

Also, a social reason to use GWT (and perhaps the real reason Wave uses it) is if you know Java and don't want to learn JavaScript. This doesn't hold much water with me - what sort of a programmer wouldn't want to learn a new language? and JS is a pretty nice one all things considered - but it's convincing for lots of other folks.

This.

One of my developers is crazy about GWT, and has shoe-horned it into a number of apps that he's primed, and it's proven to be problematic for a number of reasons.

The main problem we have is that a lot of our apps fit into a larger portal, so we have to worry a LOT more about JavaScript and CSS collisions, and generated JS/CSS just doesn't work well for that. Outside of a portalized environment, it might not be THAT bad, but for us, it's proven painful.

In the event that you find yourself duplicating your server side code and client side code, or writing excessive form validation to match server-side constraints, something like GWT is probably the way to go.

thanks for the great response

Economics can contribute to economics, so long as they are not dynamic macroeconomic theorists. That stuff is witchcraft.

huh?

They are called the "Fundamental Theorems of Welfare Economics" - look it up in any graduate econ text. Stop ranting.

You'll know when I rant. :)

Economics in the 40-50s was about Markets vs. Planning. The big economic names at the time - Hayek, Keynes, were debating the merits and faults of markets and planning. Einstein threw in his 2 cents. That he was a physicist has no bearing on his argument.

Exactly. And yet the only reason this essay has been dug up is because the greatest physicist of the day wrote it.

Hayek, Keynes, et al discussed the economics - the models, the math.

Einstein here makes a political argument, complete with hand-waving about how economics can't be used to judge his position because it's so last-wave.

This is the political version of Linus Pauling and vitamin C .

Actually Hayek and Keynes never used much math in their arguments. Keynes ideas were put into economic models by later economists with science backgrounds: Alvin Hansen, Paul Sameulson, James Tobin, etc.

Quit pointing out that Einstein was wrong. Thats obvious. The point is that a physicist can contribute to economics. Had the original poster simply pointed out the flaws of socialism, this discussion would never had taken place.

Edit: original poster = I mean the guy who said JM Keynes on QED - had he pointed out socialism flaws this discussion would have never taken place.

You are far out of your depth; this'll be my last to you.

"Quit pointing out that Einstein was wrong. Thats obvious."

Please point that out to the people trying to sleaze in the claim he was right. :)

"The point is that a physicist can contribute to economics."

To quote another physicist, you're not even wrong!

"ad the original poster simply pointed out the flaws of socialism, this discussion would never had taken place."

Well no - gnosis apparently would have never submitted this, judging by some of his other remarks.

Good thinking.

depressing =(

I dont think Mac OS X server is really anything special - Mac OS X itself can be a good server. Unless your using some specific Apple-only technology like Xgrid, plain OSX is fine. And Id only use OSX as a server over linux/bsd if you really want a GUI to administer your server as opposed to command line. The GUI really is the draw of OS X generally over Linux.

Cool idea but a rather expensive way to go (if you consider hardware + colocation fee) compared with a VPS. Although the mac mini is dedicated its laptop parts make it on par or worse than a good VPS performance-wise.

Id be more inclined to use a mac mini as a server on a local network as opposed to colocating. Another interesting application is a cluster of mac minis-that a whole other area.