The elephant in the room is that no matter how much Douglas Crockford apologizes for it and tries to brush the fact aside, JavaScript the language, the browser DOMs, HTML, and CSS make an absolutely terrible platform on which to develop large, complex, and well performing desktop-style apps.
I can't be the only one who thinks this, and it seems like Google is the only one to have any success at all in doing this on a large scale.
I think it's something I heard him say out loud in one of his earlier lectures at Yahoo. More recently, he's stated that "he used to think the browser was the most hostile programming environment ever devised, but then he found out about mobile programming" - http://www.javaworld.com/javaworld/jw-09-2007/jw-09-rwekeyno...
I've heard Crockford use that quote in a few Yahoo! talks (when I used to work there). It's probably on one of his slides from either: "Javascript: The Good Bits" or one around the same time.
You are absolutely correct but you picked the wrong venue to express this view, hence downvoting.
From a purely technical perspective web platform is approaching DOS circa 1990 accelerated by sheer amounts of open source libraries available today. But it still provides pretty much DOS experience albeit on a high-res screen.
I am not sure it's a bad thing though: when you give developers absolute freedom over your hardware they will quickly turn your PC into a typical popup-ridden WinXP box with tray area half-screen wide :)
I'm pretty sure you haven't developed for DOS in the 90ties, otherwise you wouldn't be saying that.
People that bitch and moan about HTML, CSS, Javascript and about choosing a (high-level) language/framework for server-side development ... should really try developing for DOS, having to deal with Turbo Pascal or C (since C++ in the early 90ties was a mess), having to drop to 286 or 386 assembler to do their job, having to deal with the conventional memory limitations, going through hoops choosing between extended and expanded memory, or having to deal with the protected 386 mode, and always having to reinvent the wheel because open-source practically didn't exist for DOS ... like writing their own windowing tookit, or simulate multi-threading / background workers with hardware interrupts.
The truth of the matter is, we've never had it better. People bitching and moaning about web-dev are just conservatists ... but really, they should get over it since this is 2010, not 1992.
I don't know if they are conservatists or simply too young to understand what it was really like back in the "old" days.
The web is awesome. It's the most amazing technology platform the world has ever known. HTML, CSS, and JavaScript are all standard and interoperable.
People who complain about CSS should try changing the look/feel of an entire desktop app. People who complain of HTML should try building PDFs or Word Docs on the fly. People who complain about JavaScript should try writing some Assembly: DOS, Sparc, what have you...
People who complain about web programming should try to deploy an enterprise app to 10,000 desktops. It's a nightmare. The Web is a blessing to the world. Thank you Tim!
People who complain about CSS should try changing the look/feel of an entire desktop app. People who complain of HTML should try building PDFs or Word Docs on the fly.
Actually, those two capabilities were two things that VisualWorks Smalltalk has had for something like a decade? Want your App to look OS X Aqua? One line of code, boom! Building apps on the fly? Business as usual for Smalltalk.
(To be fair, the Aqua/Motif/Windows emulated UI was half-baked appearance wise, but the underlying framework was good.)
The iPad/iPhone helps with one aspect of the problem: There is only one platform to handle. No need to support broken old browsers.
[by the way - I haven't seen any statistics but it seems Apple users are more likely to update their OS/Apps to their latest version. So supporting old versions is less of an issue]
I can't be the only one who thinks this, and it seems like Google is the only one to have any success at all in doing this on a large scale.