HTML and JavaScript dominate because of their openness and easy of transfer of knowledge. If you see anything cool that you want to know how to do, you just check out the source, that openness is why the web works.

That openness is also why search engines can crawl in and index things, something RIA's don't allow. RIA's have their place, but they'll never succeed big on the open net precisely because they themselves aren't open.

There's only one standard open method of building RIA's, Ajax! No closed proprietary technology can compete against that. They may be technically superior, but they aren't socially superior, and that's what matters most.

Additionally, right within the Flex dev environment there is an option to publish the source so a user can view it. Adobe is cool enough to automate this task where the end result is LiveDocs style documentation for your app. This is a decision left to the developer, not the underlying architecture of the shell it is running in such as AJAX/CSS/HTML within a browser which don't give developers a choice.

Lastly, do you actually think that Zoho(or any other AJAX/CSS/HTML application) will succeed or fail because one can view source? The users that matter just don't care.

I said RIA's have their place, they'll succeed a small amount on the desktop where they can be forced upon workers, but on the open internet where they have to compete against web applications, they'll get their butts kicked, as they always have, because they don't force openness.

Allowing the developer the option to hide the code is a fatal flaw, people learn by modifying examples, and that's what makes HTML/JavaScript so unbeatable and viral. When a teenager attempts to learn to program, he'll choose HTML/JavaScript over any RIA platform precisely for that reason, the barrier to entry is much lower. Given the choice to hide their code, most developers will and that creates a barrier to new developers learning the technology.

RIA's aren't new, the idea isn't new, this is what Java was all about, applets, ActiveX components from Microsoft, and Flash from Adobe have all been able to do this for years as well, and they all failed miserably to gain wide adoption. Flash found a niche doing video, but beyond that most developers hate it.

This new batch of RIA's will fail for the same reason, in the end, users don't want them, developers do, to protect their intellectual property. Users don't give a rats ass about a developers intellectual property.

Users are perfectly happy with web apps that use standard controls they understand. RIA's give developers too much freedom to deviate from the normal conventions users are accustomed too. HTML's constraints aren't a weakness to be escaped from, they're a strength that forces you to focus on providing value to the user rather than the mental masturbation of playing with and using non standard widgets that in the end mostly just confuse the average user.

RIA's will always have their place, and it'll always be secondary to HTML/JavaScript, unless an open standard is developed for RIA's that isn't owned by any one company and doesn't require any plugins to run, and doesn't allow developers the ability to hide the code from the users.

But the adoption rate of a RIA depends a lot on the platform you use for development. Gmail is a succesful RIA and it uses JavaScript and Flash. The advantage of sticking on to JavaScript/Flash over Sliverlight is that, they are already very popular. If efficiently done JavaScript and Flash can reside side by side delivering a great user experience. So for the time being,because of Silverlight's low adoption rate,a SilverLight only product may not be a great idea.

Regarding Adobe Flex, Flex its nothing more than another development tool for churning out swf files.The end product is plain old Flash. Adobe Flex allows rapid flash development for those unfamiliar with Flash CS3 (CS3's timeline concept doesn't appeal to most developers). Thanks to Adobe's marketing efforts,most people mistakingly think Flex is something RIA exclusive,while its just a development environment that churns out Flash.

SilverLight does solve a lot of headache and reduce learning curve for the traditional .NET developer when it comes to building RIAs.But from a enduser's perspective, it doesn't really bring in any specific advantage. Anything that can be done in SilverLight can be done using Flash or a combination of Flash and JavaScript. I really dont see any enduser specific advantage for using SilverLight, as it is now. Why force users to install new plugins , when the existing infrastructure(JavaScript and flash) can deliver the same user experience?

Adobe AIR is a way to convert your existing Javascript/flash/html based RIA to a desktop app. But this will require the end user to install AIR runtime first(approx 11 MB download). Compared to Mozilla Prism, Adobe AIR is super advanced in the kind of desktop integration it can impart to your web based RIA.A js/flash RIA can leverage on AIR without any tweaks.

In short, use Javascript/HTML as much as possible. Flash when you need to integrate rich media (or functionalities that require realtime user interaction like Chat ) to your js .I wont recommend Silverlight .

Flex allows easy integration of UI concepts like drag-n-drop, neat animations with low cpu usage, integration of streaming media, lowered response times thanks to model caching on the client... and the ease of development should NOT be discounted... Users tend to like an application that exists more than one that doesn't. Ease and speed of development are a huge bonus.

As for the "invonvenience" aspect, interestingly, I've found that the only people who resent Flash in some way are all geeks. The rest of the world is blithely ignorant as to the difference until you explain it to them, and even then they don't care. So, if your primary market is over-opinionated and out-of-date geeks (rather than the more forward thinking ones who are actively writing or learning Flex by now), then definitely don't use Flex.

If, on the other hand, you plan to target the rest of the world, Flex seems like a good bet!

PS: don't use Flex for writing your blog or news site! RIAs are for applications (gmail, gdocs, rescuetime, basecamp), not pure content sites (blogs, slashdot, reddit, NYTimes..).

"But they seem just like a convenience to developers and an inconvenience to many users, and don't really make anything new possible."

Not acknowledging your presumption that it's an inconvenience to users, the above could essentially be said about any AJAX/Javascript framework too.

However, the learning curve for these technologies is always steep. Now not only do I have to know php/asp.net/etc but I also have to know some kind of rich client application.

I see the sweet spot as media player for your site. Take http://www.pandora.com/ for example. This is a good way to use these rich client frameworks.

That being said, when I can't copy and paste text from these application, it really limits their usefulness. HTML/CSS/JavaScript is a pretty good stack. There will have to be something that really compels me to use them before I'll divert the resources of my shop to learn/adopt/deploy them.

Unless you really need the features slow down and evaluate your need for silverlight. Do you need it or do you want to use the new shiny stuff?

I don't even use the asp.net ajax framework, instead preferring jquery for my ajaxxy stuff. Asp.net can be a great tool but I can't really comment on silverlight as I've yet to find a real need for it.

I've started looking into moving an application to flex for two main reasons: #1 to support no javascript. As more and more ajax features creep in, it gets harder to always provide a noscript fall-back. #2 to do away with cross browser compatibility issues.

Some concerns that I have with Flex so far are #1 no html rendering built in...so there is no way to leverage existing static html/css resources without converting to flex ui (I've read that AIR comes with Webkit, but flash in the browser does not.) #2 there doesn't seem to be good support for printing, especially compared to simple print stylesheets. Oh, and the other huge concern is how prohibitively expensive flex developer is.

What these "RIA" toolkits lack is two things:

* a common, unified, pleasant look and feel. The web already has this, largely forced by the own limitations and the fact that developers now know what's sane and sensible, whereas these ugly RIAs have neither of those things and seem to be a haven for poorly thought-out applications and ideas. I've seen Microsoft showing muddled pseudo-MDI interfaces, flashy visual effects all over the place and all sorts, but none of these things actually stood out as something I'd /like/ to use. They remind me more of artist portfolio and promotional sites in Flash that no-one actually likes in the first place.

* a killer app. I've thought long and hard, but I can't actually think of anything significant (and positive) that these applications allow that the current combination of technologies (HTML, CSS, JavaScript and Flash) don't already. Yes, Silverlight provides for easier development, but were any revolutions single-handedly brought forward solely through having a more pleasant development environment rather than anything actually better for the end-user? Perhaps if Silverlight provided 3D acceleration for in-browser games it'd be a different story, as it is I'm struggling.

Adobe Flex is likely to just continue the success that Flash already has but will not make any great strides forward beyond Flash's current popularity (which I personally expect to wane as computer speeds and browsers continue to improve). Silverlight, despite all the resources that Microsoft are pushing into it, I can't see becoming widespread anytime soon, but equally it's unlikely to be abandoned by MS, so may just be left in limbo.

Of course, I probably regret these remarks in 12 months when Silverlight is the next big thing, but only time will tell but for now I remain unconvinced.

If anyone is serious about this, maybe we can do a prototype. Land one major niche-catalog retailer and the rest will follow...

To me, a killer app is usually a single product that creates such a buzz around a new or previously underused technology or concept that causes a sudden almost universal take up by both users and developers and causes long-reaching paradigm shifts.

Google Suggest did it with Ajax, Del.icio.us did it with tags, YouTube did it to online video, Wikipedia for wikis or even the web for the Internet. Not everything becomes popular thanks to 'killer apps' and some innovations are introduced more gradually (such as Flash), but they're somewhat less common and it's generally an uphill struggle before they're adopted.

It is interesting that you suggest that Google Suggest did it for Ajax (I would argue that it was Google Maps that surfaced OWA-type no-page-refresh-updates). However, I think that Ajax, like Flash is a technology. The rest of the items in your list are more ideas/concepts (e.g. making it easy for the average person to use technology for their own purposes). YouTube - easy embedded videos - a difficult problem to solve - (Flash just happened to be the enabler). And we can argue that video is/was the killer app for Flash. Email was the original killer app (and still is).

Maybe there isn't a killer app for Flash/Flex. Since it is more of a technology that supplements existing technology (OS integration) than a paradigm-changing thing.

Given Microsoft's position it's probably a good long-term bet, but I think it's quite early for use by young startups that need every user they can get.

I love developing in Flex/AIR, and I'm working on a large project now where it makes perfect sense to use it. Obviously my hope is that it is very successful.

My experience with Splashup is that a lot of people think it's AJAX, or don't have any idea what it's developed in, which is perfectly fine with me.

How, exactly, does it extend the .NET environment into the browser?

I'm suspicious of claims like this. The integration of two completely disparate technologies usually gets talked up for marketing reasons, but when you try to build something on top of it, turns out to be unwieldy if not unworkable. Kind of like how Krazy Glue never seemed to hold things together the way it did on TV. :)

I should have clarified that I'm not criticizing Silverlight (of which I'm ignorant) as such, just expressing general skepticism based on what such claims remind me of. These things usually work, but only up to a point (well enough for marketing, that is, but not really for production). If Silverlight is better than that, I'm curious to hear more. Googling hasn't helped, because "Creating a New Silverlight Application using VS 2008" is not what I'm looking for.

If you implement a CLR interpreter, by definition you support C# and any other IL languages.

I ask because for the problem I'm working on, a full Silverlight app wouldn't make sense; but if it were possible to use the Silverlight plugin to work around some of the bottlenecks we've encountered for Ajax development, I'd be interested. Given that most of the bottlenecks have to do with rendering in the browser, though, I imagine this is a long shot.

http://kontsevoy.blogspot.com/2007/11/on-flash-and-new-techn...

Adobe is trying to get Flash into the Fortune 500/Enterprise with Flex but I think Ruby on Rails has a better chance (and they are different types of technologies)

http://www.roughlydrafted.com/2007/07/08/gone-in-a-flash-mor...