Yep, passing up seeing Danger Mouse for free b/c I’m too busy. What has this world come to? Quick RIA linkywinks:
- Macromedia Flash Player Version Penetration – Flex requires Flash 7; as of Dec 03 was ~30%, although MM claims 80% by the end of the summer (for comparison, Laszlo only requires Flash 5+, which is at 96% penetration)
- Case Study: The Development Life Cycle of a Flex Application
- Summary About RIA After Evaluating Various Solutions – EXTREMELY valuable evaluation of current RIA playing field
- clarity – jimmy joe says Flex is worthless
- OpenAMF – open source Java Flash remoting
- Web Services vs. Remoting
- Remoting vs. ASP.NET performance
- Performance Comparison: .NET Remoting vs. ASP.NET Web Services
Hmm, so my take on the whole thing… The RIA-space is pretty targeted to the corporate market right now. If Macromedia can leverage the Flex platform (say, allowing compile out to Flash Lite) they could have a winner. If not, they’re going to get slaughtered come ’06 when XAML comes out. Two years is a long time however, more than enough for the FOSS community to get together and put something together. The Flash-based solutions (Flex, Laszlo) are really compelling because, well, because the Flash plug-in is just waaaay faster and better at rendering custom widgets/graphics then JVMs out there.
Now, for the rest of us (WWW), I’m not sure how much of an effect this will have in the near future (next 2 years). Laszlo has been around for the past 2 years, and groups have tried to push pretty decent webapp clients (XWT, Thinlets) with pretty limited success. Does this have to do w/ not being tied into server-side frameworks? Just that it’s easier to develop REST-based web apps?
(to interrupt for a second: for web applications (vs sites), the arguments of Google searchability, bookmarking, etc. don’t really apply; most of the time those ‘features’ are in fact liabilities that need to be worked around)
While I’d like to think that there’s a chance for DHTML to actually fulfill its promise, with 3-years of 90%+ modern-browser marketshare, we’re still pretty much nowhere. I think actually, a lot of it can be tied to the immense popularity of scripting languages like PHP (scoff all you want, it’s the most popular web development language on the web; and despite its deficiencies, not without good reason) — and the failure of anyone to create a framework that makes it as easy (or easier!) to build a more robust interface w/in PHP. The pieces are there: mod_pubsub, jsrs, a gazillion widget sets, windowing kits, but none of it seems to tie together in a complete, easy to use, integrated (w/ the backend) way.
Some thoughts on tying it all together:
- integrated PHP/JS ‘application state’ engine (handling noncing, transactions, etc)
- transparently translate REST actions into remote scripting style actions
- smart/easy complex gui-widget data binding
- dynamic key/event binding
In theory, properly designed and constructed, this could *probably* be made to degrade into traditional REST if proper JS support isn’t found…