David Deraedt

With flashplayer10 out in the wild, it's only a matter of days before the release of both FlexBuilder 3.2 and AIR 1.5 -aka Cosmo (nightly builds are already available). Although we probably shouldn't expect much besides the FP10 upgrade, it's still pretty exciting news. I don't know about you, but I still haven't compiled any application for this player just yet. I didn't want to bother tweaking my development environment, and just prefered to wait. Now, I must say I'm getting impatient... :) Here are some thoughts about what we can expect from these upgrades.

AIR 1.5

AIR 1.5 will of course include flashplayer 10. The File reference runtime access is, IMHO, the real killer feature, as far as Flex is concerned. However, AIR apps will obviously not care much about this since the runtime already provide full access to the file system.

Besides flashplayer 10, the webkit engine will also be upgraded to SquirrelFish JS engine, from which we can expect a significant performance boost.

To my knowledge, no changes to the SQLite embedded database or to the Adobe reader have be disclosed.

However, it has been said (french) that this release had some notable bug fixes, like the one concerning the relationship between the runtime and the system browser on some OS.

As usual, we'll have to upgrade our descriptor files to have our apps leverage the latest runtime features. ADM, my latest application which helps you manage your AIR application descriptor files, already provides an "upgrade" feature to automatically change the namespace from 1.1 to 1.5. As soon as AIR 1.5 is out, it will also deal with corresponding the corresponding changes in the descriptor files (I can't disclose more on this since I'm under NDA and I'm not certain exactly what informations have been made public).

Flex SDK 3.2

This release will be the first step of the Marshall plan.

I haven't heard of any specific new feature or significant change regarding the SDK or the framework (besides the usual bug fixes), so the main changes will come from flashplayer 10.

The new Vector type (think typed Array), in particular, is something that may dramatically improve both performance and coding comfort, just out of the box, for any kind of application. I've waited so long for this, now I can't wait to try it!

We shouldn't forget about the other cool little additions, such as the much needed unloadandstop method (which will help us deal with some memory performance issues), dynamic streaming (which should improve the overall media streaming experience) and Speex audio codex (which will help implementing audio conferencing solutions). I don't know what to think about RTMFP, since I understand it will only be supported by FMS.

I'm a bit skeptical about the pseudo 3d features, at least if used "as is". I imagine we'll have to wait for the oh-so many AS3 3D engines to incorporate it to see its real benefits. Just like the new hardware accelerated graphic modes and Pixel Blender filters, I'm afraid those will not concern my enterprise application developments. (However, I hope the artistic software I write will use it).

The new dynamic sound capabilities seem to disappoint so far. That's too bad.

I've put higher hopes into the new text rendering engine, although I guess that, like many other features, we'll have to wait for the next iteration of the Flex SDK (Flex 4 aka Gumbo) to have the framework really leverage those features, and actually see the benefits in our apps. As for the new drawing API which, I suppose, will be used internally for FXG.