06Oct Adobe AIR Wishes
1,867 views
While trying not to come off as an Adobe fan-boy, but Adobe AIR is truly a revolutionary desktop software platform and has the potential of going toe-to-toe with some of the markets that Visual Studio has dominated since the dawn of time. However, there are some key features that I believe could boost the power of AIR and swoon other desktop developers over:
- Direct access to USB and serial ports… this would be a massive boon to manufacturing developers who have to interface with machines
- Standalone exe option (granted there will be a higher file size since it has to contain the runtime)
- Bundle non-embeded resources into .AIR packages
- Native mysql (or like PHP’s mysqli interface) support
- Bolster performance (maybe add multithreading) so that AIR could be a viable server platform
- Optimize memory usage of the runtime (sometimes it goes mad and consumes over 300mb)
- Update to the newest Webkit (much faster now)
- AIR has very poor math performance for any desktop language… this should be improved. No company is going to embrace AIR for doing engineering or other mathematical simulations when its 10x slower than even Visual Basic.
- Game development could really really use more GPU feature support
- What else have I missed?
As it stands now, it seems like Adobe AIR does not contain any one end-user feature strength versus their competition… most of it has been developer orianted. (aside from the runtime application updating)


October 6th, 2008 at 8:42 am
FTP?
The ability to open a file with its associated application?
That’d be enough for me!
October 6th, 2008 at 8:49 am
Amen to the serial/usb port access! The Merapi project aims to solve some of these issues, but you still have to have some Java knowledge to get that to work.
October 6th, 2008 at 9:16 am
Hey Dan, Merapi is a hot project indeed (pun), but it’s an extremely resource intensive solution for applications. Both Java and AIR runtimes can consume 200mb each… this is terribly heavy for middle-sized applications. Ideally, Adobe needs to build the appropriate features into AIR itself.
October 6th, 2008 at 9:34 am
Questions:
1. Direct access to USB and serial ports… this would be a massive boon to manufacturing developers who have to interface with machines
Response: I agree that this would be useful, but this would have to be developed in a consistent cross-platform manner. In all honesty – the example that you give (interfacing with machines) is still certainly a situation where a native application makes sense. Most manufacturers wouldn’t need cross-platform compatability, and direct interaction with web services is probably not on their radar.
2. Standalone exe option (granted there will be a higher file size since it has to contain the runtime)
Response: There certainly would be some benefit to this, but I believe that over the next year AIR will become extremely commonplace, and the runtime installation will not be as big of a concern as it is now (especially with Adobe’s efforts to ensure that enterprise distribution is viable).
3. Bundle non-embeded resources into .AIR packages
Response: Currently, non-embedded resources can be included into AIR packages. They just have to be added to the compiler arguments (and the Flex and Flash tools include GUI’s for this).
4. Native mysql (or like PHP’s mysqli interface) support
Response: I understand that several people want this, and in some ways it would be a cool idea. I think this could lead to problems. It is always best to have a service layer if possible. Also, with tools like SimpleDB from Amazon (a database that does not require a service layer) you can now access a database directly from an AIR application without having to write a custom adapter (SimpleDB uses webservices).
5. Bolster performance (maybe add multithreading) so that AIR could be a viable server platform.
Response: Multiple threads will be possible through FlaCC which will allow the Flash Player / AIR runtime to execute compiled C/C++ code:
http://www.peterelst.com/blog/2008/10/04/what-is-flacc-all-about/
6. Optimize memory usage of the runtime (sometimes it goes mad and consumes over 300mb)
Response: Absolutely!
7. Update to the newest Webkit (much faster now)
Response: …
8. AIR has very poor math performance for any desktop language… this should be improved. No company is going to embrace AIR for doing engineering or other mathematical simulations when its 10x slower than even Visual Basic.
Response: The previous mentioned FlaCC addresses this as well. Compiled C/C++ code will be executed quite a bit faster than ActionScript (some estimates state that it can be 10x faster).
9. Game development could really really use more GPU feature support
Response: Agreed. The FP10 support in AIR will certainly be a step in the right direction.
David Tucker
October 6th, 2008 at 11:40 am
- dramatically improved printing
- classes libraries should be included like: save .pdf .swf .flv .mp3 .bmp .mov .xls .doc …
- more native desktop access like zinc…
October 7th, 2008 at 2:17 am
I vote for: The ability to open a file with its associated application !!!
December 17th, 2008 at 3:57 am
(some estimates state that it can be 10x faster) … i think that on their website is written it will run 10x slower