Switch from Flash to HTML5 for a new version of a large project?

We have a large web product with 11 million users (widgets that are embedded into third party web sites, and now adding iOS/Android apps开发者_如何学运维). Most of our core interface is written in ActionScript 2.0. We are planning to release a revamped version. We are now facing a choice of porting it to ActionScript 3.0 or rewriting it in HTML5. We already have about 30% of our code written in AS 3.0. Can you help us make this decision?

HTML5

Pros

1. Future proof
2. Works in browser in iOS
3. No plugin required, presumably faster loading
4. Native browser rich text rendering, including spell check and copy/paste
5. Appealing to webmasters (contemporary technology)

Cons

1. Product relies on WebSocket - no standard yet, not available in IE
2. Highly custom UI, not consistent across all browsers
3. No video/audio hardware interaction for stream capture
4. Harder to convert into iOS apps (e.g. need hybrid apps, UIWebView cannot easily call functions in the native wrapper)
5. JavaScript is like AS 1.0, less sophisticated than AS 3.0
6. More work for us given our existing codebase and skills
7. Harder to embed JS widgets into third party sites than Flash widgets

Flash

Pros

1. Already have some code base, experience
2. Video/audio stream capture, socket connection
3. Good for widgets on third party websites
4. Consistent UI across browsers and platforms
5. Easier to use the same codebase for AIR apps on iOS/Android

Cons

1. Adobe's commitment to Flash is unclear
2. Adobe feels free to force developers to change technologies often, unreliable : AS1.0, AS2.0 have been phased out, Text Layout Framework is not supported on mobile and is being phased out, not clear if our AS3.0 codebase will still be relevant in 3 years)
3. Shoddy support of Flash on Android (e.g. text input is broken in 25% of devices)
4. Does not work in browser on iOS and Metro, so need a lite HTML5 version anyway
5. Slower in browser on Android
6. Public perception of Flash as obsolete?

We are in an identical situation and I truly wish that Adobe would take the lead in answering this very question; Flash or move on?

We have an AS2 Flash RIA (RIA, a phrase from the past?) that allows our clients to edit their sites from their browser, similar to Wix. Although the sites are viewable on iOS devices via ColdFusion (example: dolcedesignsonline.com - try in IE then in an iOS borwser), we are wondering what to do next. We are in the middle of an AS3 version of our RIA and although nothing in HTML5 will ever come close to the functinality we now enjoy at our fingertips in AS3, is AS3 going to there down the road?!

I do not think that Adobe would just let flash die. A more logical next step would be make the Flash IDE open source and free to the public. I am not endorsing that idea and I do not see that happening but I believe that would happen long before Adobe would just close the doors on Flash altogether.

Tick tock... at some point we are just going to have to flip a coin.

John Glynn

My thoughts on your points:

HTML5

Pros

Future proof - this is also true of Flash (you can still run an AS1 movie in FP11)

Works in browser in iOS - but webGL requires iAds, and canvas knackers the battery, in that sense support for actual HTML5 features is pretty limited

No plugin required, presumably faster loading - depends what's loading!

Native browser rich text rendering, including spell check and copy/paste - good points, not sure what/if any support these have in AIR

Appealing to webmasters (contemporary technology) - depends on the webmaster, beside Canvas/webGL it's just plain old HTML/JS

Cons

Product relies on WebSocket - not sure how you could get around this

Highly custom UI, not consistent across all browsers - this will always be the bane of web developers - look into jQuery UI if you haven't already

No video/audio hardware interaction for stream capture - correct AFAIK

Harder to convert into iOS apps (e.g. need hybrid apps, UIWebView cannot easily call functions in the native wrapper) - correct, but it is possible to do

Javascript is like AS 1.0, less sophisticated than AS 3.0 More work for us given our existing codebase and skills Harder to embed JS widgets into third party sites than Flash widgets - JS is a very cool language but it's not really suited to building large applications which will need to be maintained by many people, debugging it can be an absolute bitch.

Flash

Pros

Already have some code base, experience - code can be organised into classes, easier to maintain and debug, AS3 also has most of the standout JS features (closures, objects, 1st-class functions) as it is derived from it.

Video/audio stream capture, socket connection - true

Good for widgets on third party websites - not sure what relevance this has to you, as not entirely sure what kind of application you're making

Consistent UI across browsers and platforms - to me, this wins by itself!

Easier to use the same codebase for AIR apps on iOS/Android - true, related to above

Cons

Adobe's commitment to Flash is unclear - this is true, but Flash player will be around for years, if not decades, yet, Stage3D has breathed a new lease of life into the platform

Adobe feels free to force developers to change technologies often, unreliable : AS1.0, AS2.0 have been phased out, Text Layout Framework is not supported on mobile and is being phased out, not clear if our AS3.0 codebase will still be relevant in 3 years) - This isn't really true as the Flash Player provides legacy support for everything ever built in Flash, it might make no sense to code in AS1 now but if you do, you can still export and run on any machine with Flash Player.

Shoddy support of Flash on Android (e.g. text input is broken in 25% of devices) - this is the major drawback, I'm not sure how well AIR addresses these problems but in my experiments with it, the only issue I've found with AIR on android is performance for intensive graphics.

Does not work in browser on iOS and Metro, so need a lite HTML5 version anyway - This may be insurmountable unless you can persuade your clients to use an app (shouldn't be too hard whilst the buzz remains around the word), then you can use AIR - and your existing codebase!

Slower in browser on Android - again, you may need to go with the AIR option, even as a Web/Flash developer I've never been over-keen on the idea of it running in mobile browser, at least with current technology

Public perception of Flash as obsolete? - This has been a real pain, but now Steve Jobs is no longer with us I think a lot of the hysteria has died down, people are looking more objectively and realise that his 'thoughts on Flash' were motivated by a desire to protect App-store revenues rather than facts, and that in terms of development speed, maintenance and cross-browser support, Flash is still miles ahead of HTML/5. Provided you can cross the mobile bridge - and that is a pain, there's really no choice but to go with AIR apps - I'd go for Flash.