Flex...anyone know about it?

[orgasmicliving!!]

It's an Adobe product...just wondering if anyone here has used it to develop anything.

[J_S_G]

orgasmicliving!! said:

It's an Adobe product...just wondering if anyone here has used it to develop anything.

Yeah, have worked quite closely with Adobe/Macromedia as we trialled the 1.0, 1.1, and 2.0 releases for building Rich Internet Apps. Just shout if you want any info.

[orgasmicliving!!]

Wow, great. Have been looking at the stuff one can do with beta 3. Absolutely amazing, and the server price is coming down from 12,000 usd to 1,000 usd!

First question:
what do you need to be running on top of Apache to host websites developed using Flex? Java? Some sort of Flex server? Neither?

DB is MySQL, which I believe is fine for the job?

[jazzybee]

The company I work for does rich imaging and is an Adobe/Macromedia partner and we are building our new application interfaces in Flex, as well as bespoke solutions for our customers. I do see this technology, along with AJAX (the open source alternative), being the base tech for the next level eCommerce sites.

In fact, if anyone is particularly skilled in Flex - I am looking for developer/trainer/PM to begin a new service practice - if any one knows someone who maybe interesed.... Hope this does not break any forum rules. Apologies if it does. I will remove this part of the message if so.

[J_S_G]

orgasmicliving!! said:

Wow, great. Have been looking at the stuff one can do with beta 3. Absolutely amazing, and the server price is coming down from 12,000 usd to 1,000 usd! First question: what do you need to be running on top of Apache to host websites developed using Flex? Java? Some sort of Flex server? Neither? DB is MySQL, which I believe is fine for the job?

With Flex 2.0, you don't need to pay for the Flex Server license at all - it's now just used for "Enterprise" features (i.e. direct back-end Java integration - decent asynchronous messaging, streaming video comms, etc). Standard "forms" type stuff needs none of that (and can still talk to Web Services, with a couple of restrictions). For those "normal" applications (ok, so "normal" will be redefined going forward), you can just compile it up and deploy it to the Web server as a Flash movie. For the "Enterprise" stuff (and for Flex 1.0/1.1), you need a Java Application Server - JBoss/JRun/etc. The Flex server then sits on top of this and manages all the communicatino.

MySQL should work depending on what you're running server-side to talk to it. If you don't go for the server license (which is still £12k or so), and just pay for the developer license (£1k or so), you'll need to write the server-side code in anything you fancy - PHP/Perl/ASP/JSP/ASP.NET, and expose the functionality "nicely" (preferably as Web Services) - so whatever OLEDB/ODBC/whatever database connection you care for. If you fork out for the server, then you can do it in Java, and it ties up neater/more restrictively depending on how you look at it. Again, JDBC/whatever to connect to data sources.

Just yell if you need anything else.

[J_S_G]

jazzybee said:

I do see this technology, along with AJAX (the open source alternative), being the base tech for the next level eCommerce sites..

Now there's a long conversation we could have! AJAX, whilst "clever", and clearly a technology that's going to gain some ground isn't a "strategy" to go with - it's a tactical solution. HTML is for structuring documents. XHTML and CSS have finally got that sorted. Ruining accessibility/compatibility/platform-independence (Web-everywhere - mobile/fridge/TV/etc) with scripting to simulate an application in an unsuited technology isn't really the way forward. It's like ASCII/ANSI art that everyone used to do... it was impressive, looked cool for what it was, but it's NOT how to represent images accurately/effectively.

Flex, on the other hand, is a platform designed for building applications. It's still limited by the Flash runtime, which itself wasn't intended for applications. But at least the runtime's been extended, and the platform's abstracted quite a bit away from that. It'll be interesting to see how Apollo takes off.

The whole Rich Internet Application area is an interesting one - whether people get it right to "augment" a decent online experience, or whether people screw it up and miss the online/offline information/application paradigm differences completely, ruining the user experience and positioning of a service/offering.

Then there's Vista... untill the uptake of that (and solutions start getting built in WPF), it'll be very difficult to judge the best online platform for RIAs. Offline/sometimes-online are much easier scenarios.

[orgasmicliving!!]

Very stimulating discussion, as I am a complete noob at this. I thought it was the future, as well. Thanks for the Ascii art analogy, makes perfect sense, as I now see it closer to trying to making websites run on top of the flash runtime player. Is that a correct interpretation of what you are saying?

I must say, though, the stuff I saw at JaveOne was INCREDIBLE. I had a Java developer working for 8 months, designing a medium-sized site. The guy at the Flex platform cobbled together a similar (and, in many ways, very much superior) thing in 1 hour flat.


>> Edited by orgasmicliving!! on Wednesday 31st May 21:29

[orgasmicliving!!]

J_S_G said:

Flex, on the other hand, is a platform designed for building applications. It's still limited by the Flash runtime, which itself wasn't intended for applications. But at least the runtime's been extended, and the platform's abstracted quite a bit away from that. It'll be interesting to see how Apollo takes off.

Hmmm, just read this para again. Seems like you are saying it is better, and therefore less of a tactical solution.

But still not a complete strategy/platform, and still untested against WPF.

Am I understanding you correctly?

[J_S_G]

orgasmicliving!! said:

J_S_G said: Flex, on the other hand, is a platform designed for building applications. It's still limited by the Flash runtime, which itself wasn't intended for applications. But at least the runtime's been extended, and the platform's abstracted quite a bit away from that. It'll be interesting to see how Apollo takes off. Hmmm, just read this para again. Seems like you are saying it is better, and therefore less of a tactical solution. But still not a complete strategy/platform, and still untested against WPF. Am I understanding you correctly?

Yep, that's closer to it. There are fundamental issues Macromedia/Adobe still need to get around. Like local storage (caching for offline running) - something Apollo's meant to fix. Rich applications on top of HTML are like ASCII art... HTML + JavaScript + a non-standard implementation of an HTTP/XML request object isn't a robust development platform! (Testing is a nightmare if nothing else) It has its place, and it's often the only solution given the medium, but it's not the way the world's going. Especially with WPF - bear in mind Microsoft are only backing AJAX (via Atlas) as they need to have a stop-gap before WPF; they need to stop Flex (and Laszlo, etc) gaining too big a foothold before they try to blow both that and AJAX out of the water. (OK, so that's simplifying it a little!)

Bear in mind anything shown at a conference will fit within the confines of what is easily achievable. ASP.NET is a great example - you can have personalisation (drag-and-dock), security, skinning, everything for free out of the box... as long as you don't want to push the envelop. Flex is the same. I tried rebuilding a real-world application and the stuff you get out of the box is great - it built 70% of the application in no time. At this point, two things became very obvious:
1. It's much better for "applets" than "applications" - splitting it up and handling events properly isn't that elegant at all.
2. The other 30% was going to be a nightmare of writing animations and event handling by hand. That 30% would easily take 90% of the overall dev. time and would involve reworking a lot of what you got for free to create anything of practical use.

If you want to go with the standard look & feel (including what you can do with CSS, etc.), and don't want to do anything too complex, it's great. For true brand-alignment, and for really funky functionality, it can become quite expensive - possibly better for those companies that can either build up an internal toolkit, or for agencies that focus on it and build their own to re-use for clients. Truly funky stuff is definitely nicer in WPF.

But you can't argue with 90%+ market penetration... there's a good place for Flex in the medium term.

[jazzybee]

J_S_G said:

Now there's a long conversation we could have! AJAX, whilst "clever", and clearly a technology that's going to gain some ground isn't a "strategy" to go with - it's a tactical solution. HTML is for structuring documents. XHTML and CSS have finally got that sorted. Ruining accessibility/compatibility/platform-independence (Web-everywhere - mobile/fridge/TV/etc) with scripting to simulate an application in an unsuited technology isn't really the way forward. It's like ASCII/ANSI art that everyone used to do... it was impressive, looked cool for what it was, but it's NOT how to represent images accurately/effectively. Flex, on the other hand, is a platform designed for building applications. It's still limited by the Flash runtime, which itself wasn't intended for applications. But at least the runtime's been extended, and the platform's abstracted quite a bit away from that. It'll be interesting to see how Apollo takes off.

Interesting... I have not been at the sharp end of development for quite a few years now, but am having to get more involved again. Very interesting point you make about Flex being a strategic platform for RIAs - I am coming up against AJAX often when talking about increasing richness of websites with Flex. Not knowing enough about AJAX yet - I would like to hear your opinions further on how they differ, as pointing out differences is much better than trying to compare. We are definately commited to the Flex route, with the frontend to one of our Hosted apps already delivered in Flex, and our core ASP application being delivered later this summer.

[J_S_G]

jazzybee said:

J_S_G said: Now there's a long conversation we could have! AJAX, whilst "clever", and clearly a technology that's going to gain some ground isn't a "strategy" to go with - it's a tactical solution. HTML is for structuring documents. XHTML and CSS have finally got that sorted. Ruining accessibility/compatibility/platform-independence (Web-everywhere - mobile/fridge/TV/etc) with scripting to simulate an application in an unsuited technology isn't really the way forward. It's like ASCII/ANSI art that everyone used to do... it was impressive, looked cool for what it was, but it's NOT how to represent images accurately/effectively. Flex, on the other hand, is a platform designed for building applications. It's still limited by the Flash runtime, which itself wasn't intended for applications. But at least the runtime's been extended, and the platform's abstracted quite a bit away from that. It'll be interesting to see how Apollo takes off. Interesting... I have not been at the sharp end of development for quite a few years now, but am having to get more involved again. Very interesting point you make about Flex being a strategic platform for RIAs - I am coming up against AJAX often when talking about increasing richness of websites with Flex. Not knowing enough about AJAX yet - I would like to hear your opinions further on how they differ, as pointing out differences is much better than trying to compare. We are definately commited to the Flex route, with the frontend to one of our Hosted apps already delivered in Flex, and our core ASP application being delivered later this summer.

Bit of a long conversation this for the forums... but I currently work in IT strategy for a well known Internet Bank - have spent a lot of time working on the Web/Rich Internet Application/Smart Client side of things (and the fundamental differences/appropriate platforms for each). Drop me a mail if you want a natter - probably easier done over the phone!

[Plotloss]

Does Flex require client side objects to run like Flash?

I've just been writing an AJAX controlled mp3 player which is fine for PC browsing but falls short on an SSR interface.

Was going to go for Flash/AMFPHP for the PDA client, might have a look at this Flex malarky if its a good fit.

[J_S_G]

Plotloss said:

Does Flex require client side objects to run like Flash? I've just been writing an AJAX controlled mp3 player which is fine for PC browsing but falls short on an SSR interface. Was going to go for Flash/AMFPHP for the PDA client, might have a look at this Flex malarky if its a good fit.

If you think of Flex as being a bit like ASP.NET (or JSP for that matter)...

You write a load of pseudo-XML (or actual XML) code to define the interface, with a code-behind for wiring up events. This all sits on the server, and when a request hits the server, it's compiled, run, and the output sent to the client.

Whereas ASP.NET is ASPX + C# (or whatever) -> HTML, and Java is JSP + EJB (depending on how complex you get/if you bother) -> HTML, Flex is MXML + ActionScript -> Flash.

So, basically, you write a UI in MXML (a custom tag based XML variant with <mx:button />, <mx:image />, etc), and ActionScript (as used in Flash) "code behinds" to handle the events - not the nicest of languages, but 2.0/3.0 is much better. The server just compiles this all up into a Flash movie that contains all the buttons, images, web-service calls, etc. DEFINITELY a better option for writing applications than pure Flash.

A few notes:
* Use 2.0, not 1.1; it allows you to "pre-compile" properly so that you can just copy the SWF movie to the web-server (OK, this was possible in 1.1, but dirty-hacky).
* 2.0 is still in Beta (should be out imminently)
* 2.0 needs the latest (about-to-be-released) Flash player - beta available, naturally
* You get the nice development studio with 2.0 (if you pay/get a trial); built on top of Eclipse rather than Dreamweaver as in Flex 1.1.
* Make sure you don't tie yourself to the £12k a license enterprise server unless ABSOLUTELY necessary (think "don't tie yourself to BizTalk when you can code something yourself in 1/2 an hour", or "avoid MQ Series if you're cobbling something together to run in your bedroom"!)

As ever, just shout if you need info (as long as we don't get into serious cone-head "why doesn't this beta release allow me to re-style the Halo interface to some of the different colour-ways that are meant to be supported?" questions )

>> Edited by J_S_G on Thursday 1st June 00:10

[Plotloss]

Oh, how very clever, I'm with you on the concept fully.

Tried it on the WM5 Flash client per chance?

[Plotloss]

Boo

Back to smashing my head against Minimo for the time being I guess...

[J_S_G]

Plotloss said:

Oh, how very clever, I'm with you on the concept fully. Tried it on the WM5 Flash client per chance?

WM5 Flash player is v7. Flex 2.0 The Virtual Machine used for ActionScript 3 (AS3) has been largely rewritten as part of Flash Player 9, as far as I know, and so will take "some time" to port at a guess. And it needs Player 9 regardless.

Given Flash Lite, and Microsoft's murmerings around WPF-E, you can put money on them trying to get 2.0 working on mobiles, though, going forward.

Now for the good news... Flex 1.x apps DO run on WM5. They're not that fast (animation is dire!), and without a 640x480 screen, it can be a bit painful, but they do run. Marvel of modern technology. (This was on an Orange SPV M2000 with a "custom" install of WM5)

I still wouldn't advise writing 1.x apps, though - totally immature and unready technology (unless you really are talking a few noddy buttons and a bit of async updating).

[J_S_G]

Plotloss said:

Boo Back to smashing my head against Minimo for the time being I guess...

Looks like you grabbed that post before I deleted it and had chance to update it properly! Had only tried against betas before... the above is correct for production releases of WM, IE, Flash 7 PPC, and Flex 1.x.