Adobe Talks Flash

SIMON BISSON TALKS TO ADOBE’S MARK ANDERS ABOUT FLEX, APOLLO AND THE FUTURE OF ALL THINGS FLASH

One of the creators of Microsoft’s ASP.NET, Mark Anders, joined Macromedia to lead the development of Flex Builder. He’s now senior principal scientist at Adobe, where he’s working on turning Flash into a platform for building next-generation Rich Internet Applications. We sit down with Anders to have a chat about Flex and the future of Flash.
It’s a cross-platform future, where Flash doesn’t care which operating system you’re using. After all, Anders himself recently switched from a Windows PC to a Macintosh and realised that he didn’t develop tools for either system – what he developed them for was Flash. He’s enthusiastic about the future of Flash and the up-coming Apollo release: “I think it’s really catching on very nicely. We’ve shown an Apollo demo from eBay and it’s gorgeous – a beautiful app completely branded.” It’s another move down the road to Rich Internet Applications, which are more closely related to desktop applications than your browser. Apollo lets developers design applications on one platform that work just as well on another. Since it’s all Flash you can have different chromes – and with the system chrome you get a Windows or Macintosh user experience.
Apollo lets you also have a custom chrome with a transparent background. This means you can build irregularly shaped applications, which used to be very hard to do in Windows. There are similarities between Apollo and what you can do with Windows’ new WPF user-interface tools, but Anders isn’t too bothered by what he has seen of Microsoft’s new tools: “It was kind of amazing because it was launching Expression and was at Flash on the Beach, so I got to see the latest apps – and by and large I don’t think they look dramatically better than the apps I’ve seen done in early Apollo.”
Taking Flash out of the browser is important. Anders describes it as pushing the boundary, by offering a “crossoperating system runtime environment that allows people to build applications using the skills that they have with existing technologies such as HTML, AJAX, Flash and PDF and to run those applications outside the browser.” It’s an interesting approach, delivering a desktop host for applications developed with those technologies. As Anders says: “We’re doing it with skills people have today, but with some new twists. The way that Flash and HTML combine today is the HTML page contains the Flash.” Apollo changes that: “The way we’re doing it with Apollo is to turn it around and embed HTML within Flash. We’ve integrated the Webkit HTML engine, which is what powers Safari on the Mac, but it’s composited through the Flash rendering pipeline. So you can take HTML and have it in a piece of Flash and apply filters to it, rotate it, skew it and use it as an integral part of a more Flash-based presentation.” Apollo will take advantage of the work done on Adobe’s Spry AJAX framework, so you’ll be able to embed Web 2.0 components (like Google Maps) in your Apollo applications.
Flex Builder began life as a proprietary tool, but the latest version is based on the open-source Eclipse IDE. Originally designed as a Java development platform, Anders describes Eclipse as “a great environment for building tools”. Working with Eclipse made Flex Builder easier to deliver. It certainly made a difference, as Anders points out: “We shipped Flex Builder almost exactly 18 months after we began it from scratch – that is beginning on a new code base.” The developers weren’t Java developers either – they were C++ programmers from Macromedia. The task was made harder as it was being built around a framework that was still under development. Flex was being ported to ActionScript 3, so every module of code was gone over and Flex went from a loosely typed language to one that was much stronger. While they went from nothing, to a full development platform in a year and a half, it still took longer than planned due to the complexity of working with so many different tools.
Some of the credit for that success has to go to Eclipse, which Anders describes as a fabulous framework. He made the initial proposal to use Eclipse for Flex Builder shortly after joining Macromedia. There were two other alternatives. One was a custom C++ development on top of Macromedia’s existing tools and frameworks. The other was to use Microsoft’s .NET platform. Anders was a big .NET fan, having been involved with it at Microsoft from the early days – but there was one big problem: it wasn’t cross-platform. As Anders says: “I’ve spent the last six years working on building a managed code platform, I thought it was goofy to build in C++. The thing I concluded about .NET is that if you’re trying to build code to run on a variety of platforms .NET could offer you nothing and in fact, where’s the cross-platform Windows form? There wasn’t one. Mono is doing a windowing toolkit, but that just makes a great environment for writing Linux applications. I didn’t want a Linux application, I wanted a cross-platform application.” For Anders, Eclipse was a better approach: “Although I liked .NET more than Java and even though I like C++ as a language, the fact is that if you’re building a tool, .NET has nothing – Eclipse is incredible for building tools.”
In fact, Eclipse nearly didn’t make it as the basis of Flex Builder. Eclipse 2.0 wasn’t quite good enough for Macromedia, so Anders and his team began to build their own platform. A month or so later Eclipse 3.0 arrived on the scene and everything changed. Anders had his platform, and work on Flex Builder could switch to the more mature Eclipse IDE.
Adobe also had to integrate Flash into Eclipse as the whole of Flex’s design view was written in Flash. Anders was pleased with the result: “The advantage is that we’re running all of the live components and so you see exactly what you’re going to get.” It’s a very different way of using Flash, Anders points out: “You’re dragging Flash components round on screen and manipulating them, but at any time you can flip back to source code.”
Flex does away with the Flash timeline, which pleases Anders: “That’s why I quit Microsoft and went to Macromedia – the timeline. I’m a developer and there was the dual realities. I would see amazing stuff done in Flash and think ‘that’s incredible’ and then I would pick up the tool and panic.” Even so, he saw a lot of potential in Flash. It’s easy to see what people are doing in Flash compared to other web development technologies, but Anders felt that Macromedia could do better – building on the work done with Flex 1.0.
There needed to be changes. Anders thought: “We could create a tool that would appeal to developers – and we wanted to make the SDK free.” Moving away from being a server was also key, as it allows Flex to be widely adopted beyond the enterprise market.
The next step is to take Flash offline, which Anders says, “Is really what Apollo is aiming for”. With Apollo you’ll be able to use whatever techniques you want: “You can use Flash, you can use HTML, if you deal with PDFs, use those.” And with Apollo, you don’t have to run in a browser or create a hybrid application with online and offline versions.
The future looks pretty bright, as Anders says: “What we’re doing with Flash is forever pushing the boundaries and enabling new things – in Flash 8 there were new visual effects, new video codecs, filters, alpha blended video and new drawing commands. People took incredible advantage of it. Flash 9 focused on memory access and a new language, with a ten times increase in performance.” That’s not the end, as Anders concludes: “We’ll always be looking to increase the richness.” Adobe’s plans for Flash are ambitious: “We’re going to be working on new stuff too. We did Flex Builder 2, start to finish, in 18 months and we’d like to ship more often than that. We’re working hard on Apollo and we want to get it out by June. Beyond that we have been showing Flash 9 and there’s some really exciting things going on there. I think you’ll see much better integration between all the tools. One of the things we’re focused on at Adobe is that we have all of these great tools and making them work well together is a critical thing.”
So what of the future of Flex? Anders plans to increase the Flex community: “I think one area where Adobe can improve is in how we talk to developers and how simple we make our message. I would like to see www.flex.org be more effective in communicating what Flex is. But overall it’s been good – I hear from people here in London there’s a huge demand for Flex programmers. Luckily there are good problems to have and bad problems to have. To me the good news is pretty much everybody I hear who takes a look at Flex and tries it out says ‘hey, it’s a great product’. To be more successful we need to do a better job of communicating what it is and who should be interested in it. The good news is it’s not the product – it’s not that we have a bad product, we just need to evangelise a little better.”