RIA Stacks Comin' 'Atcha: Silverlight + Astoria, "Apollo" + LiveCycle Data Services, & Lighting the Dark Data

Pardon me while I do a little bit of blue-skyin’…

It’s obvious but worth pointing out: Microsoft and Adobe are shoulder to wheel at delivering their own RIA stacks at the moment. You can see that Silverlight and Astoria provide the front-end and back-end, respectively for an RIA approach to web-aware application development. Adobe has Flex/“Apollo” and LiveCycle Data Services (currently Flex Data Services, or FDS). Indeed, the LDS page is none-to-shy about pointing this out:

LiveCycle Data Services provides a comprehensive set of data-enabling features for RIAs.

Now, either of those pairings could be used independently, but I’m sure they’re intended to look more handsome together.

There are, of course, other vendors and open source projects…even services. For developers, as I was chatting with a RedMonk client about this morning, all of this means the usual dichotomic “fun” of new toys mixed with the tedium of more choices. At the end of a day, if you’re doing more than having fun programming, you’ll have to choose one. Decisions are painful: meetings ensue.

And then of course, there’s Eclipse RCP, traditional GUIs, Ajax, non-Ajax web applications, and mobile interfaces. 2007 is looking like the year of the UI: everyone’s doing musical chairs now, and sure, that’s fun, but make sure you get a chair when the music stops. That cake look tasty.

Anyone remember the IE/Netcape days of yore? Doesn’t it feel like we need to dust off that script and do some search and replace for “Netscape”->”Adobe”?

RIA Musical Chairs and The SOA We Must Not Call “SOA”

While Adobe and Microsoft now both have a stake in the ground, other big ‘uns in the industry like IBM and Sun aren’t quite so clear in their roles in the area of RIA front-ends. Sun of course is always trying to be true to the dance partner(s) it brought, AWT/Swing. NetBeans has shown the benefits of sticking with your partner over the years, but there sure were some rocky times once Eclipse moved to block.

Parts of the Java community are looking towards Flex as something interesting, but my sense is that Ajax has gotten an early lead on filling in The Next UI bucket for Javaland. “Web application” and Java are closer than kissing cousins now-a-days.

Clearly, the LAMP and it’s drop-dead-hot little brother rails are purely web-heads, though using Flex and (even?) Silverlight as components in web applications is, perhaps, a viable option (more below). It’s worked for sites like flickr, with their orginizr, to great effect.

Lighting Dark Data with Middleware

I’ve mentioned the idea of Ajax middleware before (in reference to folks like The Frontside, Nexaweb, JackBe, and QEDWiki), but it’s worth reprising. The middle-ware people, like Interface21/Spring, have the potential to thrive during this year of RIA musical chairs. More important is that this thriving could be due to improving the lives of customers by providing better access to currently neglected data and processes.

The scenario is this: there’s piles of “dark data” in the enterprise, stored in dimly lit legacy systems and silos. No one knows exactly how to do fun and new stuff with these legacy systems and no one wants to get rid of them. In comes RIAs, and suddenly, as our Colgate friends have shown us with SAP and rails, if you can get just the right middle-ware that thinly layers on-top of all those legacy black-boxes, you can light all that dark data. As one of the Colgate rock-stars put it:

Yes, I know we all have already learned ABAP, but I couldn’t have written all that AJAX and drag and drop stuff in ABAP in the 2 days it took me to build the [Ruby on Rails] application.

Being the RIA middle-ware provider here is a nice, low-risk, good payoff move. As a vendor, the key is not lock yourself to any particular RIA, but instead translate all that “legacy talk” into the hip new lingo that any RIA or web UI will understand: web services and REST.

You can see that folks like NetManage (see podcast), with a little marketing and technology re-tooling could play well here too. It goes without saying — though I’ll type it — that Mule and other ESBs fit tightly in the idea-slot too.

Essentially, what we have here is the opportunity to see a real, pragmatic SOA in action. But, sshhh! don’t spoil it by telling anyone it’s an SOA! They’ll just run off screaming!

Eclipse

In another part of the dance hall, we have Eclipse. Now, Eclipse is about GUI development (desktop UIs). If you can gleam the intent and desires behind many of the projects in the upcoming Europa release, you get the idea that Eclipse wants to be the platform for doing heavy-lifting UIs, but also the platform for building platforms. That is, UI’s that are used to get work done, not just play around in social networks, listen to music, or otherwise do “consumer grade” use-cases.

Sometimes your own car is good enough to move your stuff around. Other times, you want to call in The Big Truck.

So What?

The big question for developers that might use these upcoming RIA stacks going forward is: given that you can already do web application development — you’re probably, as a team, just now “getting” Ajax or even .Net — is it worth re-learning a whole new UI framework?

Microsoft and Adobe would point towards enriching the user experience as a mega-plus; though there be much brand-as-expierience black-magic there, I sort of like it. Still, each development team needs to take out their “good enough” sweet-spot charts and rulers and figure out if even starting to think about RIA stacks is worth the rewrite, to be frank. Despite all this glowing RIA talk on my part of late, I would still choose a web application the majority of the times. But, if you gave me time to think about it, I’d at least consider RIAs now.

As I’ve mentioned before, I’ve spoken to a surprising amount of architects and developers of late who seem to be leaning heavily towards “yes” to the above question of using RIAs. The yes’s get even more strong when you talk about behind-the-firewall apps (which Nitobi could have told you long ago, as they did in a podcast episode last year). In all honesty, large, existing software projects find themselves re-writing large sections of the UI each release. Really, improving or even just re-skinning the UI can be a big ticket feature. Ajax has shown us that, if nothing else, hands down.

My main guidance for actual use of all of the above is to do it incrementally. The truth is, it’s still very much a web application world. (Heavy GUI lifting as noted above, aside, which still has it’s valid and useful place.) Seriously take a look at and understand what it means to use RIA as a part of a web application instead of the basis, or core, for a web application. If you don’t know what flickr’s orginizr looks like and how it fits into the flickr photo management workflows, check it out. Also, look into and understand the reasons that things like Twitterrific and dashboard widgets (OS X or Vista) are well liked: it’s not because they do a lot, it’s because they do just barely enough. They remove the boredom of otherwise tedious UIs in key points in the workflow.

Removing boredom, flippant as it sounds, is something well done IT can actually do that helps “the business” and improves the lives of users. Whether it be in the medium of RIA, Ajax, or GUI, getting even a bit of that is worth being hopeful about.

Disclaimer: parts of Microsoft are clients, as is Adobe, Mule, Eclipse, and IBM.

Technorati Tags: adobe, ajax, apollo, astoria, eclipse, flex, livecycle, microsoft, middleware, mix07, mule, netmanage, redmonkclients, silverlight