Blogs

RedMonk

Skip to content

Rich Internet Applications: “This Conversation Is Bullshit”

The liveliest session at RedMonk’s annual unconference at Sun Microsystems CommunityOne event on Monday was the rich Internet application discussion. Given RIA was the last chat of the day, that was pretty good going. A point that Stephen also makes here with some photographic evidence. Its always interesting to see the way our industry makes progress, without any agreement on definitions. From a personal perspective the RIA session felt like a vindication of the unconference format. It wasn’t a panel, with the “experts” telling everyone else what to think, it was a discussion where we were all peers. I took some positions to keep the conversation lively (I prefer Poor Internet Applications). Unfortunately I lost the notes I took while moderating so I can’t provide a formal writeup [fail!] but here are some impressions. The whole session was taped though, so I will provide links later.

Funnily enough, it was a strong injection of cynicism that really transformed the conversation from belly button fluff gazing to something valuable. We’d been toying with the question of RIA definitions when Dain Sundstrom of the Apache Software Foundation stepped in:

“I didn’t say that RIA is a load of b/s. I said this discussion is b/s”

Dain’s point was that we were talking about esoteric technically focused issues, when the truth is that what really defines a rich Internet application is good design. At this point I have to make an embarrassing confession – I left the notes from the session at the venue when we left, so I am writing this up from memory. Although it wasn’t his quote, Dain’s point, and the point of the discussion as a whole as that:

“People conflate RIA with good design”

It was pretty funny to see Adobe evangelists face an audience not comprised of fanboys. Without naming names I know that at least one of the Adobe guys left the session feeling that: “we got our butts kicked”. Don’t beat yourself up guy… Its great to test Adobe’s ideas in an open environment. Barack Obama will hopefully blow John McCain away – but without being fire-tested by Hilary Clinton, forget about it.

Adobe has an intriguing working definition of a RIA – “an approach that mimics real life.” RIA should learn and improve based on our experiences (that’s a core Web 2.0 pattern, as Duane pointed out) . On the flipside go to a bank and try and do something and you’ll find the existing process is sub-optimal. Or as I put it:

“Real Life Sucks”.

I really can’t understate how much I appreciate the guys that came to our show last year and then came along in 2008 again and made such a solid contribution. RedMonk super-platinum-with-diamond-sparkles card holder Savio Rodrigues said our second annual unconference rocked summed up the RIA session thusly: 

The RIA discussion benefited from the “industry experts” from Adobe and Sun (i.e. JavaFX) in attendance. Surprisingly enough, we couldn’t agree on a definition for RIA. Adobe’s James Ward suggested that “RIA is anything that lets you interact with computing in a way that you would with the real world”. Someone asked if the Wii was a RIA using James’ definition and James agreed that it was. Others felt that RIA was just another name on things the industry invented 15-30 years ago. Everyone agreed that not all applications should be RIAs. And that a pretty app that is useless is still a useless app.

I think that was the really important point – our industry has a goldfish memory, especially when it comes to learning and driving best practices. Pure eye candy can be a bad thing. Usability is far more important than glitz. Too often we forget the basic principles of good design. Some like me never learned them in the first place.

A good HTML design will always beat a poor Flash one.

One important distinction we tried to tease out was where RIA had value in the enterprise. Jeremiah Stone from SAP argued that specialised roles, and training, could and should help with better UI design, but admitted that SAP screens have a poor reputation for ease of use. I used by favourite example of rich Internet apps in the enterprise- T-Mobile US uses Flex to front end its SAP HR apps, and uses vanilla or “chrome” rather than a custom design, to help with maintainability, and allow for more rapid development.

Lauren Cooney recently left IBM to join Microsoft and brought one of her new colleagues along. Sadly we couldn’t here about Silverlight from Michael because he had to leave early. But it was cool, if not a little trangressive, to have Microsoft people at CommunityOne. Next year I will try and get more along.

I am going to post this blog now, even though it feels a bit unpolished. I want to spend some with my family… after being away most of the last two weeks.

 

Thanks for the image use Ted Leung!

Categories: Uncategorized.

Comment Feed

10 Responses

  1. What I was trying to communicate was how hard good design is, and how dependent good design is on good research data and contextual understanding of the task at hand.

    I am proud to say that the latest SAP designs are enormous strides forward from even relatively recent products, and that is a result of some very hard work from our design and product teams. The daunting thing is that I think the more we invest, the more we upskill, the harder we work, the more appreciation we gain for how poor man/machine interaction is and how difficult it is to produce a well designed application.

    While Dain was spot-on that design was what is key, he was dead wrong when he went on to say that design is easy. To be fair, I had the same attitude 3 years ago when I left development. Having walked a while on the other side, it’s simply grim how poor our methods & tools are as an industry when it comes to design.

  2. > I really can’t understate how much I appreciate the guys that came to our show last year and then came along in 2008 again and made such a solid contribution.

    James, the RedMonk unconferece content is solid because of the people who come to collaborate/argue with the 3 of you. I’d be surprised if newbies in 2008 didn’t show up in 2009.

    Hmm…provide value and people come back. Novel idea ;-)

  3. When will you host one on the East coast? :-)

  4. Hi James,

    I was one of the people involved in coining the phrase “RIA” at Macromedia in the early 2000s (along with a core group of Jeremy Allaire, Kevin Lynch, and Adam Berry if I recall. I am not not sure who first hit on the final coinage, it was the product of a series of discussions.).

    Here is the thing: it had a very clear meaning *at the time* and was a clear contrast to the prevailing mass of applications on the web. Now that the entire web has evolved dramatically, the contrast is largely gone and the phrase is not less meaningful, but certainly less useful.

    At the time, we were in a world of page based web apps. Applications that were using the page request model of the browser to deliver very limited interactivity and client side functionality, and led to frustrating repeated refreshes of the page to do anything. The iconic example we and many used at the time was the Broadmore hotel reservation site. As a Web 1.0 app, it was a long series of HTML pages just to complete a hotel reservation, and it suffered from all the problems of the day (eg, if one made an error and tried to go back, you lost all the info you had entered in the previous pages and had to start over.)

    What we saw them do was create a single screen application with rich interactivity on the client, but still all of the benefits of being a web based application (nothing to install, back end connectivity for inventory and other data using XML, use of client side media/animation to guide the user, reachable through any browser, etc.) We really looked at this as the best of web applications and the best of desktop applications: rich connectivity, platform independence, no install, lightweight as well as rich client side logic and interactivity, ability to integrate rich media and communications. But we dropped the baggage of the page based metaphor that basically required a page refresh for everything and got beyond the layout/graphics/media constraints of HTML.

    So I think the definition made quite a lot of sense, and it was I think a very valuable coinage to capture an emerging class of application that was radically better than the mainstream at the time.

    Now, at that time, much of what I described as RIA could *theoretically* be done with DHTML (now called AJAX) but the reality is that it was not yet sufficiently browser independent and there was little uptake of it after the first burst (and abuse) of DHTML in the late 90s.

    Fast forward to today. The term is less useful because it describes the mainstream. Today a large percentage (a majority?) of web applications are “single screen” and use AJAX techniques to update the screen without refreshing the page gratuitously and the major browsers and JavaScript libraries are sufficiently mature that it is quite possible to create platform/browser independent apps with AJAX. Similarly, the use of rich media, usually Flash, is widespread. Of course, over this 8 years, the Flash Platform approach has matured with richer frameworks (Flex..), tooling, components, messaging, and even richer media (H.264 video, for example). But the paradigm is still the same as we saw when we coined the phrase RIA, it just isn’t quite as “unique” and a contrast to the mainstream that it was. Debating the meaning of the phrase “RIA” has become kinda like debating the meaning of the phrase “application” because most are RIAs. So a more interesting debate (to me) would be: OK, so (a) how can we advance the state of the art to build/debug/maintain such applications rapidly and (b) what is the next major paradigm shift in a world where small screen internet connected non-PC devices out number PCs.

    Cheers,

    David
    Adobe

  5. What, you mean like Norfolk? :-)

  6. Wow. It’s like we were playing that telephone game – yet everyone was on a conference call and somehow the message still got distorted. First off, my definition of RIA is not Adobe’s definition – I don’t think Adobe has a definition. What I was really trying to say about defining RIA can be found in an old and more coherent blog:
    http://www.jamesward.org/wordpress/2007/10/17/what-is-a-rich-internet-application/

    It was a butt-kicking good time though. :)

    -James

  7. I have to say that personally having been dragged along there by Duane I enjoyed it. Working out in the real world, rather than product dev, it was good to see a discussion that included both the (to be generous) “optimistic” view of RIA as well as more robust views of what actually matters. Unlike a lot of other “panel” formats at JavaOne it was actually a decent discussion. I’m looking forward to the video.

    I never realised that the BoFs of 2000-2001 which were more open discussions than the mini-technical sessions of today were in fact “unconferences” :)

  8. Really good design takes am imagination that borders on the insane. And that seems to be what it takes to get peoples attention now. The more outlandish it is the more attention it draws. RIA fits right in there and you are right, good design is tough. Thanks for the post.



Some HTML is OK

or, reply to this post via trackback.

Continuing the Discussion

  1. […] is poor design / usability. So it has always been and so it shall always be. Anyone building RIAs (cough) should know this. Products like Flex make the eye candy easy. Only humanoids can produce […]

  2. […] most interest to me was the round table discussion about “what is an RIA?“. There were various opinions on this that I will not repeat here and let you read over on […]