RIA will lead to bad user interfaces?

Web Worker Daily makes the excellent point that the new wave of tools and technologies for building web applications could lead to some bad user interfaces:

With the wave of tools like Apollo, Microsoft’s Silverlight, and Sun’s JavaFX coming out, we’re going to have powerful visual design tools in the hands of developers who have little or no idea what works in graphic design.

Tools like widgets / gadgets have already made the whole desktop user experience much more fragmented. Does this matter?

I’d argue that too many application UIs have been built by low-level developers in the past, and have been far too complex. Back in the days when I produced software for UK schools[1], I used to think we[2] were good at this, but the fact is that we were developers and advanced users, and those kinds of individuals rarely understand what makes a good interface design.

Desktop environments like OS X and GNOME have tried to enforce consistency via user interface guidelines. I’ve read that Vista has a bunch of contradictory interfaces (“standard” buttons in different places with different icons and behaviours in different applications), but I haven’t played with it enough to know. Apple’s interfaces, and tools such as Adobe Lightroom, are far more user and task-centric (in my opinion), which is what makes them nice to use.

Will Rich Internet Applications make the computing experience more confusing for users, or will they be an unqualified success? I think there is certainly going to continue to be a case for understanding good interaction patterns and graphics and UI design.

[1] 15 years ago, fact fans…

[2] PTW Software, which I’m sure none of my readers will be aware of… we wrote educational software for RISC OS / Acorn computers

8 thoughts on “RIA will lead to bad user interfaces?”

  1. So back in 1998 I remember presenting on the role of the graphic designer and the developer in building internet solution. The key message is you leave the design to the graphic designer and the coding to the coder 🙂 , If you look at the skills and tools need to build XAML, MXML, or JavaFX script the will need to be graphic designers who know’s how to get the best from it, the commonal garden programmer will be able to work out how to do the basics but not take full advantage of these frameworks. You only need to see what can be achieved with Adobe CS etc, so the recommendation for internet facing high volumne usage solutions will be to put the right team on the project, otherwise it will fail, just like they did 10 years ago. That is why I believe it will be BAU for these frameworks on the public internet, but they are missing the requirements for intranet enterprises, who dont need flashy graphics, they just want data on a screen in the least amount of programming steps as possible.



  2. Short answer yes it will, because more toys will lead to more design chaos..

    But this has happened already with DHTML and flash. The web has more interaces than all desktop software available. A big trouble with developers and designers is that they like to make statements or think different which leads to more design chaos.

    The real answer isn’t to use a designer but rather a user interface specialist who may also be a designer. But good luck on trying to find a common style on the web, many sites even break the humble hyperlink!!

    In my opinion most developers know that they are not GUI experts, the trouble is most designers actually beleive they really are, which is more dangerous?


  3. Matt and Al, thanks for the insights.

    You’re both right of course – this is yet another technology loop. For the past 10 years I’ve observed the same thing on the server side with a lot of the enterprise middleware around today doing what CICS and other technologies did long before; this time the loop is on the UI side, and I guess I’ve been away from that for a while.

    Pretty much as soon as I posted, I realised that Flash itself introduced a lot of the web UI chaos / discrepancy that can be observed today. That’s not to say that good UIs do not exist – Flickr uses Ajax and Flash and is pretty intuitive (at least for me – an advanced user!).

    A complete side note – Skitch is interesting (I’ll blog about it soon) as it is a desktop app that breaks certain conventions and yet remains easy to use. So I’m not arguing for convention in particular.

    Al, I like your point about finding a user interface / interaction specialist who is a designer.

  4. Consistency is an interesting thing. There’s a really fine line, as if you make things too consistent, it’s easier to get lost in an interface as you don’t have any markers for where you are.

    Interesting comments, though I’d kind of like to think that enterprises should be just as interested in having things look pretty as Internet facing systems.

  5. Darren,

    You will be very suprised at how visually limiting the enterprises want the call center apps :-), I agree for some business channels like a checkout kiosk in a supermarket or a branch teller for a bank, the customer could see the interface and it should look clean and easy to use. The enterprise customers also have high turnover of staff in these roles, so again easy of use is key to quick adoption of using the system by new staff. Transaction workers who come in switch a machine on, login, and start doing a task of work, these need to be highly usable and oriented around the tasks at hand, but dont need to have the sexy slide in and out, shadows, and other visually appealing items, they just get in the way of getting the job done, the operators are more interested in keyboard shortcuts and type ahead and productivity improvements. I even had a customer say the other day, we dont want the call centers having instant messaging, otherwise they will talk to each other all day long, so they want a new UI to send message between operators and managers and that is it. I had the chief architect of a large banking organisation ping me the other day saying that he thinks this Apollo, Sliverlight and JavaFX work is missing the point of want enterprise customers need to build and deliver. I agree with him. The new breed of frameworks for enterprises needs to focus on supporting the creation of these enterprise applications as quickly and as easily as possible. I have been quoted again by some large customers it can take up to 50 days with a J2EE programming model to get a single screen from design into production and all it is doing is taking data input, validating, going through a controller into a message onto a message queue to back end serivces ie a mainframe CICS transaction, and then waiting to display the response or the next screen, so how will all this Apollo, Sliverlight and JavaFX fix that for me then !!! 🙂



  6. Meanwhile CICS has MOVED ON. Can the same be said, analalogously, in the UI space?

Leave a Reply