IUI / Web Clients vs. "Rich Clients"

If you haven't read about Microsoft's Inductive User Interface (IUI) initiative, check this out.

This should be ruffling some feathers - I've heard people talking about how they can't wait for .NET WinForm “rich clients” to end the reign of the ubiquitous web “thin client”.  The IUI approach is basically saying that web user interfaces are easier for users to learn than most desktop apps.

The model is Microsoft Money, which feels like a web app and has to be one of the friendliest apps I've ever used. Desktop development allows developers quite extensive options for UI design - enough rope to hang ourselves. Web apps, on the other hand, are constrained by both a smaller UI toolset and an implicit requirement to keep each screen simple (slow and unreliable connections dictate that user interaction is concise and incremental). Users of a web app perform one simple task per screen rather navigate their way through all their tasks on a complex screen. We've always assumed that our goal was to design a powerful UI for power-users, then make it as painless as possible of users to work their way up to power-user status. The MS Money team discovered that most users never make it to power-user status, and keep struggling with overly complex and powerful UI's they never fully utilize.

In the “real“ world, if you want to fly to Denver you talk to a travel agent who asks some simple questions, one at a time, then makes sure that you're flown where you want to go. The traditional rich client approach would drop you at the controls of a 747 with a grinning paperclip to talk you through the flight. Which makes more sense?

On a slightly related note, I heard a great quote this week: “If you ask an engineer the time, he'll tell you how to build a clock.”

1 Comment

  • You're right, Ron. That's what the MS article says, too. I extrapolated a bit, but here's my conclusion - a well designed desktop UI _should_ be better than a well designed web UI, but often developers outsmart themselves and come up with a slick paradigm that the end users don't grasp.



    Web UI design is definitely UI design with training wheels, but the fact is that users are happy with web apps - possibly not despite their limited toolset, but because of it.



    So the IUI approach says to take what works from the web app paradigm and apply it to desktop design. The cynical Jon approach (at least when writing the original blog, and more for the sake of argument) was to extrapolate that and say "we'll probably screw up a desktop app, so let's stick with web apps until we understand our users a bit better". A more realistic statement would be that end users aren't clamoring for rich UI's because they're perfectly comfortable with simple web UI's.

Comments have been disabled for this content.