Human-Computer interaction stuff

I was reading some interesting stuff about human-computer interaction the other day.

First-- someone came up with the idea of creating an "interface synthesizer" called VIS. The idea behind VIS is that programmers should write their applications in a generic way, and then use an external library to generate the user interface. This way, the user could theoretically use whatever kind of user interface he wanted to communicate with the application-- a command-line interface, or a point-and-click window interface, or even a web site.

As always, there are some practical objections to this idea, which the author goes through. Some applications are so interactive that they would have little use for this library. For example, Quake III would probably not find this library very useful. Other programmers might be unhappy about losing precise control over the user interface that was generated. It wouldn't be possible to specify exactly where each button should go, for example.

Still, it's a very interesting idea which has the potential to make certain kinds of development easier. It's a shame the guy who wrote the paper is so lazy. He feels that implementing the necessary libraries for windows would be "a tremendous amount of work" and "the author has no desire to become knowledgeable enough in such toolkits to do that part of the work."

There's a lot of interesting discussions down at the Gnome3.0 discussion area. For those not familiar, Gnome is a Desktop manager for Linux and some other operating systems. These guys spend time dealing with windows, buttons, and user-interface stuff like that.

The most interesting idea posted seems to be the idea of rethinking the whole desktop in an object-oriented way. According to "HP," one poster there, objects such as "email," "conversations," "documents," and "contacts" could be turned into "first-class objects" shared between applications. Under this model, the window manager would essentially be managing objects instead of windows and buttons.

I guess desktop environments have already taken some steps in this direction by including the "open with..." dialog box, which allows users to easily double-click on a document and have it opened in a (hopefully useful) program. But for the desktop to become fully object-oriented would require a lot of work. A lot of components would have to become integrated-- something that is next-to-impossible in the world of proprietary software, and very difficult in the world of open source as well.

Well, it's an interesting idea.