The architecture of information:

The WIMP (windows—icons—menus—pointer) user interface paradigm we’re all familiar with was originally modeled on a desktop metaphor. That is, the user manipulates documents in a virtual two-dimensional space that leverages his or her understanding of how documents behave in the real world. A word processing file is represented by an icon on a “desktop,” and opening the file reveals a window that allows the user to focus on that particular document. Many such windows can be open at any given time, allowing the user to do things like copy and paste content from one to another.

This ability to work on multiple documents of different types side-by-side is one of the reasons why this paradigm has been so successful. I’m writing this post in a text editor window, and have a web browser open in a window next to it. I keep referring to the browser as I write. These two windows — the text editor and the web browser — are the context in which I’m writing this post. When I’m done posting it, this context will go away, and I will create a new one focused on whatever task I undertake next on my computer.

When you work in a WIMP-based system, you create such task-focused mini-contexts all the time. Now Microsoft is testing a new feature in Windows 10 it calls “Tab Sets,” which formalizes these mini-contexts by grouping task-related windows in tabs. Interestingly, these sets are not constrained to your computer; if you’re logged in to Microsoft’s cloud services, tab sets travel with you to other computers so you can resume work where you left off. This video has a good overview:

Tab sets remind me of other context-focused extensions of the desktop metaphor, such as spaces. The main difference seems to be that tab sets are not constrained to a single computer.

It’s a very interesting idea, and I like to see Microsoft experimenting with WIMP fundamentals. That said, I have questions about the implementation. For one, this design seems to assume the user has the same applications installed on all the machines they’re using. While this is a safe bet when it comes to built-in apps such as the system’s web browser, what happens with apps that need to be installed, such as Office? (As shown in the video.) For another, will this work with side-by-side windows, as in my blog post example? Being able to keep multiple windows open simultaneously is one of the main advantages of a WIMP GUI; not being able to do so would make this a non-starter for me.

Via Ars Technica.