Every imaginable choice was thrown at users at once and it was up to the poor user to figure out what to do. To cram more information onto the screen, the interfaces of that era used tabs. At some point Microsoft invented the ultimate UI element - a tab with a scroll button in the end which allowed the user to page through hidden tabs.
Another philosophy of the old UI approach was that the user wants to see all information all the time. Instead of building UIs that responded to the way that people actually interacted with the tool, the user interface opened up all possible choices at all times. Naturally, this is completely overwhelming and confusing to people.
The only way to cope with complexity was to introduce a standard set of widgets, such as tables, combo boxes, check boxes, etc., so that users at least had some familiarity with UIs from program to program. But on top of that, there was a myth spread that users were stupid and would not be able to understand a non-standard UI.
The myth was supported by the fact that a lot of people do not respond well to sophisticated visualizations, like graphs, heat maps, or treemaps. While this is true, it doesn't mean that people can not figure out new user interfaces. The proof comes from Apple, which continuously innovates with new UIs for its software products. Also, recent social web applications have made a strong case for simpler, contextual user interfaces.