>>We actually had a request (and obliged) from the stores department that they wanted a flashing "led" in their receiving application whenever there is a pick-slip ready, as the computer is mostly unattended and they want to know if there is something that requires their attention when they walk by, and there is another application that has another blinking led (as requested by the users) but I can't remember what it was.
>
>Look at that. Another example where breaking the so called GUI standards makes sense (and to a paying customer who requested it). This is proof that M$ only suggests guidelines and as such, we shouldn't just blindly accept them.
I think you miss the point. Guidelines/standards cover the vast majority of cases. The idea is to follow them except when you have a valid reason to do so, as in Hugo's case.
I work a lot with applications written by other people and the randomness of UI in some of these applications is dreadful. They'd be a lot better if standards had been followed.
Also, as I mentioned elsewhere in the thread, following guidelines gives you a much better shot at creating applications that meet the needs of folks with disabilities.
I've written on this topic a fair amount. Here's a paper that covers the topic broadly:
http://www.tomorrowssolutionsllc.com/Conference%20Sessions/Best%20Practices%20For%20User%20Interfaces.pdfTamar
