If you haven’t already, please take 45 minutes or so and give a listen to Episode 15 of Windows Weekly - “Office UI Team”.
I listen to a lot of the TWiT
Podcasts Netcasts (Leo hates it when you call them Podcasts) from Leo Laporte. I listen to regularly to Windows Weekly which features Paul Thurrott. Paul gives a very fair and balanced view of the Windows platform. He is also a great source of information from inside Microsoft. He is very plugged into the product teams. In this episode Leo and Paul interview 2 members of the Office team that worked on Office 2007 and also are working on Office vNext. This was a fascinating insight into the development of the new UI for Office.
Some of the things that they pointed out should be applied to every software development project. They were:
Do qualitative and quantitative research up front. They gave insight into the studies that they did on how users “find” or “fail to find” features.
Develop tenants or guiding principles and constantly judge your decisions against them. I learned this lessons from a CIO that I used to work for. Nothing kills the emotional argument better than a guiding principal violation.
Plan for multiple Iterations. The Office UI Guys half joked about Microsoft’s reputation on not getting anything right until the 3rd try. So they built in 3 iterations.
Instrument your code. They were able to quantify the work that they had done as successful by looking at the instrumented data (you know that “share my experience data” dialog box). They cited the example that watermarking your documents was seldom used in Office 2003 and in Office 2007 use of it has really taken off (without changing the feature). Without well instrumented code they would never have known that.