Revealed: How a new feature gets into Windows
Microsoft has shared some insights into the process which produces new features in Windows. It’s a reminder of the challenge facing all operating systems: to provide as much user customization as possible without incompatibilities.
The details come in a post to the Windows 7 engineering blog. It begins by pointing out (perhaps obviously) that every new feature starts as an idea. However, the first big problem comes with analyzing how much of a need there is for that feature.
While Microsoft gathers a huge amount of telemetry data (information taken from users’ computers, which report problems, details from usability testing and so on), this info doesn’t always explain a specific problem. The post gives the example of a user’s computer which shows a series of windows being closed one after the other then reopened almost immediately. It’s not until you actually speak to the user that you find he did this because it was the only way to get the icons in the taskbar in a particular order.
In the Windows process, the next step is to use this data to produce a theory (that users would find the system easier if they could manually rearrange the order of icons), then come up with ideas for reaching this aim, and then test the ideas.
However, Microsoft has found that it needs to put as much work into designing the tests as into the actual ideas themselves. That’s because a test only tells you whether an idea actually works: it doesn’t tell you how that idea will interact with other features of Windows. It may be the case that the benefits of solving a particular problem are outweighed by the effects it has on the system as a whole.
The main purpose of the post appears to be to deal with the large number of suggestions (or demands in some case) for new features which readers have added as replies to previous entries in the blog. The message from Microsoft is that even a good idea for a specific issue might not fit into Windows because its effect on the big picture would be a negative. The firm also has to keep an eye on introducing too many options overall simply because it increases the potential for problems
Related Posts:
October 30th, 2008
“Providing as much customization as possible” is not the goal of every OS. It’s not even on Apple’s map. Love them or hate them, their philosophy is that simplicity is key to usability, and towards that goal the less customization the better. That’s why people who love to tinker hate them, and why people who don’t like to tinker love them.
November 7th, 2008
“Microsoft has shared some insights into the process which produces new features in Windows.”
Not quite – Microsoft doesn’t have “insights”. MS is a plagiarist, picking up crumbs from under the Linux and Apple tables.
“However, the first big problem comes with analyzing how much of a need there is for that feature.”
How much need is there for DRM? Everyone knows the answer to that – well, everyone except those who work at Microsoft.
“[...] the next step is to use this data to produce a theory (that users would find the system easier if they could manually rearrange the order of icons), [...]“.
That’s a joke, right? Microsoft move and change things willy-nilly in order to make a tired old code base look “new”, and what would *really* make the system easier would be if they stopped doing that. There are two rules in I.T. – rule one, be consistent; rule two, be good. If you follow the first rule, the second rule usually looks after itself. Sadly, Microsoft cannot comprehend this, and its hapless victims are doomed to the Sysiphean task of forever relearning the Windows interface (and paying for the privilege – now there’s “innovation” for you!).