Data Separation Guidelines

Application settings belong to users, not machines. This is mostly about putting people in control of their data and preferences, therefore separating what is machine data and what is user data. When done correctly:

• People can easily back up "my stuff."
• Many people can share a computer, both at home (mom, dad, and the kids) and in the workplace.
• A person can use many computers, maintaining their data and preferences on each computer.
• Administrators can easily manage corporate desktops

Data separation refers to user and machine-specific data and settings, as well as saving each to the appropriate location. The steps in implementing data separation are to classify the data into logical buckets, then map the data from this logical categorization into physical storage. In between, you obviously will use and manipulate the data and there are considerations for that as well.

The three general guidelines that are important to remember are

• When classifying and storing data, make no assumptions about a user's location or environment. The user may have roamed to an entirely different machine, or is sharing a single machine with five other users, and your application will have no way of knowing for sure.
• When trying to access data, make no assumptions about your access to the system. You may not have sufficient privilege to do anything but read.
• When opening files, always query the path beforehand to assure it hasn't changed since the last time you used that file.

Specific Windows platform implementation guidelines can be found directly below each of these sections.