The Unconfigured Process Template, or its customized version with its initializer, etc. configurations is advocated by the link that Eric referred to. My question is with the following Note
For Change and Configuration Management and Quality Management project areas, you should use the Unconfigured Process template to create a project area that will consume process from another project area because that template does not configure any process on its own. Using a template that defines its own process causes the project area that consumes process to override the process it is consuming, which is not usually the behavior that you want.
The Unconfigured Template does not provide many pieces of a Process Template that make some Templates work. For example, the SCRUM Template OOTB.  Reports, Timelines, Plans (Backlog, etc.) are all an integral part for any SCRUM-based Consumer PA Using a Provider Process itself created from the SCRUM Process Template. 
What process is being used when I create Consumer PAs with the SCRUM Template, then point them to Use the Provider's Shared Process?
I followed Sumant's approach setting Final check box for most Work Item Configuration data.
Today, I first created a Consumer using the SCRUM Process Template then pointed the Consumer PA at a Provider created earlier from the SCRUM Template itself, to which I had added some attributes.  I got the Reports, etc from SCRUM and now my Consumer shows the new Attributes I had added to the Provider.  What problems should I expect by having done that?  I thought that the Provider Process Template was copied to the Consumer and that is the Process used.