Doing a little forensics, the text came in with the Servlet 2.2 document from 1999, where it was in section 3.2. It has never been changed since.
In Servlet 2.1 from 1998 there was a completely different document, but the following text resembles it a little:
"A servlet context object is at least as unique as the host in which it resides. In a
servlet engine that handles multiple virtual hosts (for example, by using the HTTP
1.1 host header), each virtual host must be treated as a separate context. Servlet
engines can also provide context objects that are unique to a group of servlets. You
can group the servlets administratively or define them with a deployment descriptor."
So maybe the current 4.2 refers to the days before the concept of a web application (war archive) was invented and refers to that ancient concept of "grouping servlets administratively"?
I've never even read it before. This is the first time I have read the
spec from cover to cover. I've always dipped into the bit I was working
on in the past.
I'm currently leaning towards removing that section in the next pass
over the spec (there are a few minor issues to clear up after review of
the proof-reading PR).
Mark
