1. That RCS revision numbers are used for topic revs
2. That the major revision is always 1

This has two impacts. First, it makes it difficult to map any other revision numbering scheme, such as subversion's single revision number. Second, it means the major revision number in RCS can never be increment.

So why bother with it? If it's always 1, why not drop it completely? If it can be something other than 1, then why assume it is 1 in the code?

I have changed it in the DEVELOP branch so that the only place the major.minor revision number is known about is in the RCS module. Everywhere else in the core the revision number is known to be a single number, with the first rev being revision 1. RCS rev numbers (like r1.3) can still be used on input, but r1.3 is interpreted as "3". In SVN DEVELOP Revision 3335.

-- CrawfordCurrie - 03 Dec 2004