Hello all
the standard practice is to keep the trunk (cvs special tag HEAD) to
be your new code (aka mainline), and the branches for what you're
calling 'old code' (aka releases).
I think this is defined by your workflow. Here at my organization we
create branches to work on new features. The "live" code always is in
the trunk. This way other programmers can get the current code by just
do 'cvs co'. If you create a new branch for the current live code you
always have to communicate the new branchname to all programmers. This
(in my opinion) leads to problems.
After finishing the new code in the branch we then merge the code back
to the trunk. This also has the benefit, that you can work on two
different features at the same time without having to share code in the
trunk (the programmers are independent (more or less) and don't
interfere with each other if their code intermittently is broken). Two
seperate branches which than can be merged back to trunk.
But I really think this is dependent on your workflow.