source control question

[Robert William Vesterman]

Does anyone know of a source control system that is not so 
directory-centric? Most of the ones I've seen seem to have a base 
assumption that, more or less, "directory" == "project". 

But in reality, a directory could be a project, or part of a project, or 
part of many projects, or merely structural (i.e. merely to organize 
subdirectories, any of which may or may not be used in any number of 
projects, each project of which is not necessarily completely contained 
in the structural parent directory).  And a project may span many 
directories, each of which is not necessarily anywhere near the others 
in the overall repository tree structure, and whose repository tree 
"neighbors" are not necessarily parts of the same project.

For example, you may have top level repository things like "work" and 
"personal", which are completely structural.  And maybe "utils", which 
you might use in both work and personal projects.  And then if you use 
some Java, and do the standard way of making packages 
(com.mydomain.blah.blah.blah), you'll probably have a "java" directory 
outside of "work" and "personal", having a whole tree of subdirectories, 
any of which may be a complete project, part of a project, part of many 
projects, et cetera.  And a project may be spread across "personal" and 
"java" and "utils" and any number of other organizational things.

I'm sure there are ways to bend things like Subversion into kind of 
behaving the way I want, but are there any systems that are actually 
designed with this concept in mind?

Thanks,

BOb Vesterman.