> So if we've got an entire website under version control, and it's a gig of
> data, the only way to browse the Git repository is to clone the entire gig
> of data down to my [workstation], even if I just went over to a designer's
> workstation and, say, wanted to grab a copy of the website logos that were
> used 2, 4, and 7 years ago?
Git is not a backup system (
http://stackoverflow.com/questions/2576198).
If you want to use it nevertheless, I suggest to set up gitweb on the
server, or to access the server per remote login. Or you create a Git
submodule structure that separates the small parts (e.g. code and
icons) into an own repository.
Robert