Because they are completely unrelated code bases sharing nothing. With subversion you could check out subtrees but that isn't practical with git. For my perso

Because they are completely unrelated code bases sharing nothing. With subversion you could check out subtrees but that isn't practical with git. For my personal projects a common pattern is a repo for the source and a repo for generated documentation. For work projects we have an Android client (in Java), iOS client (in objective C), 3 different server components that share no code (Python), various analysis tools (Python), a client customisation layer (mini-Python), our website etc. If all those were in one large repository it would be a huge sprawling mess. Also remember that branches are repository global. There would also be constant updates because other parts have changed, nothing to do with the component you are working on.