Principles of a developing from a dist or install?

[jling]

Sorry for asking this seemingly confusing question.

In my mind, if I distribute a package for which I expect other users to 
possibly work on the source code to enhance it, certainly the best 
thing would be for them to develop on the code as it exists just after 
it is unpacked from the tarball distribution.

Is there any sense in me having the user install the package (i.e. do 
a 'make install') and then have them develop off of the code in the 
install directory? ... assuming I have the source code and headers 
copied over during the install process.

The reason I bring this up is that I have a deep directory structure 
that I develop in that I wish to preserve and use in the build process 
of the binaries, yet to make it easier for other developers I want them 
to develop the code further on a very shallow directory structure.

In order to do this I could either do work ahead of time and copy the 
source and headers into a shallow directory structure before I go 
through the entire packaging process, or perhaps I can do the copying 
during say a 'make install' as I mentioned above and have them develop 
off of that install directory?

However, because I want to build the binaries off of my deep directory 
structure, it seems to me that if I want the best of both worlds then I 
would need two sets of Makefile.am's ... one for building binaries 
using the deep structure and another set of Makefile.am's to enable the 
building of the binaries from shallow structures (so that other 
developers can code and test from there).

Perhaps the reason for doing this does not make sense.  Certainly if I 
don't want others to develop on my deep directory structure I shouldn't 
even be distributing that structure in the first place.  Nevertheless 
what is the best way or ideal way to proceed?

Thanks,
John