🚀 go-pugleaf

My opinion about how the GNU software's installation path should be set.

Currently, the GNU's products are distributed with the makefiles or

source codes that install or place the files (binaries, library's 

include files, shell scripts, manuals, documents, help files)

I currently, must rewrite every codes of GNU's when I install them.
Firstly because the GNU tools are not organized in this manner.
Secondly, I want seperate GNU's directory for the ease of maintainance.
(The GNU products updates very rapidly, so making tape backup of only
GNU's sources and binary is much easier than they are mixed into
the others in /usr/local. )
Thirdly, it becomes much clear which is Free software
when the path name is /usr/gnu. If one wants to use gnu tools he just add
/usr/gnu/bin in his command search path and he can place the directory
ahead of even /bin or /usr/bin. /usr/local should be come after
those.

My opinion:

1) The base path should be "/usr/gnu" like "/usr/ucb"
   rather than "/usr/local" since GNU's software is not
   just PDS but is Free Software that has different copyright(copyleft)
   policy than the others.

   Thus,
	GNU source codes is placed under   /usr/gnu/src
	binary will be placed under 	   /usr/gnu/bin,
	shell script may be placed under   /usr/gnu/bin, (But I think script 
files should be placed differently from the binary)
	libraries will be placed under 	   /usr/gnu/lib,
	manuals pages will be placed under /usr/gnu/man/man*
	include files be under             /usr/gnu/include
	document files be                  /usr/gnu/doc
	help files be under                /usr/gnu/help
					   /usr/gnu/etc

2) Unifying the base path above (not under /usr/local )
   greatly reduces the user's complication about the maintainance
   of the GNU software.

   For example, we can compile the source code on one site, and then
   move it by transferring only under one directory /usr/gnu with sources.

   For the networked machine we can easily set the permission for GNU's
   binary unreadable to avoid only the binary files are taken.

   Writting the shell scripts using full path name becomes portable
   since we can assume the intended binary executable is placed on
   the same directory path.

3) The other idea is that all the programs are made so that
it refers the user's environment variable as the base directory
of the GNU products. For example if the environment variable
GNUPATH is set to /usr/gnu, then everything is relative to that
path. If the GNUPATH is set to /usr/local, then ... so on.
However I think floating path is not as good as it sounds.

How do you think?