[XML-SIG] Inconsistent module names

Nik O niko@cmsplatform.com
Thu, 22 Jul 1999 17:07:38 -0600

Scott Cotton wrote:
>I've switched back and forth between both repository organization
>types for several projects at work, and separating the source from
>the README, etc. has turned out to be *much* better in practice.

Absolutely agreed.  My years of developing software and managing groups of
progs and non-progs (e.g. artists, tech writers) have convinced me that the
most functional project/program/thing tree structure is:
    ./foo    <== .bin, .exe, .html, .ini, _whatever_it_takes_to_run_foo
        /docuser    <== "end-user" doc: readme, .doc, .txt, .html,
        /docdev    <== "developer" doc: readme, .doc, .txt, .html,
        /source    <== .c, .h, .rc, .py, .lib, .mak, .prj,
        /object    <== .obj, .map, _any_build_foo_process_temps

Sometimes a "./foo/images", or "./foo/prefs", or something similar is useful
to further organize the "_whatever_it_takes_to_run_foo" stuff.  With the
above structure, it is very simple to prune branches that aren't pertinent
to a specific release package, e.g., "end-users" would get only "./foo",
"./foo/docuser", and "./foo/images", whilst a developer might get the entire
tree, and a tech writer might get just the "./foo" and "./foo/doc*"

Hope y'all will forgive my obvious C bias -- i don't really use Python that
much, but XML is XML, and software engineering is software engineering.

-Nik O, Content Mgmt Solutions, Jackson, Wyo.