question re. installing software & init system

[Miles Fidelman <mfidelman@meetinghouse.net>]

Hi Folks,

One thing that is very murky in the documentation is how to install 
software that isn't packaged.  I can't seem to find any clear 
documentation on the process.

Several specific questions:

The recent PPT presentation on GUIX (to Inria) implies that packaging 
can be as simple as writing a config file that tells GUIX to do a 
"config; make; make install" on a source file built with the GNU tools
- do I NEED to package something or can I simply do a basic download, 
untar, config, make, make install on a running system?
- does dmd handle standard sysvinit files (as are still most commonly 
included in source packages)?
- and the there's the whole set of issues, recently raised, related to 
language systems that maintain their own repos and build systems (e.g., 
CPAN) -- somehow, partial import of dependencies into the GUIX 
environment does not seem usable

And a general comment on the documentation:  Given that this is a 
(sort-of) new distro, that does things VERY differently from previous 
distros - it sure would be helpful to have the install documentation 
provide both a very clear overview of the conceptual 
approach/architecture (vs. items spread around various ppts), AND very 
clear step-by-step instructions for:
- basic install & configuration
- installing & configuring packages (both those that run as services and 
those that don't; including later re-configuration)
- installing & configuring software that isn't packaged
- with particular attention to how the installer, package system, and 
init system work together
- and with attention to how these all work with other build systems
You know - something like the Debian install instructions or the FreeBSD 
handbook.

Granted that documentation generally follows code; when doing things 
radically differently, there's a lot to be said for writing 
documentation FIRST - doing so provides a really good check on 
conceptual clarity and usability.  (It kind of makes it hard to do any 
kind of testing, evaluating, or contributing without a good starting point.)

Thanks,

Miles Fidelman

-- 
In theory, there is no difference between theory and practice.
In practice, there is.  .... Yogi Berra