[Nix-dev] Nix Presentation at Syd-FP [Off-Topic]

[Jeevakan Suresh]

> However, for system administrators and developers it is a different
case

I'd say this is where our 'niche' is - nix eases the task of system
management. Any reasonably large organisation will eventually home-brew
their own 'version' of nix - the fact is this is a problem which is
naturally declarative.  Nix just encapsulates what most sys-admins have
been doing in an ad-hoc fashion anyway (writing kick-start scripts,
versioning config files (and sometimes binaries) etc.).

Although it would be nice if nix (eventually) ruled the world - the more
likely scenario is that it will become the goto tool for sys-admins.
That itself would be quite awesome :)



-----Original Message-----
From: Pjotr Prins [mailto:pjotr2008 at thebird.nl] 
Sent: Monday, November 17, 2008 9:07 PM
To: Jeevakan Suresh
Cc: nix-dev at cs.uu.nl
Subject: Re: [Nix-dev] Nix Presentation at Syd-FP [Off-Topic]

Hi Jeeva,

On Mon, Nov 17, 2008 at 02:23:26PM +1100, Jeevakan Suresh wrote:
> burnt. At the same time however they wouldn't consider switching to 
> nix just yet as they felt it wasn't mature enough to warrant the time 
> investment.

That is the reality for end-users (one reason few are weaned off
Microsoft Windows). However, for system administrators and developers it
is a different case. Those two groups can afford an investment when they
know it pays in the longer term. System administrators are good for Nix,
as they can help maintain and test packages.  Developers are good, as
they can improve their specific packages and the development
environment. If a major packager (read 'developer'), like Ubuntu, would
opt for Nix it would be extremely useful.  End-users, at this stage, are
maybe not so useful, except for testing. I think we don't have the
structure to support and help end-users.

If it were to integrate Nix with Debian packages (or Fedora), we would
easily popularize it. Unfortunately, as we know, that is impossible
other than Debian adopting Nix. 

I am happy with my hybrid system of both package managers, but I think
it will be almost impossible to popularize Nix with end-users. The three
routes available are:

1. Get NixOS to the level Debian is and popularize it

2. Integrate Nix with Debian (or similar) - the hybrid option improved

3. Have Debian use Nix

(1) is hard, a lot of work, and still suffers from end-user
'investment'. (2) is probably impossible to do well and (3) is
political.

So, ruling out (3) until we hear different, and (1) as it is too hard,
it leaves (2).

Maybe if we were to change Debian tools to support Nix packages we could
invite (3) at some point. How about providing a front-end for standard
Debian tools (or Fedora, whatever) where we add a --nix option. I.e.

apt-get --nix update
  updating Nix database
apt-cache pkgnames vim
  vim
	vim (nix)
	vim-python
	vim-full (nix)
	(...)

apt-get --nix install vim-full

That would lower the threshold of use significantly.

Pj.


NOTICE
This e-mail and any attachments are confidential and may contain copyright material of Macquarie Group Limited or third parties. If you are not the intended recipient of this email you should not read, print, re-transmit, store or act in reliance on this e-mail or any attachments, and should destroy all copies of them. Macquarie Group Limited does not guarantee the integrity of any emails or any attached files. The views or opinions expressed are the author's own and may not reflect the views or opinions of Macquarie Group Limited.