cfengine-develop
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Cfengine-develop] Development plan / meeting


From: Mark Burgess
Subject: Re: [Cfengine-develop] Development plan / meeting
Date: Mon, 3 Mar 2003 09:31:43 +0100 (MET)

> Everyone, what role do you see cfengine as fulfilling
> in your network?  Framework, tool, or somewhere in between?  Where is
> it now?

Everything we do exccept package management (a common request on the
lists -- and something that Red Hat is looking at) is done in cfengine.

I think /bin/ls is a tool. Emacs is a tool.

I am not sure what is to be gained by saying that cfengine is not a
framework. Here's my bottom line.

  BIGGEST GOAL OF ALL:

     Cfengine should attack the problem of autonomous system
     administration, based on solid and innovative research.
     It should never just "go with the flow". If we don't know
     what is best, then we need to research it and find out,
     test it, simulate it, write lots of papers, whatever. IMO
     the worst programs out there (esp in sysadm) are those where
     the author begins "In my opinion..." and then they share with
     us their narrow vision.

     In cfengine now there are two types of feature:

     * well tested, well understood mechanisms with predictable
       properties.

     * Experimental parts (like cfenvd) that are "research in progress".

This is the way I want to continue. I loathe opinions. I want
predictable properties that can be combined flexibly, without making
things so difficult for the user that they lose interest in doing things
in the way that extensive research says is "best".

Mark

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Work: +47 22453272            Email:  address@hidden
Fax : +47 22453205            WWW  :  http://www.iu.hio.no/~mark
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~





reply via email to

[Prev in Thread] Current Thread [Next in Thread]