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

Re: File Systems.

On Thu, 16 Mar 2000, Robert W. Current wrote:

> I would like to open a discussion on the merit of moving "packaged"
> userland software out of /usr/bin and to a different FHS compliant area.
> The reasons for this are to allow more structured and organized
> standards of systems administration, similar to those found in some of
> the BSD flavors.  Hopefully, this can be done without making Linux into
> a BSD system.
> Although I understand that the de facto standard for some of the larger
> Linux distributions uses /usr/bin as the default location for software,
> I feel that /usr/bin growing to several G of software is a real problem
> point for system administration.  Furthermore, the convention is not
> completely standardized, shown by the use of /usr/local by distributions
> such as Slackware.

Please define what is wrong with having all this software in /usr/bin? Is
it the fact that there are so many programs? Is it the fact that all being
in the same directory, they are not neatly organized? What precisely?

Are we agreed that putting all this software in /usr/bin does _not_
violate FHS? If it _does_ violate FHS, then 1) precisely how, and 2) how
does this then become an issue for LSB to resolve?

Also, why is it problematic to have several gigs worth of software in
/usr/bin? Also, do you actually know of any systems which have or even
approach this much software in /usr/bin? I'm not talking about the future;
I mean _now_.

I'm playing devil's advocate to your devil's advocate because most of the
reasoning I've seen in the last several days has been very general or
somehow a priori, without prior examination.

BTW, much of your complaint about the filesystems really sounds like a
complaint about the general nature of the FHS. If so, I must agree. You
were onto something at one point when you said something like, "You should
be able to unmount /usr and still boot/admin the system." (Did I get that
right?) This is the type of thing that FHS needs: rather than general
language, specific testable criteria.

Paul M. Foster

Reply to: