File Systems.

[Robert W. Current]

Erik Troan wrote:
> > No distribution vendor is ever going to embrace this philosophy.  It's
> > completely needlessly restrictive.
> 
> I strongly suspect Red Hat's users would hate it if we started littering
> /opt with things from our distribution. /usr is correct for packages that
> come with the operating system, /opt is correct for 3rd party packages, and
> /usr/local should be left alone for system administrator's use.
> 
> Let's standardize the well-accepted tenets of Linux. They've evolved for
> a reason, 20 million people are comfortable with them, and it will speed
> the adoption of the LSB.

Is this Erik's stance (speaking for himself or Red Hat?), or that of the
LSB?

If it is the stance of the LSB, many (including myself) will walk
completely away from the LSB if this is the case.  That's not a threat,
that's not a flame, that's not a discussion item.

I am simply saying, if the LSB turns into "what's good for Red Hat is
good for the LSB, what's bad for Red Hat is bad for the LSB," then the
LSB is completely useless to me, I might as well go buy a $1.99 CD for
Red Hat if I want "Linux" (in quotes).  And in fact stuff I am working
for my job and projects will not benefit, because they are focused on
other markets, ones that don't consider Netscape, Apache, and Emacs to
be essential parts of "Linux."

The part where "/opt is correct for 3rd party packages" and
"well-accepted tenets of Linux" added to "20 million people are
comfortable with them" all just sounded like "let the Red Hat de facto
standard reign, it's easier" IMHO.  I doubt Patrick Volkerding would
never accept putting one distributions benefits above standardization
(even his own), the embedded market won't care for this approach, and it
sounds like the LSB would will only address the desktops market.

The whole issue of /opt aside, Netscape on a Red Hat CD doesn't, and
probably never will make me believe Netscape is a base component of
Linux.

I had hoped that the LSB would benefit more than that, and be useful to
Linux Router Project, muLinux Project, Calcaria Linux7k, CAJUN,
DragonLinux, Trinux, LinuxCE, LinuxSH3, Linux Embedded Project, PDAs,
device controllers, and hundreds of other projects that are taking Linux
(apparently only "the Linux kernel" and not the above LSB definition of
Linux) into brave new uses and markets.

If that is the case, that's fine.  I don't mind, and like I said, I
don't think it's worth fighting over.  It's just not productive to
continue to debate this for another 4 days.  I would hope that the LSB
would make it more clear that it's really LSD (Desktop) not LSB in this
case as to not waste others time.  There are a lot of cool other
projects out there, and I'd hate to see them waste their time on this as
well.

Good luck, I wish you all well.  Good luck on your quest to topple
Microsoft.  I'll be happy for you if you win.  It's a big big world out
there, there are things I'd rather do than worry about world domination
of the desktop.

-- 
+---- CurrenTEK --- http://www.currentek.com ----+
|     "Robert W Current" <rob at currentek.com>     |
| Founder, President of Research and Development |
+------------------------------------------------+