Home Messages Index
[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index

Re: [News] A Look at the Linux-based Challenger to ZFS

On Sun, 2 Nov 2008 07:14:12 -0500
"Moshe Goldfarb." <brick.n.straw@xxxxxxxxx> wrote:

> On Sun, 2 Nov 2008 02:19:01 -0500, Michael B. Trausch wrote:
> > On Sat, 1 Nov 2008 21:27:58 -0400
> > "Moshe Goldfarb." <brick.n.straw@xxxxxxxxx> wrote:
> > 
> >> Here is what the Linux audio developers have to say about too many
> >> Linux sound systems:
> > 
> > What, pray tell, do multiple file systems all adhering to the same
> > API have _anything_ to do with the multiple sound systems all with
> > their own APIs?  Being able to use multiple filesystems is a
> > transparent process under Linux.  Plug in a FAT16 formatted device,
> > it works.  Plug
> It's about too much choice not always being a good thing.
> I notice you snipped the part about the Chinese wishing that the many
> Linux distributions could at least be compatible with each other.

Linux distributions _are_ compatible with each other.  There is usually
binary compatability, but in the case that there isn't, there _is_
source level compatability.  That's the important part anyway, since
Linux works on so many architectures.

There is some level of interactions which vary from system to system
based on custom setup and all of that.  But you have this in the
Windows world, too; it's entirely possible for a program to fail on
Windows by making assumptions that aren't true for every machine in the
class of "Windows machines."  The same is true for any other operating
system---so citing it as an issue only on Linux systems is quite
misleading.  Good programmers (no matter what environment they come
from) realize this and closely guard against assumptions that they make
and document them.

No software will be compatible with anything directly outside of its
tested execution environment if the software assumes that the tested
execution environment is a perfect image of what users will have.

> Face it, you can slice and dice and dissect down to the minutia any
> part of this choice topic all you wish but the facts are on the table.
> Linux is doing horrible as a desktop system, and too much choice is  a
> major part of that.

You are talking about your opinion, and stating it as a fact.  That's
not a great way to make an argument.

I personally think that Linux is a *wonderful* desktop system.  That
is, of course, why I use it on both my desktops and my servers.  My
girlfriends (and my family for that matter) happen to agree with me,
because their systems have just worked since they were setup.  And
they'll continue to do so until something breaks at the hardware
level.  You can't get that kind of robustness and reliability out of an
operating system where the common programmer's mindset is "give me
elevation" and the design makes the barrier to entry for malware very

> Pool your resources, stop the distribution of the hour, focus on a
> single desktop that actually works (unlike the hodgepodge of kde and
> gnome each trying to out do the other), stop the internal bickering
> and focus on beating Microsoft at their own game.
> That is the only way Linux will top Microsoft.

Mayhaps in your own opinion.  But free software isn't about popularity,
so your point is kind of lost.  I _still_ think that choice is the most
important thing we have.  Any human being has this as a basic
fundamental right, whether they choose to exercise it or not.  If a
person chooses to use Windows, this is their right, just as it is my
right to choose to use Ubuntu.

If people choose to use Windows because there are too many choices in
the Linux world, I would submit that these are the same people that
don't look for software online and head to the store to ask the guy at
Best Buy what they should buy because they want to do something in
particular.  These are the kinds of people that like cookie cutters.
That's _fine_.  It's a very different way of thinking.  The philosophy
behind UNIX and UNIX-like systems is that there are lots of very small
tools that are very excellent at their jobs.  This philosophy has been
proven useful for 40+ years.  Windows doesn't follow this philosophy,
and that's perfectly fine.  So what?  Windows is only a little older
than Linux, and far younger than UNIX.  It remains to be seen whether
Windows will endure 40 years without having a major shift towards a
slightly more UNIX-like way of thinking at the system level.

The way it looks, Microsoft's R&D would like to see WIndows go away,
too, and replace it with a system that follows the small toolkit
philosophy with a different spin on how to apply it (using managed
code throughout the entire operating system and small libraries and
things similar to PowerShell for interfacing with the system on a
technical level). I think it will be interesting to see, and hey, you
know what---if the system is any good, I'll be happy to see it enter the
scene.  If all operating systems were equally functional and robust, we
wouldn't have a need for the pimply-faced teenager stereotype nerd to
fix grandma's computer when she gets a virus.  (So many choices when it
comes to viruses, which one shall we get today?)

> If you guys want all this choice, something will suffer.
> I would 17+ years of Linux failing to make a dent on the desktop is
> pretty good evidence that your choice method isn't working.

Hrm, the first production version of Linux was released in early 1994.
It's late 2008.  That would seem to amount to about 14½ years.

And Linux is, and has been, a _huge_ success in every way measurable,
particularly since popularity really isn't the goal.  The goal is to
have a great system in place for people that want to use that system.
Don't like the ones that are available?  You have the freedom to build
your own!  It's not a black box.

Before calling Linux a failure, take a look at the LKML.  There's
_lots_ of activity there.  If Linux were dying anytime soon, the LKML
would be a ghost town.  Take a look at the companies and the
individuals that are putting their time and money into the system.  All
signs point to Linux being very much alive and well.  The GNU system,
too.  After all, most Linux systems out there use the GNU userland
stack (libc, coreutils, etc).  It's not _necessary_, but it's a great
thing to have.

	--- Mike

My sigfile ran away and is on hiatus.

[Date Prev][Date Next][Thread Prev][Thread Next]
Author IndexDate IndexThread Index