[geeks] Why Linux and not *BSD?

[shannon at widomaker.com]

On Wednesday 03 August 2005 09:13, James Fogg wrote:

> Then the problem is clear. If someone like myself thinks that BSD
> doesn't have a well rounded selection of software then how can we expect
> average joes to know. Without a "company" and a marketing budget I don't
> know how to get the word out, but it happened for Linux and I'm sure it
> could happened for the BSD's.

BSD *DOES NOT* have as well rounded a selection of software in some areas 
though, especially on the desktop, and certain realtime niches.  Just because 
the code is out there for UNIX, doesn't necessarily mean it is well tested 
outside of the origin system.

A lot of new code now is developed on Linux rather than Solaris, DEC UNIX, 
*BSD, etc.  It is therefore usually much more fully debugged and polished on 
Linux platforms.  Things work smoother, have fewer system related bugs, and 
are usually released for Linux first.  

Linux is also pragmatic.  Yes, there are some very vocal zealots who will 
cripple software for political reasons, but most of them aren't like that.  
Their primary drive is traditional UNIX: get work done.

BSD, especially in earlier years, was often hung up in ivory-tower snobbery, 
and a reluctance to make drastic but necessary changes to the OS.  I think 
the fast-forward drive of Linux has helped force BSD to progress faster, and 
despite the chaos of FreeBSD 5.x, recently forward moves by both FreeBSD and 
NetBSD have been good.  Obviously NetBSD's team did a smoother job (they rule 
of course), but I think FreeBSD 5.x is now starting to show the benefits of 
the changes.

Not everything about Linux is bad nor is its popularity all marketing.  It 
does a very good job in a lot of areas, and there are perfectly valid reasons 
for using it.  The 2.6 kernel series in particular made big leaps forward, 
especially on high end hardware, and their varied filesystems and large Linux 
based projects have given people a lot of experience with high end 
applications and the changes needed to support it.

I think in the near future we are going to see more of BSD, and already I've 
noticed more participation in UNIX desktop programming from BSD hackers.
NetBSD 3.x and FreeBSD 5.x are critical releases, and should get more 
attention than past releases.  I'm hoping that we'll see more visibility in 
spite of Linux corporate stupidity.

My personal hope is to one day be able to move my desktop to NetBSD.  Right 
now I just can't get the hardware support I need.  All my servers run NetBSD 
since I started replacing Solaris with NetBSD about 5 years ago.  

But for PCs, right now, Linux does a better job.  Driver support is the big 
issue, and better testing of desktop applications.