Recently it was announced that NetBSD
is adding PAM (pluggable authentication modules) to core.
This comes not long after
a new logo for NetBSD was announced - a
logo confusingly invocative of RedFlag Linux
Flaws in PAM account for a disproporationate
number of security problems in Linux.
Throwing away their numbering system,
throwing away their excellent logo in
favor of something generic and
not even unique... what the hell is wrong
More recently, a widely circulated article spawning from a NetBSD user group that questioned the continued relavency of NetBSD
This should shed some light on what the
heck is going on with NetBSD.
Quoting from near the beginning of that article, Ten years ago neither Windows or Linux were serious competitors both in the functionality and stability axes. Now both offer more features than we do, and they have behind them the resources of very large commercial organizations.
It looks at Microsoft Windows, which has longer and longer release cycles.
It looks at FreeBSD, whose quest for advanced
features such as fine grained kernel
locking is preventing them from releasing
a production quality system.
Then they look at OpenBSD and cite lack of
development manpower and its mismanagement.
And, according to this article,
Linux keeps re-writing major portions of the kernel and has stability issues
I should stop for a moment and say I'm a long
time NetBSD user, running it since 0.9.
I just barely missed when NetBSD and FreeBSD
were both still 386BSD.
Linux was maybe 1.0 at this time and was
completely unusable for anything except
a toy - Minix was more mature.
So I'm not just taking a chance to berate
NetBSD unneccesarily - this is from the heart.
Reading their list of complaints with
other operations that justify the existance
of NetBSD, you'd think the plan was to
keep the code stable, and keep developers working
on core features such as stability in lieu of
going on a feature hunting quest.
But to quote this article,
We have our work cut out for us and we need work hard in the areas we are behind in order to remain relevant.
They're planning on doing the exact opposite.
They're going to cram in more features,
work on major projects such as the fine grained
locking they say is making FreeBSD not
production quality, which will naturally
expand release cycles.
They're going to work hard to catch up to Windows, Linux, and FreeBSD's instability and code churn.
So in other words, because Windows, Linux, and
FreeBSD all have lots of features but also
serious code churn and resultant stability
issues, it's only natural that NetBSD does
exactly the same thing.
I kind of suspected NetBSD priorities had
been slipping - I have a Sparc I have to admin
that goes offline when it can't recover from
too many device goofiness and has to be hard
booted. I have two Macs that crash for no
reason I can tell (too hard to attach a serial
debugger to a Mac). And NetBSD 2.0 on x86
just had way too many of those little glitches
that really make your stomach churn.
I've started migrating away from NetBSD
until I - or they - can deal with some
of this stuff.
But now it's pretty apparent I won't be
going back to NetBSD.
I won't be able to.
And it isn't for lack of features - this
blows my mind, but Linux is actually getting
to the point where it's more stable than NetBSD
through Linux slowly improving and NetBSD
It appears that this point passed some
time ago and no one really noticed it...
after too many years of the Linux camp thinking
Linux is stable when it isn't
and years of NetBSD being used to being
stable, it kind of makes sence if the roles
were reversed it would go unnoticed.