Date: Thu, 17 Feb 2000 15:28:08 -0800 From: Alfred Perlstein <bright@wintelcom.net> To: Paul Robinson <wigstah@akitanet.co.uk> Cc: Jonathon McKitrick <j_mckitrick@usa.net>, chat@FreeBSD.ORG Subject: Re: BSD vs Linux comments Message-ID: <20000217152808.U3509@fw.wintelcom.net> In-Reply-To: <Pine.BSF.4.10.10002172205410.51357-100000@elwood.akitanet.co.uk>; from wigstah@akitanet.co.uk on Thu, Feb 17, 2000 at 10:14:22PM %2B0000 References: <20000217130241.T3509@fw.wintelcom.net> <Pine.BSF.4.10.10002172205410.51357-100000@elwood.akitanet.co.uk>
next in thread | previous in thread | raw e-mail | index | archive | help
* Paul Robinson <wigstah@akitanet.co.uk> [000217 14:35] wrote:
> On Thu, 17 Feb 2000, Alfred Perlstein wrote:
>
> > I think your friends are mistaken, BSD isn't stagnating in 80s/early
> > 90s technology, BSD got it right in the early 90s. Linux is
> > still playing catch-up and seemingly not doing a very good job at
> > it. It's 2000, which subsystem in Linux is going to need to be
> > gutted and redone from scratch this month?
>
> I think that's the crux of the argument really - it appears that BSD is
> generally thought about and planned, and the people doing the coding
> attempt to think about it before coding. Linux appears to be whatever was
> covered in Dr. Dobb's last month, or what some 18-year old was taught in
> his first-year 101 university lecture the semester before... not
> particularly scientific...
>
> At least linux got the *marketing* right... :)
Right in what sense? If you mean decieving users, then yes, very good,
probably better than Microsoft as you'll hear many people mistakenly
praise or defend NT, but hardly ever go as far as to flame/attack people
that dislike NT unlike recent incidents with Linux.
It even amazes me, I've had some pretty technical people tell me
that Linux will scale to 32 processors or more, yadda yadda, now
this was over a year ago. The current Linux SMP implementation is
_sorta_ ok, meaing it's better than nothing, but a year ago it was
hardly better than nothing.
There's still major codepaths in Linux that look something like this:
lock_bar(); lock_baz(); lock_kernel();
do something;
unlock_kernel(); unlock_baz(); unlock_bar();
At over 3000 clocks per locked cycle ending with a 'lock everything'
I hardly see this as being as scalable as the hype indicates.
I'm really suprised how the Linux propoganda machine works, it
seems that spreading mis-information and hype about oneself really
works.
Unfortunatly although tempting, it's not the right thing to do.
In my previous email I mentioned about the fsync problem in Linux,
it seems that one can mount a Linux partition in 'sync' mode, whether
that means sync metadata or sync _all data_ I'm unsure, but the real
point is that it ships that way in order to give the appearance of
speed, conviently forgetting about reliability and the bottom line,
doing the right thing.
From the FreeBSD mount manpage:
-o Options are specified with a -o flag followed by a comma separat-
ed string of options. In case of conflicting options being spec-
ified, the rightmost option takes effect. The following options
are available:
async All I/O to the file system should be done asynchronously.
This is a dangerous flag to set, and should not be used
unless you are prepared to recreate the file system
should your system crash.
The linux default install is async, the user is _not_ warned.
One of my early posts to the lists in re using FreeBSD was why the
heck doing file ops where so darn slow, people explained the concept
of meta-data to me and the implications of setting the bit to enable
this. With Linux I probably would have never posted, then wondered
why my filesystems turned into mush when someone kicked out the power.
But, hey, then I would of had a good excuse to try out another Linux
distro. :)
I like FreeBSD because if something toasts a user I can usually tell
them that they were given ample warning not to do XYZ, Linux enables
A-Z and expects you to fend for yourself.
Please don't bring reiser to the table unless you've read and
understood: http://devlinux.com/projects/reiserfs/8_1.html
Look, I'm not even really saying that FreeBSD is better, I'm just saying
that some truth would be a refreshing change.
thanks,
-Alfred
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-chat" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20000217152808.U3509>