On 26 Mar 1994 20:46:22 GMT, Herb Peyerl (hpeyerl@sidney) wrote:

Probably.  If you want the novice to have a working system out of the
box.  Which is not necessarily a bad thing.  :)

lilo

From: l...@slip-2-70.ots.utexas.edu (lilo)
Date: 29 Mar 1994 23:04:07 GMT

On 26 Mar 1994 20:46:22 GMT, Herb Peyerl (hpeyerl@sidney) wrote:

Probably.  If you want the novice to have a working system out of the
box.  Which is not necessarily a bad thing.  :)

lilo

In <2n2smo$...@eve.adam.com.au> st...@adam.com.au (Stephen White) writes:

This is possibly a bit misleading: while there is a core group of
people who control changes to the official source tree (I think that
while I speak of my experience with FreeBSD, the same would apply to
NetBSD) the core group is very open to outside contributions, and in
fact actively encourages this.  Note for instance that Paul Kranenburg
is neither a member of the FreeBSD nor the NetBSD core team, yet both
products use his shared libraries.  

Neither of the *BSD projects is a closed effort.  Sure, if you want
your favourite patch of the day put into the distribution, you have to
champion it with someone who has commit rights.
Remember also that most of the problems occuring with *BSD never see
the light of day with normal users - the users of *BSD-current subject
themselves to the joys of finding all of these bugs ;-)

Geoff.
--
 Geoff Rehmet, Computer Science Department,   | ____   _ o         /\
 Rhodes University,  South Africa             |___  _-\_<,        /\/\/\
   email : c...@cs.ru.ac.za                   |    (*)/'(*)    /\/\/\/\/\
         : ge...@neptune.ru.ac.za             |

|
| As I see it the major difference (from a user impact perspective) is that
| FreeBSD has better network code and Linux has shared libraries. Give FreeBSD
| shared libraries and I would probably pick it over Linux...
|

Well I suggest you stop using Linux now as FreeBSD *does* have shared
libraries. Which implementation is snappier will be left to you to decide
whether to come back to Linux.

Mitch

In <2n8laa$...@lace.Colorado.EDU> atk@agua (Alan Krantz) writes:

(c.o.3.m has seen a lot of discussions ?flamage? re shlibs, so at the
risk of making myself a lightning rod:)

Both *BSD in their -current flavour have Sun-style shared libraries
(written by Paul Kranenburg), and these shlibs are available in
FreeBSD-1.1BETA.

Geoff.
--
 Geoff Rehmet, Computer Science Department,   | ____   _ o         /\
 Rhodes University,  South Africa             |___  _-\_<,        /\/\/\
   email : c...@cs.ru.ac.za                   |    (*)/'(*)    /\/\/\/\/\
         : ge...@neptune.ru.ac.za             |

Odd, considering that last time I rebuilt the world on my FreeBSD machine,
it was done within roughly 50MB of disk space... (add 70MB for the full
source code).

Your figures seem really _way_ off the mark (or the *BSD teams have added
gigabytes of source code since the last release...)

--
Francois Berjon                         Francois.Ber...@gwynedd.frmug.fr.net

How good are all these? I tried the MSDOS emulation before but it was so full
of bugs and limited that I'd rather reboot.

Linux is full of these experimental features and they keep on piling them
quickly without much thought. Soo Linux will be overwhelmed. It now has many
versions of file systems but still none with the capablility of FFS because
Linus hate it.
        The last time I studied Linux networking code, it was adapted Net/2
code with hacks to make it work for Linux. Otherwise how could they get those
networking utilities up so quickly.

Don't be fooled by features that you do not  need. Keep it slim and bug free.

Linux is ONLY for X86, whereas BSD is for ALL. If you want a toy to play with,
go for Linux, but if you want to go on to greater things, try BSD.
SABAH is HEAVEN. Beautiful islands, mountains and jungles are next to 5 star
hotels. There are no natural and very few man-made disasters,
               BUT for how long will it last?
Disclaimer: I only speak for myself

In <2n9f90$...@great-miami.iac.net> skyh...@iac.net (Chris Thompson) writes:

Believe me, it's far easier to say "make world" every month or two, and
back in the morning to find everything sitting there installed - new.  This
is a lot easier than trying to track all of the bugs that are fixed, and
all of the components that are upgraded replaced.
(Those people who receive the commit messages for FreeBSD will attest to
the number and rate at which changes come in. - I assume the same applies
to NetBSD.)

As I said - far easier than tracking every bug-fix.

An MSDOS emulator for *BSD is under development.  WINE is the same across
Linux and *BSD.  As far as iBCS2 binary compatibility goes - I'm sure that
if someone were interested in developing that there would be interest ;-)

Geoff.
--
 Geoff Rehmet, Computer Science Department,   | ____   _ o         /\
 Rhodes University,  South Africa             |___  _-\_<,        /\/\/\
   email : c...@cs.ru.ac.za                   |    (*)/'(*)    /\/\/\/\/\
         : ge...@neptune.ru.ac.za             |

I play wolfenstein in it quite regularly, use it to access a novell network
and with the exception of 386 protected mode DOS stuff its good. I've never
for example got netware clients to run inside of the interactive syste,.
Linux has had no new features added for well over a month to help refine a
stable 1.0 release. The new stuff will start to go into 1.1, and we will
all be running our disks twice as fast and running SCO binaries if we wish.
Despite that at least one machine here will stay running known stable releases.
Linux doesn't support FFS because nobody felt like wading through pages of
turgid BSD specific code that might in fact belong to USL anyway (and one
or two files it turned out did and will need rewriting for BSD4.4 lite.). I'd
be interested when/if someone does a BSD FFS for Linux to benchmark it against
ext2. At the moment with the two machines here with fast scsi disks and adaptec
cards Linux + clustering ext2fs is faster than BSD FFS is faster than Linux
without clustering. Some of this will be very application dependant.

There is an old saying that Usenet works best when people who don't know what
they are talking about shut up.... You obviously know so little about Linux
that you are contributing nothing of use.

Fascinating. The Linux network code was written from scratch by a large group
of people. It has no Net/2 code in it at all. A lot of the utilities are
BSD based for two reasons
1) While often not too portable (BSD rather than Posix tty etc) most of the
BSD networking tools are quite good and worth the porting effort
2) Several of them are not even described properly by RFC's because the
older BSD hackers (pre 386BSD) never bothered documenting things - even
the lpr protocol was documented afterwards by someone else.

What will also be interesting is seeing how well BSD networking and Linux
networking adapt to IPng.

Thats good advice. Linux is moving towards loadable modules. BSD has gone
from a fat splodge of code to quite clean refined code.

Linux is X86 and 68K under development. When the 68K port is finished we will
have the portability. This is following the path Unix (thus BSD) took very
closely. V7 was not portable until people started doing ports to System3 and
friends (There is a good CACM paper on this).

Alan

You misunderstand. Thats what it would take to build Linux _ all the tools
and packages I have from scratch (ie gcc,g++,gdb,X386,Khoros,Magic,gpc(alpha),
numerous kits of our own, xntp, [and so on]).

Alan

Imminent death of Linux predicted.

Exactly what capabilities are you talking about?

Interesting.  None of the kernel code is BSD (BNR/2) derived, except for a few of
the constants.  Some of the net utilities are BSD derived.  From what I've seen,
there are changes to make them work for Linux, but not every deviation from the
one true BSD way is a hack.

Oh, *BSD are bug free?  And SLIM?  Care to back up your assertion with some
proof?

Actually, the Linux port to the AMiga (and theoretically to other 68k
machines) is progressing at a fast clip.  Although I admit NetBSD (but neither
of the other free *BSD) supports many platforms with varying degrees of
completeness, it certainly doesn't run on "ALL."

--
 _g,  '96 --->>>>>>>>>>   gt81...@prism.gatech.edu  <<<<<<<<<---  CompSci  ,g_
W@@@W__        |-\      ^        | disclaimer:  <---> "Bow before ZOD!" __W@@@W
W@@@@**~~~'  ro|-<ert s/_\ nders |   who am I???  ^  from Superman  '~~~**@@@@W
`*MV' hi,ocie! |-/ad! /   \ss!!  | ooga ooga!!    |    II (cool)!         `VW*'

: Perhaps that can be rephrased to say "I have a long way to go before
: I'm ready for *BSD". :-)

: *BSD is an operating system. kermit is an application. XFree86 is also
: an application as is TeX and many other things.  Should every conceivable
: application be shipped with the operating system?  

: As far as 'tip' goes; typing "man tip" points you in the right direction
: for setting up the "remote" and "phones" files.

: What if I don't *want* my users to have POP accounts?

...among other garbage.

At any rate, your post didn't tell me anything I didn't know, nor
anything I didn't figure out BEFORE I ever tried to install *BSD.
My entire post claimed ONLY that I felt much more comfortable with
a installation package like Slackware than I did with the *BSD
install.

As I am STILL having problems with Linux' networking code, I'm
STILL hoping someone will put together such a beast.  But until
there is a way to painlessly install *BSD AND lots of utilities
I need, it's not for me.

Thanks for the lecture, but next time, at least say something
that hasn't been said.

Rob

In article <2neomp$...@clarknet.clark.net>,

I guess the concept of "no pain, no gain" is foreign to you?

I've noticed that everything in life has positive and negative aspects.  Linux
has a much nicer install procedure and supports a wide variety of interface
cards; BSD has better networking (so I've heard).  You can see that each of
the different groups has focused their energies in different spots.  Many
people don't have any problems installing *BSD on their systems; it looks like
the problem is with you.

To be honest, it's hard for me to have any sympathy for you; here you have
two FREE operating systems to choose from, with FULL source, and when something
doesn't work all you can do is whine about it.  Why don't you try fixing some
of these problems yourself?  What, you don't know how?  Then learn.  If you
want your hand held, then get a commercial operating system.

--Ken

:
:>Linux is full of these experimental features and they keep on piling them
:>quickly without much thought. Soo Linux will be overwhelmed. It now has many
:
:Imminent death of Linux predicted.
:
:>versions of file systems but still none with the capablility of FFS because
:>Linus hate it.

    Personally, having worked on BSD systems for years, I prefer Linux.
    BSD has always felt, well, stuffy.  From a comparative standpoint,
    at least for PC-based UNIXs, Linux is the most compatible and one is
    likely to see drivers for new cards developed on it before anything
    else.  Also from a comparative standpoint, BSD-specific code tends to
    be rather archaic... a lot of it is still K&R C (rather than ANSI C),
    and a lot of it tends to makes BSD-specific assumptions for system
    calls that are incompatible with ANSI C.

    No, thank you.

    From a stability standpoint, the only reboots I do nowadays occur when
    I install a new kernel.

    Considering that Linux was alive and well long before it could have
    been said to be reasonably complete, now that it IS reasonably complete
    it is highly unlikely that it will undergo an imminent demise.

    That said, there ARE a few exceptions when it comes to BSD code... I'm
    running sendmail 8.6.8 and it compiled without a hitch under Linux.

    The primary difference between BSD and Linux, apart from Linux's
    SYSVish syscall and tty interface, is in the support programs.  BSD
    has a very predictable set of programs, whereas Linux has a melange.
    For example, there are at least four different getty programs for
    Linux, two major system layout configurations (SysVish or BSDish),
    two or three different password-related utility sets (though that's
    finally been normalized with the incorporation of shadow passwords
    in the official shared C lib), etc.  

    There are also a whole lot of kernel enhancement patches, which I believe
    was a sticking point for many arguing for BSD.  It really isn't... most
    people do NOT bother with the enhancement patches until after they have
    been officially incorporated into the kernel by Linus, at which point
    they become the standard.  On the otherhand, if you need a particular
    feature instantly, there is a path you can follow to get it.  This
    path does not exist under a BSD system.

    You have to be a bit more knowledgeable to setup a Linux system than a BSD
    system.  I believe the result is well worth it.

                                                -Matt

--

    Matthew Dillon              dil...@apollo.west.oic.com
    1005 Apollo Way
    Incline Village, NV. 89451  ham: KC6LVW (no mail drop)
    USA                         Sandel-Avery Engineering (702)831-8000
    [always include a portion of the original email in any response!]

Best learning experience I ever had was moving from a single large
root partition to a small root partition and large /usr partition,
when I decided that I was going to figure out what the heck was going
on and move everything over by hand, rather than reinstall.  Took an
entire evening, but well worth it.

I fully agree.  Both OS's are excellent.  I use Linux because at the
time I made my decision, none of the BSD variants had dynamically
linked libraries, and now I've got too much invested to switch for
what would be minor advantages and disadvantages to me.
--
Eric J. Schwertfeger, man...@cs.unlv.edu

       Personally, having worked on BSD systems for years, I prefer Linux.
       BSD has always felt, well, stuffy.  From a comparative standpoint,
       at least for PC-based UNIXs, Linux is the most compatible and one is
       likely to see drivers for new cards developed on it before anything
       else.  Also from a comparative standpoint, BSD-specific code tends to
       be rather archaic... a lot of it is still K&R C (rather than ANSI C),
       and a lot of it tends to makes BSD-specific assumptions for system
       calls that are incompatible with ANSI C.

While I highly respect some of the work you have done on the Amiga
front, this statement leads me to believe that you've never really
looked at the code you're criticising.  Do you really think we'd
settle for non-ANSI compliant code?  Much of both the FreeBSD and
NetBSD teams' effort has been in adding extensive prototyping, and we
run the entire codebase through `gcc -Wall' periodically.  The
mainline efforts in *BSD are anything but archaic, and this strikes me
as simply more of the same unthinking bigotry that people in both
camps periodically exhibit.  It's both innaccurate and unnecessary.

                                Jordan
--
Jordan K. Hubbard       FreeBSD core team       Electric Bivalves Anonymous
On the net, no one can hear you scream.

In article <HJSTEIN.94Mar24111...@sunset.huji.ac.il> hjst...@sunset.huji.ac.il

        I have seen this thread go on and on, and the big question that I could
never figure out is why anyone would want to recompile their entire system from
scratch.

        The one thing that I came up with was that with *BSD until quite
recently there were no shared libraries, so any change to libc required that
you completely rebuild the system if you want to have the change propogated to
all of your binaries.  Also, I suspect that the people on the BSD core team and
the people following the Free-current source tree (i.e. using the shared
libraries) would want to do a complete rebuild time a problem related to the
new shared libraries was found.

        It was quite similar just over a year ago when the first DLL linux
shared libraries were being developed.  You would want to recompile everything
so that you could test the system and see if you could find any bugs.  Now that
the linux linker is stable, there is simply no need for massive recompilations
of the entire system.

        I am not arguing that there is anything wrong with a unified source
tree - it does have some advantages and it is a worthwhile goal, but personally
I thiuk that the advantages are not as big as some of the BSD people would lead
you to believe.

-Eric

--
"The woods are lovely, dark and deep.  But I have promises to keep,
And lines to code before I sleep, And lines to code before I sleep."

In article <JKH.94Apr2120...@whisker.hubbard.ie> j...@whisker.hubbard.ie (Jordan K. Hubbard) writes:

    Well, I wasn't refering to FreeBSD, but rather the Berkeley BSD release
    and the Berkeley source tree (which I have extensive experience with).  
    You are correct in saying that I have not looked at the FreeBSD release,
    and I *did* assume it was a straight port of BSD for which I apologize.

                                    -Matt

--

    Matthew Dillon              dil...@apollo.west.oic.com
    1005 Apollo Way
    Incline Village, NV. 89451  ham: KC6LVW (no mail drop)
    USA                         Sandel-Avery Engineering (702)831-8000
    [always include a portion of the original email in any response!]

In article <2nf0fo$...@sbus.entropic.com>,

Sorry.  I dislike these flame wars intensely, and I just happen to find
the above "concept" one of the more disgusting ones ("Hey, I'm *macho*,
I use an operating system that is PAINFUL because that way I gain
much").

Next I guess you'll tell me "you get what you pay for".  I haven't been
this close to flaming someone in a *long* time,

                Linus

PS.  I tested out Slackware a few weeks ago, and my reaction was "hey,
this install looks like a DOS program".  I'm not ashamed to admit that I
was immensely pleased.

In article <2nm6tb$...@klaava.Helsinki.FI>,

Unfortunately, in case *BSD, "Lots of pain, no gain" to most people.

I strongly agree with Linus. It is like a car mechanic saying the same
thing : "No pain, no gain"  with a problem car, implying that the driver
should know how to rebuild the engine, how to fix when oil leaks, etc...
In fact, if the instructions are given correctly, I can do it, but
in case of *BSD, most the instructions are either outdated, incorrect, or
not clear.

Don't worry. For some people, "No flame, no gain".  :)

In fact, I installed it a week ago freshly, and it went very
smooth.  I think that the total installation procedure was much more
simpler than SUN workstation's case.
But when I hooked it up with network, I felt that the network performance
was not really satisfactory.  I felt it quite sluggish than even SUN3s.
(Linus, do you possibley know why ?).

Here is my experience in installing NetBSD.  I somewhat succeeded, so
if people are interested in, they may continue to read.

I think that installing DOS/UNIX in one disk is quite typical, and
if you read the volumenous 386BSD FAQ or succinct NetBSD Install guide,
nothing is clear about it.

First I followed the NetBSD Install guide : It doesn't work.  So, I tried
installing in whole disk. In case it works.  But I really needed DOS
partition, I don't have enough money just like standard *BSD users, and
I feel that 540M is enough for my use (for DOS and NetBSD together).

So I read 386BSD FAQ, and there is a very interesting (quite amusing,
and looks like a joke to Linux users.  Every Linux users should see
386BSD FAQ 2.5.3 - how to install multiple OS in one drive).

How good if that hassle would just work. If you followed the instruction
correctly, it doesn't work either.

After wasting lots of time in figuring out using trial-and-error method,
I finally found a way how to install DOS/*BSD in one IDE drive :

--------------

0) Preparation :
   You must have the following available :
       pfdisk  (from linux dosutils in tsx-11).
       fdisk  from DOS 5 or higher.
       Bootable DOS.
       NetBSD Kernel disk, file1 disk, file2 disk.
   You must back up any important data in your HD.

1) Since NetBSD wants you to start NetBSD partition to start from
   cylinder boundary ( Be careful !  The installation instruction tells
   you to prepare calculator and use math for cylinder/head/sector stuff,
   but it will eventually wipe out DOS partition),
   use pfdisk to partition the disk in cylider number AND to edit the
   partition id (you can give a5 to NetBSD in later step).

2) If this is not your first try and you were screwed up previous time,
   you will need this step.  For safety, you can just include this step.
   a) Use fdisk, and see the partition table.
   b) Write the partition sizes in Megabytes on a scatch pad
            as seen in the table.
   c) Do 'fdisk /mbr' to remove NetBSD boot which possibly exists.
   d) Use fdisk and remove all partitions. And reboot.
   e) Use fdisk to repartition the disk (now you need the scratch pad
      in b)).  Use the partition size in Megabytes in b) to partition
      the disk.  Reboot.
   f) Use pfdisk and see partition table.  
      For your safty, check whether it started from the cylinder boundary
      (as same as the numbers shown in 1)).

      Write all the informations
        - starting sector, partition size in sector, ...etc.
         on your scratch pad.
   g) Give partition id 165 (a5 in hex) to NetBSD using the pfdisk. Reboot.
   h) format c:/s/u (if you wish).

3) Use kernel diskette.
   Reboot the machine.
   ... the step is same until you copy the kernel in floppy to HD.....
   Insert file system 1 diskette.
   Use the informations you got from 2-f)  to prepare HD.
   Rboot with kernel diskette.
   >Copy
   >wd0a

4) Now, if you follow the instruction in NetBSD installation guide,
   it won't work.
   First boot with the kernel diskette.
   when the propmt waits on :-,  quickly (within 5 sec) type
   wd(0,a)/netbsd, which means you are booting from the copied kernel
   in HD (the root partition).
   It will ask you to insert the file system 2 diskette.
   And just follow the simple instructions.

5) When you next boot, do the same thing :
     type wd(0,a)/netbsd.
   If you want single user mode, type 'wd(0,a)/netbsd -s' while the
   prompt is waitin on :- (for a few seconds).

6) Now you may continue to install using 'extract' command.

7) You may wish to hack around the boot sector using os-bxxx, but
   in my case, I was afraid of losing what I have done, so I decided to
   make boot diskette for myself.

8) However, in next step - to update the old 0.9 version to Current,
   The FAQ's procedure (3.1.8) led me to "Bus error. core dump".   :(
   It seems to me that I should give up.
   Too bad that I realized too late that I shouldn't start *BSD stuff at all
    :(

   In fact, if the instructions are given correctly, I can do it, but
   in case of *BSD, most the instructions are either outdated, incorrect, or
   not clear.

Oh dear.  I think it's time for me to get into some serious capital
letter usage here!

I find it somewhat annoying that you say `*BSD' here and then proceed
to later admit that you've only tried NetBSD.  Not to flame NetBSD at
all, but each group is different and has different strengths and
weaknesses.  Please do not tar us all with the same brush!

That aside, and speaking more generally, many of the instructions
WON'T GET ANY BETTER IF PEOPLE DON'T HELP US TO IMPROVE THEM!  It
never ceases to amaze me how many people seem to regard 100%
perfection in something they have not paid one penny for as a
god-given right, and for every one person that sends in constructive
criticisms that we can _WORK WITH_, ten more simply whine at us for
not doing a `better job', never even stopping to consider how they
might contribute to the effort themselves!

It also seems to me that Linux is blessed with the twin advantages of
both a very large user base (several times larger than FreeBSD's) and
one that is willing to PITCH IN AND HELP OUT.  In the FreeBSD group,
at least, we are a VERY SMALL group of SERIOUSLY OVERWORKED
individuals who are now giving up what amounts to pretty much *ALL* of
their free time to bring the general public these releases!  I ask
you, would you yourself feel perhaps just a little bit bitter about
the amount of flaming and destructive comparison that goes on here if
you were in my shoes?

Even in the worst situation, where one has tried hard to install some
version of *BSD and had no luck, what do you think that the best thing
to then do is?  Jump on the net and say "xxxBSD is a heap of flaming
horse exhaust!" or drop one of our frequently published mailing lists
a line saying "I've had this bad problem trying to install, can anyone
help me?  Is there anything I can do to help make it work?"  I think
the answer's pretty clear.  We're not commercial software
organizations here, and we can't do it without public help and
support.  If we get fed up and go away, where will that have gotten
everyone?

                                        Jordan

--
Jordan K. Hubbard       FreeBSD core team       Raving lunatic

J> without public help and support.  If we get fed up and go away,
J> where will that have gotten everyone?

Well, everyone will be using Linux.

-Magnus

As a long-time Linux user, I'd like to apologize to the *BSD community
for this person's stupid snide comment.

Frankly, I think it's good in the long run that we now have _3_
distinct evolving free-unix communities --- the cross-fertilization
will be good for all three, at least for a while.  Their work does us
all good, and the loss of one community will harm us all.

Please don't stop.

        -Randy