[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[Victor]

On Aug 22,  2:18am, Vadim Antonov wrote:
 Subject: Re: [comp.os.linux.misc] Help wanted, Plan9 a piece of junk!
>My own research in O.S. architectures in an attempt to find
>a complete set of kernel primitives led to a discovery of an
>architecture which can work even on analog machines or support
>process communication by means of ICBM exchange :)

This was certainly a fad in OS design for a while, but the 
result indicates that the clean simple complete set of kernel
primitives may not exist. 

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[dhog]

>> You seek... a GRAIL?
>No I just have an open mind to the diversity of the world I live in.

``Why not indeed?''  -- Basil Fawlty

The OS which you are requesting information on is in fact called Grail.
I have a hard copy of a paper written by Vadim on my desk; I have had
it for a few years, and have quite forgotten how I obtained it.  Maybe
Vadim can help.  It's called ``In Search Of The Holy Grail''.

It seems like quite an interesting system.  Yet another example of taking
an idea to its logical extreme (just a different idea...)

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[Johnson]

> 
> You seek... a GRAIL?
> 
No I just have an open mind to the diversity of the world I live in.

thanks.jluqaz

-- 
Johnson Lukose                          jluqaz@hicom.po.my
HICOM Communications Sdn. Bhd.          Tel: +60 3 202 8800
Suite 3.5, Wisma HICOM                  Fax: +60 3 202 8899
Jalan U1/8, Off Persiaran Kerjaya
40000 Shah Alam
MALAYSIA

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[dhog]

>Out of interest you did not mention which OS you are refering to.
>Perhaps you could be so kind as to give a brief description.  Hope
>this is acceptable to the P9 group.

You seek... a GRAIL?

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[Johnson]

On Wed, 23 Aug 1995 11:13:46 -0400, "Victor Yodaiken" <yodaiken@sphinx.cs.nmt.edu> wrote:
> On Aug 22,  2:18am, Vadim Antonov wrote:
>  Subject: Re: [comp.os.linux.misc] Help wanted, Plan9 a piece of junk!
> >My own research in O.S. architectures in an attempt to find
> >a complete set of kernel primitives led to a discovery of an
> >architecture which can work even on analog machines or support
> >process communication by means of ICBM exchange :)
> 
Out of interest you did not mention which OS you are refering to.
Perhaps you could be so kind as to give a brief description.  Hope
this is acceptable to the P9 group.

thanks.jluqaz

-- 
Johnson Lukose                          jluqaz@hicom.po.my
HICOM Communications Sdn. Bhd.          Tel: +60 3 202 8800
Suite 3.5, Wisma HICOM                  Fax: +60 3 202 8899
Jalan U1/8, Off Persiaran Kerjaya
40000 Shah Alam
MALAYSIA

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[dhog]

avg@postman.ncube.com (Vadim Antonov) writes:
>The ideal system is "functionally complete" in regard to the
>class of applications.  Then there wouldn't be ugly extensions
>and creeping featurism.  Note that this approach is kind of
>contradictory to the pure minimalism, which is to do absoulte
>minimum to solve the problems at hand.

The only problem is that you need an infinite amount of storage.
A "functionally complete" system must be able to understand all
the file formats of the world (audio, video, data compression, MS
Word documents etc) including ones yet to be devised.  It must
contain all the applications that anyone ever wants to be able to
use a computer for, including ones yet to be devised.  No system
can provide all the functionality that users are going to want.

Look at X windows.  It is a system which _tries_ to be functionally
complete.  Yet every new release provides yet another batch of ugly
features that were "missing" in the last.  The toolkit & widget sets
try to do everything for the programmer, who is forced to spend
hours writing messy code which describes just what the widgets
should do, only to find that some bug in the libraries takes even
longer to find a workaround for.  This pattern is duplicated by many
other software systems, written by well meaning programmers
with your philosophy.

So Seventh Edition Unix didn't provide support for networking.
Can't really blame its authors, networking was pretty non-existant
in those days.  I suppose that in another 10-20 years, people like
you will be bemoaning the fact that the original versions of Plan 9
didn't include support for (say) direct brain interfaces, so that there
are multiple conflicting implementations of them...

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[Vadim]

>The only problem is that you need an infinite amount of storage.
>A "functionally complete" system must be able to understand all
>the file formats of the world (audio, video, data compression, MS
>Word documents etc) including ones yet to be devised.

Not. Check any math book for the definition of the term.

>Look at X windows.  It is a system which _tries_ to be functionally
>complete.  Yet every new release provides yet another batch of ugly
>features that were "missing" in the last. 

You obviously confuse completeness with complexity.  May
i suggest elementary text on a math logics?

The problems of X are exactly because they did poor design of
the system, and instead of rethinking it to make things simplier
and more orthogonal (loosely speaking, to look for orthogonal
basis) they choose to patch it by adding more complexity.

In a sense, functional completeness is very close to minimality,
since it is hard to build a system which is complete and
*not* minimal, as gaining confidence in completeness is becoming
harder for larger systems.

My own research in O.S. architectures in an attempt to find
a complete set of kernel primitives led to a discovery of an
architecture which can work even on analog machines or support
process communication by means of ICBM exchange :)

--vadim

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[forsyth]

>>that i can once again regard sparcs, 680x0 boxes, PCs, etc. as

>Is it anything new?  My first large-scale project (i was in a kind

no, that's why i carefully said `again'.  i intend to keep it this time,
though.

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[Vadim]

>Plan 9 has its own conventions that it adheres to on *all hardware
>platforms*.  this is a strength, not a weakness.
>as someone operating a plan 9 installation, i can tell you from
>experience that it is a real joy to have something that is so
>plug compatible in its interfaces and details of management
>that i can once again regard sparcs, 680x0 boxes, PCs, etc. as
>just a source of faster (or slower) computing power.

Is it anything new?  My first large-scale project (i was in a kind
of leading position in the project, it had no formal "organization"
behind it for a long time) was building a family of Unix-like systems
(called DEMOS, addreviation for "Interactive Unified Portable Operating
System").  It was running on eight different architectures (including
clones of IBM/370 and some machines with no Western prototypes), and
was built from the same source tree (modulo hardware dependencies and
stuff like support for 3270s).  Needless to say all machines were
"the same".  It also was (and still is) the only effort in complete
internationalization of Unix.

I guess it is more like that Plan 9 has a single organization
producing releases.  If the system will be successful it is
inevitably going to change as i do not suppose that Bell Labs
people will be interested in supporting the commercial releases
(it is a hell lot of tedious work).

When it will happen the history with Unix will repeat itself,
as the root cause of the divergence of revisions wasn't fixed.
That root cause is the functional incompleteness.  Nobody in 
his own mind makes systems incompatible for the sheer hell of it.
Rather, people add things to fix their particular problems which
weren't adequately addressed in the original system.

The ideal system is "functionally complete" in regard to the
class of applications.  Then there wouldn't be ugly extensions
and creeping featurism.  Note that this approach is kind of
contradictory to the pure minimalism, which is to do absoulte
minimum to solve the problems at hand.

--vadim

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[forsyth]

several people have commented on the reliability of the system.
it is wrong to conclude from installation problems (or even some
operational problems on the PC) that plan 9 is an `alpha' or
`early version' of something, and that it is unsurprising to have it
fail regularly.  once installed on (or adapted to) a particular hardware
configuration, most of the software that is running is software with
a good few years behind it.  the bulk of the code is portable code
that has run on many different platforms, not just the PC.
it certainly is not error free, but it is far better than `alpha' or `beta' quality.

of course, there are newer, more experimental parts, that
are less well tried, but as in the original Unix systems,
the documentation is forthright about potential problems and limitations.

as one example, see the health warnings in the documentation
for Mothra and the Panel library on which it is built.
nevertheless, i've been using the Panel library
in a program i'm building and i regret to say that the errors have
been in my code, not td's.  (it is a very nice package, by the way.)

in general, compared to most other systems i know, there is typically several orders of magnitude less
source code to do more work (and more interesting work at that),
and consequently there is less space for the bugs to hide.

you really do have to read the documentation, though.
if you approach the system with too many preconceptions about what
it `obviously' will do, it will trip you up.  in particular,
the PC is just a hardware platform on which Plan 9 runs.
it isn't the system's primary platform (it hasn't got one!).

Plan 9 has its own conventions that it adheres to on *all hardware
platforms*.  this is a strength, not a weakness.
as someone operating a plan 9 installation, i can tell you from
experience that it is a real joy to have something that is so
plug compatible in its interfaces and details of management
that i can once again regard sparcs, 680x0 boxes, PCs, etc. as
just a source of faster (or slower) computing power.

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[pete]

The irony here is that the current plan9 release is doing precisely the things
Linux was doing back in the 0.11/0.12 days.... yet people seem to expect some
magically reliable piece of software.

pete (sitting there looking at a pile of diskettes and a nice empty IDE drive,
      and considering making the plunge tomorrow afternoon!)

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[Scott]

dhog@plan9.cs.su.oz.au writes:
| (2) open() can set errstr as a side effect, even when the call succeeds. 

Ouch:  ``...it is understood that the error string is altered only if an
error occurs.'' -- intro(2).

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[Steve_Kilbane]

forsyth writes:
> anyhow, i expect in consequence to see lots of articles about Plan 9 in
> the local computing press!

Possibly, which is a shame, because I get the impression that the Plan 9
release is an attempt to get back to the early days of UNIX, with all
that that entailed: You want a copy? Here's a dump of our system; it's
up to you to figure out how to get it going at your end.

Personally, I can't complain about how things are. Plan 9 is supposed
to be unsupported, but you can't deny that the plan9.att.com folks have
been helpful on this list.

There appear to be two basic mistakes, here:
	- the 4-disk installation seems to cope for most people, but in
	some cases, it's out of its depth, and causes damage. Personally,
	I was surprised that so much effort had been put into the PC
	installation arrangement in the first place. Perhaps the PC
	program should have asked for more confirmation?

	- people still aren't backing up their disks. I know it's not
	much comfort to say things like this when folks have just
	trashed months of work, but to be honest, this is often the
	only way that many people learn. In my case, I *removed* the
	disk from the machine; no way was I going to risk losing data,
	just because the software did something I didn't expect.

I'm probably not expressing my point too well here, but what the hell.
Think of Plan 9 as an Alpha, because that's near enough the case. Remember
early versions of, oh, Mosaic or Netscape? Crashes all over the place. Well,
this is a similar situation, except Plan 9 is an OS, and when an OS goes
down, well, you'd just better have done your backups.

If this sounds negative, sorry. It's not supposed to be. I just get a
bit annoyed when people start treating all software as the same.
Plan 9 != gcc != Microsoft Word, so don't expect them to have similar
attributes.

steve, somewhat incoherent

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[dhog]

>Love that subject...

He can't even spell Linux...  ;-)

>Since everything in Plan 9 is a file, the message 'file does not
>exist' is probably the least useful and most seen message.

I've noticed that sometimes the message is printed even when
there is no actual file that couldn't be opened.  The reason for
this appears to be:

(1) Some programs use errstr when reporting errors which don't actually set errstr
(2) open() can set errstr as a side effect, even when the call succeeds.  This is
  due to the implementation of union directories -- walk() tries each directory
  in a union in turn until it succeeds, and errstr is set to 'file does not exist' as
  a side effect of a failed attempt.  Thus any program which, say, uses files in
  /dev, will have errstr set to 'file does not exist' instead of 'no error'.  In fact,
  it's worse than that, since "/" is a union directory, opening any file in the
  namespace will set errstr!

I'm not sure what the best way to fix this is.  Ideally, open() should not change
errstr at all if it succeeds, but this would require wasteful copying of errstr in
every call to walk() on a union directory.  The alternative is to just blindly clear
u->error[0] on a sucessful open() -- cheaper, but is it The Right Thing?

Anyway, here is a program to demonstrate the problem:

-----------------
#include <u.h>
#include <libc.h>

void
try(char *file)
{
	char err[ERRLEN];

	strcpy(err, "no error, dude");
	fprint(2, "\nbefore opening %s, set errstr = '%s'\n", file, err);
	errstr(err);
	if (open(file, OREAD) < 0)
		fprint(2, "open %s failed!: %r\n", file);
	else
		fprint(2, "after opening %s, errstr = '%r'\n", file);
}

void
main(int argc, char *argv[])
{
	fprint(2, "on program entry, errstr = '%r'\n");

	try("/dev/null");
	try("/adm/users");
	try("#p/1/status");
	exits(0);
}
--------------------
Here's the output when I run it:

on program entry, errstr = 'file does not exist'

before opening /dev/null, set errstr = 'no error, dude'
after opening /dev/null, errstr = 'file does not exist'

before opening /adm/users, set errstr = 'no error, dude'
after opening /adm/users, errstr = 'file does not exist'

before opening #p/1/status, set errstr = 'no error, dude'
after opening #p/1/status, errstr = 'no error, dude'
--------------------

P.S.  If Plan9 is junk, what does that make Windows 3.1?

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[dhog]

Perhaps in the interests of world peace, the versions of disk1 on the ftp
site (and mirrors) containing the original disk/prep should be made
inaccessible?  Though that still leaves the physical copies that have been
(or will be) sent out to licensees -- will the Labs be issuing some sort of
errata sheet/warning with bought copies of the distribution?  I mean, these
people are potentially less likely to look at the ftp site's errata list or the
README in /plan9/pcdist

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[forsyth]

	You should be glad that bridge fell down --
	I was planning to build 13 more to the same design.
		Remark attributed to I.K. Brunel, addressing the
		Directors of the Great Western Railway

(i think i saw it first in one of P J Brown's books, but i'm not sure.)

i've been told that windows 95 will also scribble over a linux partition
given half a chance.  clearly, people have it in for linux.

anyhow, i expect in consequence to see lots of articles about Plan 9 in
the local computing press!

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[Bill]

"l" == lsson  <Ture> writes:

  >> >[...]
  >> >I want to recover my more than half-year work, since I don't have tape
  >> >backup.
  >> 

  l> Seriously, I don't like the way the 4-disk set is supposed to
  l> magically figure out what partition on my disk to install on. I have a
  l> fairly complicated partition setup, and I want to tell Plan 9 to "go
  l> onto /dev/hda3, period!" (Or whatever Plan 9 calls the third partition
  l> on the first disk). ...


    Absolutely.


    Bill

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[presotto]

Love that subject...

Since everything in Plan 9 is a file, the message 'file does not
exist' is probably the least useful and most seen message.
It would be nice to know what file is not copying over.  Also,
it would be helpful to see what if anything does manage to get
into c:\plan9.

You don't happen to have a DOS compressed file system do you?

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[Greg]

> Seriously, I don't like the way the 4-disk set is supposed to
> magically figure out what partition on my disk to install on.

For what it's worth, after testing the Plan 9 4-disk set on a machine
here at MIT, we also noticed the Linux partition wiped, even though we
had cleared up room on the first IDE disk after the DOS partition.
(We're good about not keeping important data on non-backed-up storage,
though; there wasn't really anything there besides the operating
system.)

Bad business, very bad.

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[jmk]

But wait! There's more!

	Path: alice!allegra!uunet!in2.uu.net!news.ios.com!tribeca.ios.com!hvaidya
	From: hvaidya@tribeca.ios.com (Hemant Vaidya)
	Newsgroups: alt.uu.comp.os.linux.questions,comp.os.linux.misc
	Subject: Installing Plan9 corrupted Linux root partition
	Date: 11 Aug 1995 18:12:59 GMT
	Organization: Internet Online Services
	Lines: 43
	Message-ID: <40g6jb$raj@news.ios.com>
	NNTP-Posting-Host: tribeca.ios.com
	X-Newsreader: TIN [version 1.2 PL2]
	
	
	Hi Linux Experts,
		
		When I installed Plan9 on my PC it seems to have corrupted root
		partition on my drive. I get following error when booting from 
		LILO or floppy. DOS partition seems OK.
	
		The kernel version is 1.2.0 gcc 2.6.3
		PC configuration:
		P90, 16MB,1MB SCSI-II, BusLogic PCI SCSI-II controller (BT946C)
	
		Partition check:
		sda: sda1 sda2 sda3 sda4 <sda5>
		[MS-DOS FS Rel. 12, FAT 102,check=n,conv=b,uid=0,gid=0,umask=022,bmap]
		[me=0x10,cs=0,#f=0,fs=0,fl=0,ds=0,de=53505,data=0,se=32798,
		ts=-63599110,ls=252]
		Transaction block size = 512
		Kernel panic: VFS :Unable to mount root fs on 08:02
	
		sda1 is DOS primary starting at 1 to 100
		sda2 is root partition starting at cylinder 101
		sda3 is swap 
		sda4 I think was created by Plan9
		sda5 is DOS logical drive D:
	
		I can still boot from slackware distribution without mounting root
		and using ramdisk.
	
		If I then go into fdisk  (linux) I see partition info as before.
		If I try to mount /dev/sda2 /mnt I get error: /dev/sda2 not mountable
		something like that.
		
		I can mount sda1 under /mnt without problem.
	
		Is there a way to salvage my root partition or fix the problem. 
	
		If you think more info is needed I can try and get it.
	
	
		You can reply at: hvaidya@tribeca.ios.com
	
	--Hemant

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[Ture]

> >[...]
> >I want to recover my more than half-year work, since I don't have tape
> >backup.
> 
> HAHAHAHA!!  Stop it, you're killing me!

Seriously, I don't like the way the 4-disk set is supposed to
magically figure out what partition on my disk to install on. I have a
farily complicated partition setup, and I want to tell Plan 9 to "go
onto /dev/hda3, period!" (Or whatever Plan 9 calls the third partition
on the first disk).

However, I haven't managed to get as far as actually installing the
thing. During the fist stage, copying files to C:\PLAN9, it says
"problem copying file: file does not exist" about a file which most
certainly is on the floppy -- both DOS and Linux can read
it. (Naturally, I left my notes at home so I don't remember the exact
name of the file...) After rebooting into DOS, there's no trace of
Plan9 on C:, but when I reboot from the first Plan9 floppy it tells me
that Plan9 is already on my disk, and asks whether I want to
reconfigure.

I'm beginning to suspect that the installation program somehow gets
confused by my setup, and tries to copy the floppy onto itself, or
something similarly stupid. Is that possible?

I _am_ two megabytes showt of RAM (i.e, I have only 6 MB). Is 8 Mbytes
a "hard" limit, or is it just a "it-will-run-but-not-very-fast" limit?
(Even Solaris can run in 8 Mbytes, and I though Plan9 was supposed to
be small?)

My setup is: Noname asian-made 386+387sx, AMI BIOS, Trident 8900C, two
Seagate IDE disks (102 + 545 Mbytes, on the same controller), one 3.5"
(known by dos as A:) and one 5.25" floppy.

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[David]

> HAHAHAHA!!  Stop it, you're killing me!

It's unfair to mock the afflicted.

I must say it's nice they're teaching basic logic so well at Rutgers:

Premise 1: I never back anything up.
Premise 2: I don't understand the difference between an async port and my
elbow.
Conclusion: the operating system sucks.

Very sound.
	Dave.

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[dhog]

>From: xke@paul.rutgers.edu (Xiao Ke)
>Newsgroups: comp.os.linux.misc
>[...]
>I want to recover my more than half-year work, since I don't have tape
>backup.

HAHAHAHA!!  Stop it, you're killing me!

>Please help! Any suggestion will be mostly appreciated.
>I guess this is doable, since my linux partition occupies 400Mbytes, 
>while plan9 only occupies 20Mbytes according to its installation notes,
>so it looks like the rest 380Mbytes should be able to recoverd, sounds
>reasonable?

Maybe he should try Norton Utilities <snicker!>

[comp.os.linux.misc] Help wanted, Plan9 a piece of junk!

[Bill]

------- Start of forwarded message -------
From: xke@paul.rutgers.edu (Xiao Ke)
Newsgroups: comp.os.linux.misc
Subject: Help wanted, Plan9 a piece of junk!
Date: 15 Aug 1995 00:10:54 -0400
Organization: Rutgers University LCSR
NNTP-Posting-Host: paul.rutgers.edu

Hi, there,

Help wanted, Plan9, the so boasted next generation os is just a piece of
junk. This time, AT&T really picked a good name for plan9.

I've run linux1.2.0 for about a half year, everything seems OK.
Today, I download the plan9 to play with, it looks working too.
However, actually when I was installing plan9, it overwrote my linux
partition without any pre-warning.

My linux partition is not bootable now, and I try to boot from floppy,
and do fsck on it, it says: "bad magic number". So definately
plan9 overwrote my linux partition.

I want to recover my more than half-year work, since I don't have tape
backup. Please help! Any suggestion will be mostly appreciated.
I guess this is doable, since my linux partition occupies 400Mbytes, 
while plan9 only occupies 20Mbytes according to its installation notes,
so it looks like the rest 380Mbytes should be able to recoverd, sounds
reasonable?

Thanks,


-----Xiao Ke
xke@paul.rutgers.edu 
------- End of forwarded message -------