From: Borja Marcos <borjamar@sarenet.es>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] What makes Plan 9 unique?
Date: Fri, 26 Oct 2001 17:36:11 +0200 [thread overview]
Message-ID: <100411057301@192.148.167.16> (raw)
In-Reply-To: <9r783d$s5l$1@newpoisson.nosc.mil>
On Thursday 25 October 2001 11:00, you wrote:
> What makes Plan 9 truly unique? What about it makes it better than
> Solaris, or other UNIX systems?
I don't think Plan 9 can or should be compared to Unix systems. Plan 9 is
a distributed system, and Unix is not.
One of the things that impressed me the most the first time I read a
description (some years ago, long before AT&T decided to sell research
licenses), was its capability to work in *heterogeneous* networks. It is
perhaps the most significant feature in Plan 9.
Another distributed operating system, Amoeba, converts a whole network
into a sort of huge multiprocessor computer (from the user's point of
view), but it does not take into account the differences between networks,
so Amoeba may be good for organizations with a high speed local network.
Plan 9 does not have automatic process migration like Amoeba, but, for
the designer of a distributed application, Plan 9 offers the ability to
decide how to use the network depending on the available resources. The
figure 1 in the original Plan 9 paper shows an example; CPU and file
servers linked by a high-speed optic network, and remote terminals working
from a phone line. The speeds can range from a 9600 bps line to a gigabit
network.
Plan 9 offers a per-process namespace that can be configured to make the
network topology completely invisible to the user or the application
program. Moreover, it has taken the "everything is a file" idea to the
extreme. The network services, for example, are seen as files. To open a
connection you only copy one "master" file, and the new file obtained will
be the new connection.
Together with this, Plan 9 offers a very clean network filesystem. It can
export *anything*. One of my favourite examples is a machine "mounting"
the TCP/IP stack of another through a network, or a process in one machine
being debugged by a debugger in another. Classic network filesystems don't
allow this flexibility.
The window system makes a heavy use of the "everything is a file" notion,
and it is extremely flexible and simple. It does not have a specialized
protocol for network operations such as X11. Offering a file interface, it
uses 9P. In fact, as the screen itself is a file, you can run one window
system inside a window.
This is what I liked most about Plan 9. There are obviously more things.
The user interface is not the typical interface you see in the rest of
environments, it shows new ideas.
I even love the documentation; a set of two books in the old Unix
fashion. A set of very well written papers describing the different
components and tools, and the manpages.
Summarizing, in a time when you don't see anything new from commercial
operating system vendors, and when everything is either Unix (note that I
am a heavy FreeBSD user) or (argghhh) Windows, an operating system with a
truly innovative design is fresh air ;-)
Borja.
next prev parent reply other threads:[~2001-10-26 15:36 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2001-10-25 9:00 Matt Senecal
2001-10-25 9:36 ` Lucio De Re
2001-10-26 9:25 ` Douglas A. Gwyn
2001-10-26 14:03 ` Matt Senecal
2001-10-26 15:36 ` Borja Marcos [this message]
2001-10-25 11:56 Russ Cox
2001-10-26 9:25 ` Douglas A. Gwyn
2001-10-26 15:01 Russ Cox
2001-10-26 16:48 ` Thomas Bushnell, BSG
2001-10-29 9:04 ` pac
2001-10-26 16:58 presotto
2001-10-29 10:16 ` Ozan Yigit
2001-10-29 20:54 ` Skip Tavakkolian
2001-10-30 16:50 ` Dan Cross
2001-10-26 17:09 forsyth
2001-10-29 1:56 okamoto
2001-10-29 13:07 bwc
2001-10-29 18:54 presotto
2001-10-29 20:54 David Gordon Hogan
2001-10-29 21:28 David Gordon Hogan
2001-11-05 14:59 ` Jonadab the Unsightly One
2001-11-06 10:22 ` Jonadab the Unsightly One
2001-11-05 22:14 David Gordon Hogan
2001-11-06 10:22 ` Jonadab the Unsightly One
2001-11-06 11:01 geoff
2001-11-06 16:45 Russ Cox
2001-11-06 17:53 ` Thomas Bushnell, BSG
2001-11-06 18:28 ` William Josephson
2001-11-07 2:46 ` Boyd Roberts
2001-11-06 17:08 anothy
2001-11-07 1:02 David Gordon Hogan
2001-11-07 6:34 Russ Cox
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=100411057301@192.148.167.16 \
--to=borjamar@sarenet.es \
--cc=9fans@cse.psu.edu \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).