From owner-freebsd-chat Mon Dec 17 2:49:29 2001 Delivered-To: freebsd-chat@freebsd.org Received: from scaup.prod.itd.earthlink.net (scaup.mail.pas.earthlink.net [207.217.120.49]) by hub.freebsd.org (Postfix) with ESMTP id 08C5837B417 for ; Mon, 17 Dec 2001 02:49:25 -0800 (PST) Received: from pool0020.cvx21-bradley.dialup.earthlink.net ([209.179.192.20] helo=mindspring.com) by scaup.prod.itd.earthlink.net with esmtp (Exim 3.33 #1) id 16FvKc-00073M-00; Mon, 17 Dec 2001 02:49:15 -0800 Message-ID: <3C1DCDAC.CEA3DEAF@mindspring.com> Date: Mon, 17 Dec 2001 02:49:16 -0800 From: Terry Lambert X-Mailer: Mozilla 4.7 [en]C-CCK-MCD {Sony} (Win98; U) X-Accept-Language: en MIME-Version: 1.0 To: Anthony Atkielski Cc: "f.johan.beisser" , FreeBSD Chat Subject: Re: UNIX on the Desktop (was: Re: Why no Indians and Arabs?) References: <20011216112759.U16958-100000@localhost> <002f01c1866e$1e4d9510$0a00000a@atkielski.com> <3C1DB7EB.9232204A@mindspring.com> <001101c186dd$5ab94430$0a00000a@atkielski.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Sender: owner-freebsd-chat@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org Anthony Atkielski wrote: > UNIX was designed as a multiuser, text-based, server timesharing > system; Actually, it was designed as a single user Multics replacement to serve as a loader and emulator for already written "space war" and other games for PDP hardware, that Ken and Dennis wanted to be able to play. It evolved into a control system for phone switches when Ken had to justify the work. You really need to read a history of UNIX, and you need to read the original Bell Labs Technical Journals where early UNIX was chornicled; any good university technical library should have archive copies in the stacks. > Windows and the Mac were designed as single-user, GUI-based, desktop > systems. It should be self-evident that the latter would naturally > tend to fit into the desktop environment better than the former. It's not. The single user nature was a direct result of the lack of credentials in the predecessor OSs. It was a mistake. If you have followed the evolution of CIFS over the years, you would know that there is now the possibility of passing credential information over a single multiplex channel to a file server. Please do not confuse "single user" with "single credential". The current crop of Windows desktops are "single user" (so was NeXTStep, due to Display Postscript proxying limitations, unless you ran other atypical applications -- as with Telnet, whos daemon uses the Windows NT "Impersonate()" call in order to switch credentials, and only provides "multiuser" access because of its ability to run non-standard shells). FWIW: I uses to run DOS machines "multiuser", using a timer based TSR facility and the serial port redirection available to handle COM port based I/O, which surfaces in MS-DOS 2.11 (I did this on Leading Edge 8086 boxes). The resulting machines were "multiuser", but NOT "multicredentialed". > > This is actually a minus, since credential > > domains are a significant barrier to having > > your system "owned", aqnd contain damage when > > the worst does happen (crackers, virus, worms, > > etc.). > > The worst rarely happens, and most desktop users prefer convenience to > security. Insofar as they limit this to the desktop, there is little > reason not to indulge them. THere's really nothing inconvenient about credential enforcement, when it is done correctly. It only ever becomes obtrusive when you are being "owned", and a dialog box pops up and asks you for your password to permit the write of the DLL to the C:\windows\system directory. I run with a VXD that hooks the IFSMgr calls below the IFSMgr layer on my Windows box, precisely to let me deny such writes, which should only be necessary during software installation (which I always do offline). So even without "multiuser" or "multicredential", I get the same level of enforcement that yo state is the primary reason to not have "multiuser" or "multicredential" support in a desktop. > > Actually, you probaqbly weren't there for > > Windows through 95 ... > > I was there for all versions of Windows. Then you were well aware that Windows was not an intrinsic part of the OS, but was instead an application program that ran as a graphical user shell, capable of "fork/exec" type operations, and that you boot to DOS, not Windows, and the Windows startup has more to do with the initial command loaded being "command" or "win". > > Lindows. MacOS X. > > I'm talking about UNIX and Windows/Mac (the conventional Mac OS), not > hybrids. Of course, since once again, they defeat your binary view of the universe... 8^). > > The ability to perform "point-and-click" for > > trivial administrative tasks (such as minimal > > firewall installation, or minimal mail server > > configuration) far, far outweighs "raw power" > > in most cases. > > Point-and-click gets very tiring very quickly when you have to do a lot of > system administration, especially at a distance. Just modifying a text file > can be a lot faster and simpler. Sure. That's what scripting languages are for. Most people don't need to do that sort of thing, though, for a non-enterprise installation, and even if they do, the number of people they have to support is small enough that they can "live with the pain" of GUI administration. And since small businesses grow up to be big businesses, a GUI administration facility is a requisite bridge that make market penetration significantly easier. It's definitely not an accident that Apple is (or is going to be) the UNIX vendor with the single largest installation count of all time. Apple is all about small shop ease of use. I rather expect Apple to start selling rack-mount systems as OS/X becomes more popular... -- Terry To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-chat" in the body of the message