Article View: comp.windows.x
Article #23782Re: Mac System 7.0 vs. Unix
From: Douglas S. Rand
Date: Mon, 17 Jun 1991 16:30
Date: Mon, 17 Jun 1991 16:30
127 lines
6914 bytes
6914 bytes
In article <7871@segue.segue.com> gene@segue.segue.com (Gene Hightower) writes: This should go someplace other then comp.windows.x. I agree but there is some applicability. In article <1068@witch.witchcraft.sf.ca.us> chappell@witchcraft.sf.ca.us (Tom Chappell) writes: >larryc@puente.jpl.nasa.gov (Larry Carroll) writes: >>This is a kind of silly thing to ask, but my supervisor doesn't think so: >>How does Apple's System 7.0 operating system compare with Unix? ^^^^^^^^^^^^^^^^^ So lets compare it with Unix. First let me say that by "Unix" I am talking about SunOS 4.x and HP-UX from first hand knowlage. I can't speak for all the systems that are (or have ever been) called Unix. I've used BSD, SUN and SYSV, I've also used Macs for a long time. >Man, oh man, I can't believe you just said that. I won't bother with a full >list, but apart from the "cosmetic" changes, System 7 includes: > >1. Peer-To-Peer networking via AppleTalk (LocalTalk, TokenTalk, or EtherTalk) >2. Inter-Application Communication, INCLUDING IAC BETWEEN APPLICATIONS RUNNING Unix systems have networking. NFS provides file sharing, but there are lots of ways that Unix systems use the ability to network. With a window system such as X, programs can run on one machine and display graphics on the screen of another machine. The X window system is not unique to Unix, but the Mac OS makes it hard to build more then the server side of the system. In other words, to use Macintish A you have to sit in front of Mac A and use it's keyboard and screen. You can't access its CPU power from another computer. Not true. An application called Timbuktu does exactly this. Also X is not UNIX is not X. X is a protocol, X runs on VMS and even my old OS PRIMOS. It isn't clear to me why running X clients on a Mac would make any sense anyway. The major use for X is as a server to allow a user to run applications which require special resources on another machine. The MAC is hardly a compute resource. Apple's file sharing is much kinder then NFS. NFS has lots of flaws and if you want a list I'll send it. Apple doesn't require the user to know what's going on. Most UNIX networking is not very sophisticated and takes place right at the TCP or UDP layers, and don't tell me about OSI since I know it's a commie plot ( large grin here ). >3. Support for File, Folder, and Volume Aliases, similar to UNIX links, except The same as "symbolic" links. Symbolic links can point to files on other filesystems and via NFS to files on remote machines. Not the same. Apple's aliases are just like symbolic links for files on the same machine, but for files on a different machine they store all the information to find that specific file. UNIX requires much to have this work (for example, setting up the automounter on SUN OS). >4. Publish and Subscribe. One Macintosh user selects a portion of a document, This feature is mostly a feature of the application you are running, i.e. the drawing or word processing program. System 7 may provide support to such operations, but it does not provide the operations directly. This is a non-trivial support job. I have yet to see any Unix applications supporting live links. It really is helpful for a hardware vendor to specify how such things should work. It binds the different software suppliers to one mind of how to do the work. Apple has been the most successful example of this to date. There really isn't anything in the Mac that requires all Mac applications to look the same. They could choose not to have, for example, the menu bar. But Inside Mac says you shall do it and I don't remember the last application I ran, if ever, which didn't. Unix applications have had the OS support (in the form of IPC) for some time. Right. Just go look at the work done on even the current ICCCM work in X11R4 and tell me that this approaches the ease of use of the Mac clipboard. On the Mac I can take and cut something from app A and paste into app B and usually not loose very much. At worst the cut becomes a PICT resource. IPC on UNIX is very non-standard. You have Berkeley (SUN) RPC on quite a few machines. You have NCS on others. The two don't interoperate. The only absolute guarantee on all workstations is TCP/IP and even that might not exist on some older Apollo boxes (with Domain) or (not up to date on this one) some VAX stations with DECNET. >5. Virtual Memory. Allows Macintosh users to actually run several applications Well, almost virtual memory. All of the programs run in the same address space. The programs don't move around. If you run programs A, B and C and program B needs more space, you are out of luck. You could give each program a huge block of address space, but it's not as good as each program having its own space. It has VM. It doesn't yet have some other memory management features and memory protection isn't yet turned on. The inability to move programs around suprises me though. You sure? Another problem with this way of doing things is that one program can screw up not only itself, but any other program on the system, or the OS itself. Not a necessary limitation on a MAC with a PMMU or a 68030 or better. But Apple hasn't turned it on yet. System 7 also swaps to the filesystem. This has drawbacks. Speed mostly. OS/2 does this also. The big reason is that folks can upgrade to the new virtual memory OS without rearanging hard disk space. This is a red herring. The SUN OS does this also. You can definitely use the filesystem for swapping without incurring (serious) performance problems. After all, the OS can superoptimize the swapping file, for example, make all the blocks contiguous as Sun does. Oh well. That's enough comments for me. Someone else can comment on the rest. -- Douglas S. Rand Internet: <dsrand@mitre.org> Snail: MITRE, Burlington Road, Bedford, MA Disclaimer: MITRE might agree with me - then again... Amateur Radio: KC1KJ
Technical Details
Message-ID:
<DSR.91Jun17123035@mir.mitre.org>Path:
arc1-ca.pugleaf.net!utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!thunder.mcrcim.mcgill.edu!snorkelwacker.mit.edu!linus!linus!linus!dsrReferences:
<1991Jun7.201220.19510@jpl-devvax.jpl.nasa.gov> <1068@witch.witchcraft.sf.ca.us> <7871@segue.segue.com>Show Full Headers
Complete Article Headers
Path: arc1-ca.pugleaf.net!utzoo!utgpu!news-server.csri.toronto.edu!bonnie.concordia.ca!thunder.mcrcim.mcgill.edu!snorkelwacker.mit.edu!linus!linus!linus!dsr From: dsr@mitre.org (Douglas S. Rand) Newsgroups: comp.windows.x Subject: Re: Mac System 7.0 vs. Unix Message-ID: <DSR.91Jun17123035@mir.mitre.org> Date: 17 Jun 91 16:30:35 GMT References: <1991Jun7.201220.19510@jpl-devvax.jpl.nasa.gov> <1068@witch.witchcraft.sf.ca.us> <7871@segue.segue.com> Sender: news@linus.mitre.org (News Service) Organization: The MITRE Corporation, Bedford, MA Lines: 127 In-Reply-To: gene@segue.segue.com's message of 14 Jun 91 23:52:44 GMT Nntp-Posting-Host: mir.mitre.org