Does Microsoft lie about the Linux features?

Discussion in 'Linux Networking' started by totojepast, Sep 12, 2003.

  1. -----BEGIN PGP SIGNED MESSAGE-----
    Hash: SHA1

    How does root go about clearing out the cruft then? Like cronjobs
    cleaning up /tmp?

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.2.3 (GNU/Linux)

    iD8DBQE/bQofd90bcYOAWPYRArjtAJ9SL3RNO8BiI+EdCzHt+nocBVG0QwCgiHj6
    qWQOs+tEdTrLMRZJ/iN5rqk=
    =WoHY
    -----END PGP SIGNATURE-----
     
    Jim Richardson, Sep 21, 2003
    1. Advertisements

  2. One way would be for root to take ownership. Another would be to run the
    job as an account with "backup" priviledges.
     
    Erik Funkenbusch, Sep 21, 2003
    1. Advertisements

  3. -----BEGIN PGP SIGNED MESSAGE-----
    Hash: SHA1


    If no delete privs exist, how would the backup account clean out the old
    cruft?

    Using the root method, you'd have to take ownership, and then delete the
    files? seems kinda clunky. Can't you just give delete perms to root
    and the owner only? after all, that's what you get with the sticky bit,
    automatically.

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.2.3 (GNU/Linux)

    iD8DBQE/bU/Id90bcYOAWPYRAiYWAKDaHeA40HBW2ESZTxKAxfa2S8gDqgCg29Xl
    6ZOe/WK1nTg4MzmRgFeFL3c=
    =pDJ+
    -----END PGP SIGNATURE-----
     
    Jim Richardson, Sep 21, 2003
  4. The backup priviledge must have the right to read and write (and delete and
    create) files it doesn't own or otherwise have permission to access,
    otherwise backups wouldn't work.
    Deny rights take precedence over Allow rights. If you deny everyone, then
    that includes administrators unless they take ownership, even if you have a
    specific "allow" delete permission.

    You'd have to Deny some group that Administrators don't belong to in order
    to accomplish this.
     
    Erik Funkenbusch, Sep 21, 2003
  5. totojepast

    James Knott Guest

    When Gary Killdall wrote CP/M, there were very few options other than the 8
    bit CPUs, such as 8080, Z80, 6800, 6502 etc., for personal computer use.
    There was no built in support for multitasking etc. In that context,
    CP/M-80 was quite something. CP/M-86 was designed to take advantage of the
    8086 hardware and would have been more capable than CP/M-80. DOS however,
    was stuck on the CP/M-80 model. Given that DOS was originally meant to
    develop hardware, it lacked a lot of features of a real OS. DOS 1 looked a
    lot like CP/M, because that's what Tim Paterson at SCP was trying to
    emulate, so that the hardware would be ready, when CP/M-86 came out.
    Thanks to Bill Gates, we got stuck with a hardware development system,
    rather than a proper OS.

    --

    Fundamentalism is fundamentally wrong.

    To reply to this message, replace everything to the left of "@" with
    james.knott.
     
    James Knott, Sep 21, 2003
  6. totojepast

    James Knott Guest

    FWIW, there's an interesting article in the June 1983 issue of Byte, "An
    Inside Look at MS-DOS", which discusses a lot of the short comings in the
    package sold by MS. (remember Edlin?). It shows a lot of the design faults
    that were there, simply because it wasn't originally intended as a
    commercial package. There are also some interesting articles on CP/M, MP/M
    and Xenix in the June 1981 issue.

    BTW, I have *EVERY* print issue of Byte Magazine on my shelves, going back
    to Vol 1 #1, Sept 1975. I bought the first 3 issues, in person from the
    original publisher (Wayne Green of 73 Magazine), at an amateur radio
    convention in Ottawa, back in Oct 1975. It's interesting to browse through
    them and recall the events of past years. I recently took the 1st issue
    into work, and mentioned to one of my co-workers, that the magazine was
    older than she was!!!

    --

    Fundamentalism is fundamentally wrong.

    To reply to this message, replace everything to the left of "@" with
    james.knott.
     
    James Knott, Sep 21, 2003
  7. -----BEGIN PGP SIGNED MESSAGE-----
    Hash: SHA1

    ["Followup-To:" header set to comp.os.linux.advocacy.]
    and it can do this in a directory in which there *are* no delete privs?
    So what you are saying, is that the original point is correct, that NTFS's ACLs
    can't accomplish this as easily as the *nixoid ugo?

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.2.3 (GNU/Linux)

    iD8DBQE/bcHQd90bcYOAWPYRAneBAKCLH3NSGX5ARwgw0eoyTW32vjrmdgCfafgO
    GRO2oczj0yxaRz2E+N/Y+4I=
    =BMVY
    -----END PGP SIGNATURE-----
     
    Jim Richardson, Sep 21, 2003
  8. totojepast

    DM Guest

    MS's modus operandi consists of distorting the truth, cajoling and
    otherwise arm-twisting in order to suit their ambitions.
     
    DM, Sep 21, 2003
  9. A few points to put this into perspective though...

    In the 1970's, and certainly in 1980, Intel was designing
    "device controllers", not "computer" chips. That was a
    purposeful decision, and caused a group of Intel engineers to
    break off and form Zilog to design computer cpu's. (The hoot
    is, Zilog's biggest successes were device controllers, and
    Intel went on to rule the world with computer cpu's.)

    Hence Intel came out with the "16 bit" 8086 in June 1978, and
    one year later followed it with the 8088, an "8 bit" version
    that was more compatible with the 8085 and 8080. And *none* of
    them were designed to be used in computers! But IBM saw
    software continuity as a valid direction, and used the 8088
    rather than the 8086, in their new PC design. The support
    hardware for an 8088 was less expensive, but the significant
    point was that 8085/8080 software could literally be translated,
    rather than re-written, to run on the 8088.

    So one year later, in early 1980, finds Tim Patterson, working
    as Seattle Computer Products (SCP) in the business translating
    8080 code to 8088 code with a set of software translation tools
    he had written. It _did_ work well, just as Intel had said! In
    fact, it would even translate the source code for the CP/M
    operation system... which is _exactly_ what Patterson did to
    produce QDOS and 86-DOS. He sold 86-DOS to Bill Gates, who
    transformed it into MS-DOS/PC-DOS for IBM. (All of it in violation
    of Gary Kildall's copyright.)

    Intel had decided they could beat out Motorola by guaranteeing
    software continuity from one chip to the next, even if it wasn't
    the most powerful cpu. IBM chose the Intel chips because of
    that. Microsoft provided an OS *now* in an effort to beat out
    Digital Research, even if it wasn't technically the best OS.
    IBM bought the Microsoft product because of that.

    But even if they weren't then, competition since has forced two
    of those three companies to become *very* concerned about
    producing the _best_ product... And it appears that the Open
    Source Software movement might just Microsoft into the fold soon
    enough too.
     
    Floyd Davidson, Sep 21, 2003
  10. totojepast

    Les Mikesell Guest

    I guess that explains a lot about the Intel instruction set. The part
    I don't get is why they still don't make chips that can virtualize
    themselves without the undocumented magic of VMware. Don't
    they understand yet that the chips are being used in computers and
    the operating system should be able to virtualize everything?
    I think you are mistaken about the relative importance of these
    issues in the mass market. Motorola continued to price the
    68xxx series for the minicomputer market long after Intel had
    taken over the desktop with microcomputers built from the
    cheap controller chipset. Just like cheap windows software
    easily beat $1,000/box unix on the desktop, cheap chipsets
    took the volume market in spite of the reduced functionality.
     
    Les Mikesell, Sep 21, 2003
  11. But IBM and Intel didn't look at it that way in 1980, which was
    what I meant that to say. They thought the compatibility issue
    was extremely significant at the time, and both were publishing
    advertisements and "canned" reviews (in Byte magazine, for
    example, where it was not totally obvious that the "review" was
    written by the company producing the product) that made a big
    deal out of Intel's promise of compatible cpu's for several
    generations, and IBM's promise that the IBM-PC would continue to
    use those compatible cpu's.

    Personally, at the time, I thought they were nuts in the head.
    Given the way it worked out, I'm not so sure now...
     
    Floyd Davidson, Sep 21, 2003
  12. totojepast

    James Knott Guest

    The 8086 & 8088 use the same software, so upgrading from CP/M would be
    similar for both. However, as you say, the hardware support was cheaper
    for the 8088, because of the 8 bit I/O bus.
    Anyone else recall the P-codo, that was also supposed to be a PC OS?


    --

    Fundamentalism is fundamentally wrong.

    To reply to this message, replace everything to the left of "@" with
    james.knott.
     
    James Knott, Sep 21, 2003
  13. totojepast

    Les Mikesell Guest

    Personally, I've always thought that up until the late 90's IBM
    did not want cheap desktop PC's (even their own) to be powerful
    enough to perform any jobs that could take away their mainframe
    business. As long as it could run a 3270 emulator, IBM probably
    wanted the least useful CPU they could find on the desktop. Note
    that IBM also manufactured 68xxx based machines at the time,
    but designed and priced for the scientific market.
     
    Les Mikesell, Sep 21, 2003
  14. totojepast

    Lew Pitcher Guest

    I didn't say "Windows ACLs", I just said "ACLs". <grin>

    I don't know much about "Windows ACLs", so I don't have a "stickybit" workalike.


    --
    Lew Pitcher

    Master Codewright and JOAT-in-training
    Registered Linux User #112576 (http://counter.li.org/)
    Slackware - Because I know what I'm doing.
     
    Lew Pitcher, Sep 22, 2003
  15. -----BEGIN PGP SIGNED MESSAGE-----
    Hash: SHA1

    That's OK, Erik allready pointed out how hard it was to come close to
    this functionality with NTFS ACLs... :)

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.2.3 (GNU/Linux)

    iD8DBQE/bpEvd90bcYOAWPYRAiKyAKCfaNmfj3AlTS7NFcdcT3e/qk2ghwCghGZw
    FUwWsSvy6wlo7879d7DG9us=
    =8RCT
    -----END PGP SIGNATURE-----
     
    Jim Richardson, Sep 22, 2003
  16. ["Followup-To:" header set to comp.os.linux.hardware.]

    Umm... buy another hard drive? :)
     
    John Thompson, Sep 23, 2003
  17. totojepast

    wild bill Guest

    The difference is PROTECTED MODE. What happened
    that nobody, ms, ibm, figured on was that gates & co would
    decipher making things run in protected mode during the
    re-coding from 286 to 386. By leaving ibm running a lot
    of real mode code in OS/2, they knew protected mode
    windows would blow it away, performance wise. Once they
    got the bugs down to where it would run at all.
    Gates has figured he's smarter than ANY lawyer virtually
    all his life. Read the contracts he made his family sign when
    he was a kid. The IBM lawyers were an overconfident bunch
    of fools for him to work on. And he outsmarted them. Going
    into any contest where you don't respect your enemy makes
    you a fool. Use ms products, or avoid them, But don't take on
    microsoft unless you're prepared for the fight of your life.

    That's what makes Linux so interesting. It's like some sort
    of corrosive stuff on the floor ms is standing on. Threatening
    to dissolve the evil empire from the ground up. ''That's the
    stupidist fucking thing I've ever heard of''. But..... true!

    And then there's Torvalds, as if nothing bothered him at all.
    What's a bully like gates supposed to do to get his attention?
    Gates is becoming irrelevant, and I don't think he likes it.


    Bill
     
    wild bill, Sep 24, 2003
  18. totojepast

    Olav Dahlum Guest

    Are you nuts? if there's one thing I have to say about M$ filsystems it
    must be this: that's not true!!!
    All their filesystems, even the so-called high-end NTFS sucks!!!
    Haven't you noticed that it self-destructs from time to time or quite often?

    I had an really messed up 486 DXL4 100 Mhz that was running vanilla
    RedHat 6.0 and it worked really well. Treated that installation like
    shit and the filesystem (ext2) lived happily trough my massive abuse.
     
    Olav Dahlum, Nov 15, 2003
  19. totojepast

    Olav Dahlum Guest

    I like you. Ever heard of Skolelinux vs. the M$ FUD machine?

    Olav
     
    Olav Dahlum, Nov 15, 2003
  20. totojepast

    Leon. Guest

    Most Microsoft "features" arent used except by nobody, and so they dont
    exist either.
     
    Leon., Nov 23, 2003
    1. Advertisements

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.