Linux 64Bit - Linux

This is a discussion on Linux 64Bit - Linux ; There have been several posts (mainly from the usual suspects, Hadron Quark and Tim Smith) that linux in 64bit version uses a *lot* more memory. Recently I was forced to recompile a Web-Server module (written in C++) for 32bit, as ...

+ Reply to Thread
Results 1 to 15 of 15

Thread: Linux 64Bit

  1. Linux 64Bit

    There have been several posts (mainly from the usual suspects, Hadron Quark
    and Tim Smith) that linux in 64bit version uses a *lot* more memory.

    Recently I was forced to recompile a Web-Server module (written in C++) for
    32bit, as it had to run on a non-64bit machine (which serves as a backup).
    Both versions were compiled with the exact same setup, just in one case for
    32bits.

    Today I came around to look at the memory footprint.
    Shock, horror, OMG..

    The 32bit version needs 186 KBytes
    And the 64bit version a whopping 200 KBytes.

    Immediate swapping will occur. After all, an increase of around 8% will
    break every memory calculation in no time at all

    Shocking, really
    --
    It's not about, 'Where do you want to go today?' It's more like,
    'Where am I allowed to go today?'


  2. Re: Linux 64Bit

    Peter Köhlmann writes:

    > There have been several posts (mainly from the usual suspects, Hadron Quark
    > and Tim Smith) that linux in 64bit version uses a *lot* more memory.


    Nope. We said it can use more memory. And even stated in what cases. You
    denied this. As usual you were wrong.

    >
    > Recently I was forced to recompile a Web-Server module (written in C++) for
    > 32bit, as it had to run on a non-64bit machine (which serves as a backup).
    > Both versions were compiled with the exact same setup, just in one case for
    > 32bits.
    >
    > Today I came around to look at the memory footprint.
    > Shock, horror, OMG..
    >
    > The 32bit version needs 186 KBytes
    > And the 64bit version a whopping 200 KBytes.
    >
    > Immediate swapping will occur. After all, an increase of around 8% will
    > break every memory calculation in no time at all
    >
    > Shocking, really


    It must be for you. How stupid do you feel now?

    How about the "blazingly fast" figures? Or did you not do any
    benchmarks? I mean, you seemed to think that there are some tremendous
    optimisations in C/C++ compilers over the past couple of years which
    totally dispel the benchmarks showing similarly clocked 64 bit machines
    running slower in many cases from a while ago.

    Well? Where are they?

    Poor Peter. Time to scrape your boots clean again.

  3. Re: Linux 64Bit

    Hadron wrote:

    > Peter Köhlmann writes:
    >
    >> There have been several posts (mainly from the usual suspects, Hadron
    >> Quark and Tim Smith) that linux in 64bit version uses a *lot* more
    >> memory.

    >
    > Nope. We said it can use more memory. And even stated in what cases. You
    > denied this. As usual you were wrong.


    You both said it will use *lots* more memory
    You were wrong

    >>
    >> Recently I was forced to recompile a Web-Server module (written in C++)
    >> for 32bit, as it had to run on a non-64bit machine (which serves as a
    >> backup). Both versions were compiled with the exact same setup, just in
    >> one case for 32bits.
    >>
    >> Today I came around to look at the memory footprint.
    >> Shock, horror, OMG..
    >>
    >> The 32bit version needs 186 KBytes
    >> And the 64bit version a whopping 200 KBytes.
    >>
    >> Immediate swapping will occur. After all, an increase of around 8% will
    >> break every memory calculation in no time at all
    >>
    >> Shocking, really

    >
    > It must be for you. How stupid do you feel now?
    >
    > How about the "blazingly fast" figures? Or did you not do any
    > benchmarks?


    Actually, no, I did not. And I don't intend to do so. It is pointless for
    that service. As is the increase of 14KBytes of memory footprint.
    The module runs usually on the 64bit server. Which has currently 4GBytes of
    RAM, and will get another 4GBytes shortly. You certainly can do the math
    what percentage 14KBytes will do to such a machine, can you?

    As I already pointed out, I recompiled it for a backup machine which
    hopefully will never kick in

    > I mean, you seemed to think that there are some tremendous
    > optimisations in C/C++ compilers over the past couple of years which
    > totally dispel the benchmarks showing similarly clocked 64 bit machines
    > running slower in many cases from a while ago.


    Yes. Windows machines
    Linux machines were always faster in the 64bit version

    > Well? Where are they?
    >
    > Poor Peter. Time to scrape your boots clean again.


    Idiot
    --
    Only two things are infinite,
    the Universe and Stupidity.
    And I'm not quite sure about the former.
    - Albert Einstein


  4. Re: Linux 64Bit

    On Wed, 06 Feb 2008 01:25:23 +0100, Peter Köhlmann
    wrote:


    >Linux machines were always faster in the 64bit version


    More lies from deluded Linux daydreamers.

    Check the facts:

    http://www.osnews.com/story/5768

  5. Re: Linux 64Bit

    On Feb 5, 7:25 pm, Peter Köhlmann wrote:
    > Hadron wrote:
    > > Peter Köhlmann writes:


    > >> Today I came around to look at the memory footprint.
    > >> Shock, horror, OMG..


    > >> The 32bit version needs 186 KBytes
    > >> And the 64bit version a whopping 200 KBytes.


    This was for a web server, which is mostly working with character
    strings.

    > >> Immediate swapping will occur. After all, an increase of around 8% will
    > >> break every memory calculation in no time at all


    > >> Shocking, really


    > > It must be for you. How stupid do you feel now?


    > > How about the "blazingly fast" figures? Or did you not do any
    > > benchmarks?


    Have you seen any good benchmarks comparing 64 bit and 32 bit Linux?

    I have installed OpenSuse in both 32 bit and 64 bit versions on my
    Z61p with 4 Gb of RAM and 160 Gb 7200 RPM hard drive. The 64 bit
    machine seems faster, and much more responsive, but there are times
    when some apps get disk bound and then they run about the same. The
    big problem, for me, is that my WiFi device is only supported in 32
    bit mode. So I guess I have to go back to the 1990s for a few more
    months.

    Will Atheros EVER get their act together? There STILL isn't good
    support for the a/b/g/n chipset in 64 bit mode. Of course, even in
    Windows XP and Vista, the 802.11n mode doesn't work properly. The
    connection keeps dropping and doesn't auto-recover.

    > As I already pointed out, I recompiled it for a backup machine which
    > hopefully will never kick in


    It's backing up a Linux machine - need I say more? Why not go hot-
    hot?

    > > I mean, you seemed to think that there are some tremendous
    > > optimisations in C/C++ compilers over the past couple of years which
    > > totally dispel the benchmarks showing similarly clocked 64 bit machines
    > > running slower in many cases from a while ago.


    It depends on what you are trying to do. A 64 bit machine is a mixed
    blessing. You can crunch 64 bit integers faster, but 32 bit registers
    and stack operations still take 64 bits per argument. 64 bit
    addressing means that you can access more real and virtual memory, and
    64 bit file systems means you can create files that are 16 billion
    gigabytes. With Linux/Unix sparse files, that can be a nice feature.

    > Yes. Windows machines
    > Linux machines were always faster in the 64bit version


    Have to be careful here. Complex string manipulations on 8 bit
    characters can go slower. On the other hand, barrel shifts and other
    graphics functions can be much faster in 64 bits. This is why many
    graphics cards have 128 bits or even 256 bits.

    > > Well? Where are they?

    Some benchmarks

    http://tinyurl.com/2fk7jc

    http://tinyurl.com/y69hk3

    Sorry, the results are barely readable - lousy color choices
    http://www.linux.com/feature/114024


    http://www.phoronix.com/scan.php?pag...item=616&num=3

    Pretty much, the verdict seems to be that if you don't need huge
    amounts of RAM, huge files, or huge number crunching, then you may not
    get much benefit from 64 bits.

    On the other hand, if you are thinking about hosting both Windows and
    Linux as VMs, or you want to use large virtual memory swap partitions,
    then you might still want to look at 64 bit as the host system.

    > > Poor Peter. Time to scrape your boots clean again.


    Peter isn't completely wrong. 64 bit Linux takes a modest amount of
    extra memory, and only runs slightly faster when looking at the common
    32 bit benchmarks.

    Rex Ballard
    http://www.open4success.org


  6. Re: Linux 64Bit

    * OK peremptorily fired off this memo:

    > On Wed, 06 Feb 2008 01:25:23 +0100, Peter Köhlmann
    > wrote:
    >
    >>Linux machines were always faster in the 64bit version

    >
    > More lies from deluded Linux daydreamers.
    >
    > Check the facts:
    >
    > http://www.osnews.com/story/5768


    1. He's testing Solaris 9
    2. On a SPARC (not an x86)
    3. He's also, effectively, testing gcc 3.3 as provided by
    sunfreeware.com

    --
    Men are so simple and so much inclined to obey immediate needs that a deceiver
    will never lack victims for his deceptions.
    -- Niccolo Machiavelli

  7. Re: Linux 64Bit

    Peter Köhlmann wrote:

    >There have been several posts (mainly from the usual suspects, Hadron Quark
    >and Tim Smith) that linux in 64bit version uses a *lot* more memory.
    >
    >Recently I was forced to recompile a Web-Server module (written in C++) for
    >32bit, as it had to run on a non-64bit machine (which serves as a backup).
    >Both versions were compiled with the exact same setup, just in one case for
    >32bits.
    >
    >Today I came around to look at the memory footprint.
    >Shock, horror, OMG..
    >
    >The 32bit version needs 186 KBytes
    >And the 64bit version a whopping 200 KBytes.
    >
    >Immediate swapping will occur. After all, an increase of around 8% will
    >break every memory calculation in no time at all
    >
    >Shocking, really


    Well, with the all-powerful Micro$oft Corp having so much difficulty
    fielding a decent 64-bit OS, they just HAVE to FUD the OSS
    competition...

    The "M$ Guide to Dirty Business Tactics", states "If M$ product is
    totally inferior to the competition, use FUD and 'vaporware' tactics
    to limit the damage".

    Of course, FUD and vaporware comprise their own chapters of the
    book...


  8. Re: Linux 64Bit

    Verily I say unto thee, that Linonut spake thusly:
    > * OK peremptorily fired off this memo:
    >> On Wed, 06 Feb 2008 01:25:23 +0100, Peter Köhlmann
    >> wrote:


    >>> Linux machines were always faster in the 64bit version

    >> More lies from deluded Linux daydreamers.
    >>
    >> Check the facts:
    >>
    >> http://www.osnews.com/story/5768

    >
    > 1. He's testing Solaris 9
    > 2. On a SPARC (not an x86)
    > 3. He's also, effectively, testing gcc 3.3 as provided by
    > sunfreeware.com


    >


    LOL! Looks like Hardon's passed the "stupid" virus to Otto.

    --
    K.
    http://slated.org

    ..----
    | "There would be in the license, a caveat which forbids publishing a
    | distro until it has been approved by an OSS committee." ~ Comrade
    | Hardon Quirk, communist party member.
    `----

    Fedora release 8 (Werewolf) on sky, running kernel 2.6.23.8-63.fc8
    20:10:16 up 48 days, 17:46, 5 users, load average: 0.07, 0.13, 0.09

  9. Re: Linux 64Bit

    Linonut wrote:

    >* OK peremptorily fired off this memo:
    >
    >> On Wed, 06 Feb 2008 01:25:23 +0100, Peter Köhlmann
    >> wrote:
    >>
    >>>Linux machines were always faster in the 64bit version

    >>
    >> More lies from deluded Linux daydreamers.
    >>
    >> Check the facts:
    >>
    >> http://www.osnews.com/story/5768

    >
    > 1. He's testing Solaris 9
    > 2. On a SPARC (not an x86)
    > 3. He's also, effectively, testing gcc 3.3 as provided by
    > sunfreeware.com


    Will the troll now apologize for saying that you lied?


  10. Re: Linux 64Bit

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

    On Thu, 07 Feb 2008 20:10:35 +0000,
    [H]omer wrote:
    > Verily I say unto thee, that Linonut spake thusly:
    >> * OK peremptorily fired off this memo:
    >>> On Wed, 06 Feb 2008 01:25:23 +0100, Peter Köhlmann
    >>> wrote:

    >
    >>>> Linux machines were always faster in the 64bit version
    >>> More lies from deluded Linux daydreamers.
    >>>
    >>> Check the facts:
    >>>
    >>> http://www.osnews.com/story/5768

    >>
    >> 1. He's testing Solaris 9
    >> 2. On a SPARC (not an x86)
    >> 3. He's also, effectively, testing gcc 3.3 as provided by
    >> sunfreeware.com

    >
    >>

    >
    > LOL! Looks like Hardon's passed the "stupid" virus to Otto.
    >



    don't expect the trolls to *read* what they are replying to!

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQFHq32Dd90bcYOAWPYRAhrvAJ9/mnLqcNdIzr62rjjUNyyfha8b3gCeKhPx
    GngJLBbfN0lYfPyoInZua3U=
    =sWYU
    -----END PGP SIGNATURE-----

    --
    Jim Richardson http://www.eskimo.com/~warlock
    I have an understanding with my local police--I have them outgunned, but
    they have me outnumbered.

  11. Re: Linux 64Bit

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

    On Thu, 07 Feb 2008 15:48:11 -0600,
    chrisv wrote:
    > Linonut wrote:
    >
    >>* OK peremptorily fired off this memo:
    >>
    >>> On Wed, 06 Feb 2008 01:25:23 +0100, Peter Köhlmann
    >>> wrote:
    >>>
    >>>>Linux machines were always faster in the 64bit version
    >>>
    >>> More lies from deluded Linux daydreamers.
    >>>
    >>> Check the facts:
    >>>
    >>> http://www.osnews.com/story/5768

    >>
    >> 1. He's testing Solaris 9
    >> 2. On a SPARC (not an x86)
    >> 3. He's also, effectively, testing gcc 3.3 as provided by
    >> sunfreeware.com

    >
    > Will the troll now apologize for saying that you lied?
    >



    Na, more likely, having been spanked, the Troll will slink away.

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQFHq4sJd90bcYOAWPYRAtPUAJ9De/QPjFxgmZpekzaQ5pqags/u4QCgg1Gf
    84pLkxNFllQUKLxRVSyoDrg=
    =GWO0
    -----END PGP SIGNATURE-----

    --
    Jim Richardson http://www.eskimo.com/~warlock
    "There is no reason for any individual to have a computer in their home."
    -- Ken Olson, President of DEC, World Future Society Convention, 1977

  12. Re: Linux 64Bit

    Rex Ballard wrote:

    > When the XP uptake was slow, and companies were seriously considering
    > how many of their desktops they could convert to Linux, back in 2001,
    > Microsoft did their typical "Vaporware" tactic, and announced that
    > "Longhorn" would be out "real soon".
    >
    > MS-DOS 4.0 vapor-ware promised to match the true multitasking of DR-
    > DOS.


    Proof?

    > Windows 1.0 through 3.0 promised to match the capabilities of a MAC
    > (even 3.1 wasn't as reliable).


    Proof?



    > Windows NT was vapor-ware to prevent major corporations from going
    > with Sun and X-Terminals.


    Proof?



    > Windows 95 plug-n-play was vapor-ware announced when Linux with "plug-
    > and-play" threatened NT 3.x.


    Proof?


    > Windows NT 4.0 was announced just as people started realizing that
    > Linux was more secure than 95.


    Proof?



    > Windows 2000 was vapor-wared as NT 5, when people noticed that Linux
    > was more reliable than NT 4.


    Proof?


    > Corporate customers were even offered free upgrades.
    >
    > Windows XP was vapor-wared when Linux started to take off in
    > 1998-2000, offering the power and security of NT with the low end
    > hardware of Windows 95, and Win4Lin offered Windows 95 compatibility
    > for Linux users.


    Proof?



    > When Microsoft tried to "Force Feed" XP to users who were happy with
    > Windows 2000, they started talking about Longhorn, mostly to keep
    > corporate customers from terminating or lapsing their support
    > agreements.


    Proof?




    > Windows 2008 server was vapor-wared when corporate customers started
    > switching from NT 4 to Linux instead of Windows 2003.


    Proof?



    > In each case, Microsoft saw the threat very early, quickly isolated
    > the critical hot-buttons that competitors were using to capture
    > interest and sales, and crafted a vapor-ware announcement of a new
    > versions of Windows that would have those hot new features "real soon
    > now". Usually, the new product was promised for the following year,
    > then would be announced as being delayed for another year, then
    > released in very raw and buggy beta form to a select group of people,
    > and delayed for up to another year, then finally released in beta
    > versions to MSDN and then to the general public about a year later.
    >
    > Each vapor-ware announcement, delay announcement, buggy beta, MSDN
    > beta, and GA release was carefully orchestrated to maximize profits,
    > by minimizing the number of customers who switched to the competitors,
    > while also minimizing the number of actual innovations and releases
    > that had to be rolled out.
    >
    > If you were a Microsoft stock-holder, you would have to marvel at
    > Microsoft's brilliance. Throwing just enough bones at the dogs to
    > minimize the amount of raw meat taken out of Microsoft by competitors.
    >
    > The irony is that all of the features Microsoft announced, including
    > those that were not delivered, and those that were delivered late,
    > were already available in the competitor's products at the time the
    > announcement was first made. In many cases, features introduced in
    > Windows 2000, XP, and Vista were available in Linux or UNIX as far
    > back as 1994. UNIX was doing 3D animations back when Microsoft was
    > still trying to put multitasking into MS-DOS 4.0.



    You make a lot of bogus, blanket claims about MS, but you can't provide any
    3rd-party proof of any of them. Not one. Which means it's just Ramblin'
    Rex And His Cola Bull**** Show.

    Your silly words aren't even opinions - you claim MS did this and MS sold
    that and Windows costs this and MS promised that. But almost nothing you
    say has any basis in reality.







  13. Re: Linux 64Bit

    On Feb 13, 10:08 pm, "DFS" wrote:
    > Rex Ballard wrote:
    > > When the XP uptake was slow, and companies were seriously considering
    > > how many of their desktops they could convert to Linux, back in 2001,
    > > Microsoft did their typical "Vaporware" tactic, and announced that
    > > "Longhorn" would be out "real soon".


    > > MS-DOS 4.0 vapor-ware promised to match the true multitasking of DR-
    > > DOS.

    >
    > Proof?


    It was well documented in periodicals published in the 1980s. You
    probably won't find it on Google, but you might check the
    preannouncments for MS-DOS 4.0 as well as the feature lists - probably
    around 1985-1986.

    > > Windows 1.0 through 3.0 promised to match the capabilities of a MAC
    > > (even 3.1 wasn't as reliable).


    > Proof?


    The unreliability of Windows 3.0 was well documented, again in
    publications such as e-week, information week, and other
    publications. Microsoft even added an "auto-save" feature that saved
    documents every 5-10 minutes, because crashes as frequent as every 20
    minutes were quite common and customers were complaining about loss of
    time and productivity due to these crashes.

    > > Windows NT was vapor-ware to prevent major corporations from going
    > > with Sun and X-Terminals.

    > Proof?


    In the 1991 CES, SUN was the big buzz atttraction. Microsoft was also
    offering previews of Windows 3.1, and that wasn't so interesting.
    Bill had experienced this before, when the Apple ][ eclipsed Microsoft
    Basic on the Altair back in the late 1970s. By this point, Bill had
    become quite adept at having his people attend the presentations,
    listen to the crowd, and see what features were "hot buttons". Bill
    gave the keynote speech and announced that NT would be a "Better UNIX
    than UNIX", which was a direct response to the threat of the Sun SLC
    and IPC workstations which were competitive with Microsoft on price/
    performance when a second X-terminal was used by a second employee.

    Again, all this predates Google archives, but if you go to a major
    public library in a major city like New York, Denver, Chicago, or Los
    Angeles, you should be able to access the content on Microfilm or
    Microfiche. It was covered in E-Week, PC-Week, Byte, and a few other
    weekly publications of the day. It wasn't front page news, more like
    5-10 pages in.

    > > Windows 95 plug-n-play was vapor-ware announced when Linux with "plug-
    > > and-play" threatened NT 3.x.

    > Proof?


    Historical references to Yaddragsil plug-and-play can be found quite
    easily. Red Hat also offered Plug-and-play in their "Mother's Day"
    release. Both were fully implemented and working quite effectively on
    ISA, EISA, MicroChannel, and VESA bus machines, as well as generic PCI
    machines. Microsoft announced delays in the release of Windows 95,
    much of which was to get "plug-n-play" to work correctly. Microsoft
    assigned the vendor codes and device ID codes, and insisted that these
    codes be kept protected under strict non-disclosure agreements.

    Eventually, Adaptec, who notified Microsoft of the breach of contract,
    released all of the codes it new about, to Red Hat, allowing Red Hat
    (and other Linux distributors) to configure PCI plug-and-play using
    these device codes. Eventually, the Linux user community was able to
    derive and publish a substantial catalog of Vendor and Device codes
    for both PCI and USB.

    > > Windows NT 4.0 was announced just as people started realizing that
    > > Linux was more secure than 95.

    > Proof?


    One of the big vulnerabilities of Windows 95 shares was that anybody
    could access the FAT32 directories. You could be prompted for a
    userid and password, but you didn't have to actually enter it. Early
    cable-modem and DSL internet providers often found that they could
    discover dozens of other computers "sharing" the network, allowing
    them to explore other people's PCs. Corporate hubs were also a
    problem, especially when a subordinate could explore a supervisor's PC
    and get confidential information about coworkers, or even his
    manager. It was becoming a liability problem, especially in 1996, and
    many companies looking at more secure alternatives such as NetWare and
    Linux.

    Linux on the other hand, had the same types of security mechanisms as
    UNIX, including protection and user, group, and "others" levels. In
    addition, Linux supported setuid capabilities allowing a restricted
    group of people to execute administrative scripts that could perform
    controlled actions that the user himself could not (MS-Windows still
    does not have that capability).

    Microsoft assured corporate users that NT 4.0 would have a more
    comprehensive security capability, and would still be software
    compatible with Windows and MS-DOS, which turned out to be only partly
    true. The security model was a bit weaker than Linux/Unix, but did
    provide superior security than Windows 95.

    Again, all of this was well documented in periodicals of the time,
    including publications such as E-week, Infoworld, Byte, and several
    other weekly periodicals, and again, it wasn't front-page news, more
    like pages 6-20.

    > > Windows 2000 was vapor-wared as NT 5, when people noticed that Linux
    > > was more reliable than NT 4.


    I think there are even some You-Tube clips of Bill Gates trying to
    demonstrate beta versions of Windows NT-5, getting a blue-screen of
    death, and promising that the production versions would be more
    reliable. Microsoft was also trying to tout NT 4.0 as a replacement
    for UNIX servers, and was getting hammered in the reliability
    department. The biggest problem was that process-to-process context
    switching on NT 4.0 was very slow, so developers had to depend on
    putting lots of threads into a single process to get performance.
    Applications were plagued by race conditions and deadlocks, most of
    which didn't show up until the applications were subjected to
    production level loads.

    In Windows 2000, Microsoft placed more emphasis on apartment threads,
    which was more like UNIX fork() functionality, but used more efficient
    context switching than traditional NT context switching. UNIX had
    such fast process-to-process context switching that it could run 10
    times more processes and users without having to resort to threading.
    Microsoft also placed more emphasis on the use of transaction services
    (MTS) as well as message queuing (MSMQ) to reduce the number of
    deadlocks and race conditions.

    Again, much of this was discussed in usenet newsgroups in both COLA
    and COWNA, and was also covered in a number of competitive benchmarks.

    Ironically, this was also about the same time that Linus replaced the
    single system level spin-locks with a queue system similar to the
    system used on IBM Mainframe systems such as MVS and CICS, this
    improved the performance of SMP systems significantly.

    > > Corporate customers were even offered free upgrades.


    This was also well documented in printed periodicals. Microsoft was
    getting resistance to NT 4.0 in the corporate market, and couldn't
    afford to let NT fail again (Windows NT 3.1 and 3.5 had failed
    miserably). As an incentive, Microsoft offered corporate customers
    the ability to get automatic upgrades to NT 5.0 when it was released,
    if they ordered NT 4.0 in 1997. By 1998, Microsoft was adopting the
    Linux revenue model, offering corporate customers service contracts on
    Microsoft products ranging from Windows to Office to Project and
    Visio. These were similar to MSDN, but covered all updates and
    patches for all employees. One of the promises was that subscribers
    to these service plans would automatically get upgraded to NT 5.0 at
    no additional cost.

    When Windows 2000 was released, Microsoft "shipped" license key
    authorizations to corporate customers for Windows 2000, along with an
    image that could be installed from a corporate server. Ironically,
    this was very similar in approach to the distribution tools used by
    most Linux distributors such as Red Hat, Turbo-Linux, Caldera, and
    SUSE at the time.

    > > Windows XP was vapor-wared when Linux started to take off in
    > > 1998-2000, offering the power and security of NT with the low end
    > > hardware of Windows 95, and Win4Lin offered Windows 95 compatibility
    > > for Linux users.

    >
    > Proof?


    Microsoft announced Windows 98, and one of the most significant
    introductions was USB. Most of the older Windows 95 machines were now
    obsolete, and as a result, there was suddenly a glut of old Windows 95
    machines that were still completely functional. End-users were
    converting these older machines to Linux, and interest in Linux began
    to surge. By late 1999, when Red Hat went public, there was a big
    surge in interest, and soon Corel was offering it's version of Linux
    as well.

    Microsoft released Windows ME and Windows 2000 at about the same time,
    within a few months of each other, and Windows ME was a disaster. The
    PC market was in a slump, the Y2K panic was over, and users who tried
    ME, or purchased new machines with Windows ME hated it so much that
    they wanted to "down-grade" back to Windows 98. Interest in Linux was
    also starting to surge again, partly because Linux had cracked
    Microsoft's USB device and vendor codes, and partly because IBM was
    now a very visible supporter of Linux. In addition, HP and Dell were
    producing more "Linux Ready" machines, most of which were "top-of-the-
    line" machines, and they were finding that corporate customers and
    professionals were quite willing to pay a little extra for Linux
    compatibility. Even though the machines were sold with Windows ME or
    Windows 2000, the end-users were often creating dual-boot machines.

    Win4Lin, and other "Windows Emulators", most based on enhanced
    versions of WINE, increased the capability to run Windows applications
    under Linux. By the time XP was released, most Windows 98
    applications could be run under these enhanced emulators.

    There was lots of coverage, especially in publications like E-Week and
    ComputerWorld, as well as other publications with a "full IT spectrum"
    focus. By this time, byte was now just a web site, but they also
    provided coverage. ComputerWorld did a survey and found that 85% of
    the corporate IT respondents were exploring alternatives to Windows on
    the desktop.



    > > When Microsoft tried to "Force Feed" XP to users who were happy with
    > > Windows 2000, they started talking about Longhorn, mostly to keep
    > > corporate customers from terminating or lapsing their support
    > > agreements.

    > Proof?



    In 2003, many Microsoft service contracts were about to expire. Many
    IT departments had noticed that Microsoft offered automatic update
    services for XP as part of it's standard offering, and that many of
    the patches being pushed to the corporations had to be blocked because
    they kept breaking 3rd party applications. IBM had to block XP
    service pack 2 for almost 18 months because it broke key strategic
    applications such as Lotus Notes, Lotus SmartSuite, SameTime, and some
    corporate tools used for booking travel, tracking time, and filing
    expense reports. I don't know exactly which applications were broken
    or when they were patched or upgraded, but it took about 18 months to
    get it all sorted out.

    I suppose that if someone had CD-ROM archives or DVD-Archives of their
    Lotus Notes databases, they might be able to find the references.
    Several companies did make public statements to this effect. Several
    other companies did announce publicly that they had decided to opt out
    of the "platnum" support contracts, especially when Microsoft wanted
    to substantially increase the subscription rate (after increasing it
    for Windows XP).

    Again, much of this was covered in print periodicals of the day.
    Unfortunately, most of these publishers don't maintain long-term
    archives of their entire content. They might have a few "choice"
    articles available, but it's not very easy to search or browse a
    specific back-issue even 5 years back. Some Ziff-Davis publications
    purge their archives after 90 days.

    > > Windows 2008 server was vapor-wared when corporate customers started
    > > switching from NT 4 to Linux instead of Windows 2003.

    > Proof?


    Microsoft tried to "expire" Windows NT 4.0 several times, starting as
    early as 2001. They were originally only going to give corporate
    customers a year to make the transition to Windows 2000. This
    expiration kept being extended out and eventually, Microsoft did cut
    support completely. Shortly after this, Microsoft started announcing
    new features of the "Longhorn Server". They promised that several
    features, which had been cut from "Longhorn Workstation" would be
    implemented in the server version.

    About this same period of time, there was a huge surge in Windows to
    Linux server transitions. Many of the services used on these Windows
    servers were either databases such as DB2 and Oracle, which could
    easily be moved from Windows to Linux, or they were Java 2
    applications, which could also be easily moved from Windows to Linux.
    Oracle, IBM, and other major vendors had made Linux a primary target
    platform, often releasing the Linux versions at the same time as the
    Windows and UNIX versions.

    In addition, AIX, Solaris, and HP_UX had improved the ability to
    partition large servers into several smaller "virtual servers". By
    using these LPARs, a single AIX server could do the work of 20-30 4-
    processor Windows NT servers, in a much smaller footprint, at much
    lower cost in terms of power and cooling, and at a much lower
    maintenance cost.

    > > In each case, Microsoft saw the threat very early, quickly isolated
    > > the critical hot-buttons that competitors were using to capture
    > > interest and sales, and crafted a vapor-ware announcement of a new
    > > versions of Windows that would have those hot new features "real soon
    > > now". Usually, the new product was promised for the following year,
    > > then would be announced as being delayed for another year, then
    > > released in very raw and buggy beta form to a select group of people,
    > > and delayed for up to another year, then finally released in beta
    > > versions to MSDN and then to the general public about a year later.

    >
    > > Each vapor-ware announcement, delay announcement, buggy beta, MSDN
    > > beta, and GA release was carefully orchestrated to maximize profits,
    > > by minimizing the number of customers who switched to the competitors,
    > > while also minimizing the number of actual innovations and releases
    > > that had to be rolled out.

    >
    > > If you were a Microsoft stock-holder, you would have to marvel at
    > > Microsoft's brilliance. Throwing just enough bones at the dogs to
    > > minimize the amount of raw meat taken out of Microsoft by competitors.

    >
    > > The irony is that all of the features Microsoft announced, including
    > > those that were not delivered, and those that were delivered late,
    > > were already available in the competitor's products at the time the
    > > announcement was first made. In many cases, features introduced in
    > > Windows 2000, XP, and Vista were available in Linux or UNIX as far
    > > back as 1994. UNIX was doing 3D animations back when Microsoft was
    > > still trying to put multitasking into MS-DOS 4.0.


    > You make a lot of bogus, blanket claims about MS, but you can't provide any
    > 3rd-party proof of any of them. Not one. Which means it's just Ramblin'
    > Rex And His Cola Bull**** Show.


    Maybe an archive guru like Roy Schestowitz can dig up the references
    for you.

    I don't have time to go digging through 3 million google hits trying
    to find an old ziff-davis article that I read in a print version 20
    years ago, or 10 years ago, or even 5 years ago. If I kept all the
    hard-copy periodicals I've read in my 30 year career, I would have to
    have a warehouse to store them all. At one time, I had over 2,000
    cubic feet of old periodicals, manuals, and old hardware parts,
    including every copy of Byte from 1979 to 1993, but I finally had to
    let all that go after paying $100/month for 12 years to store it all
    in a storage garage.

    > Your silly words aren't even opinions - you claim MS did this and MS sold
    > that and Windows costs this and MS promised that. But almost nothing you
    > say has any basis in reality.


    I have a peculiar knack for reading entire periodicals, during lunch
    breaks, during waiting periods, and while flying on airplanes. I tend
    to remember the really interesting articles in these trade
    publicatons, especially when they are related to Unix, Linux, and
    Microsoft.

    Most people only get one perspective. They might have a vague
    recollection of activities done by Microsoft during a particular
    period of time, but they had no interest in Unix or Linux at the time
    (many were only in high school at the time and only had access to
    Windows or MS-DOS), therefore they have no context of the broader
    spectrum of technology and how the UNIX community was attempting to
    compete with proprietary products like MS-DOS, VMS, MVS, VM/CMS,
    Windows, and Mac.

    Show me a good RELIABLE way to search these ancient archives for the
    original content in their entirety, and offer you a link to that
    publication, so that I don't have to violate copyrights, and I'd be
    willing to try doing some digging.


  14. Re: Linux 64Bit

    "Rex Ballard" schreef in bericht
    news:3efe06ad-30f5-4e05-a2c2-d865b483990f@i7g2000prf.googlegroups.com...
    > On Feb 13, 10:08 pm, "DFS" wrote:
    >> Rex Ballard wrote:
    >> > When the XP uptake was slow, and companies were seriously considering
    >> > how many of their desktops they could convert to Linux, back in 2001,
    >> > Microsoft did their typical "Vaporware" tactic, and announced that
    >> > "Longhorn" would be out "real soon".

    >
    >> > MS-DOS 4.0 vapor-ware promised to match the true multitasking of DR-
    >> > DOS.

    >>
    >> Proof?

    >
    > It was well documented in periodicals published in the 1980s.


    ZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZ










  15. Re: Linux 64Bit

    On Feb 18, 1:22 pm, "Hans" wrote:
    > "Rex Ballard" schreef in berichtnews:3efe06ad-30f5-4e05-a2c2-d865b483990f@i7g2000prf.googlegroups.com...
    >
    > > On Feb 13, 10:08 pm, "DFS" wrote:
    > >> Rex Ballard wrote:
    > >> > When the XP uptake was slow, and companies were seriously considering
    > >> > how many of their desktops they could convert to Linux, back in 2001,
    > >> > Microsoft did their typical "Vaporware" tactic, and announced that
    > >> > "Longhorn" would be out "real soon".

    >
    > >> > MS-DOS 4.0 vapor-ware promised to match the true multitasking of DR-
    > >> > DOS.

    >
    > >> Proof?

    >
    > > It was well documented in periodicals published in the 1980s.

    >
    > ZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZZ


    Agreed.
    Really is boring trying to *prove* something that actually happened
    10-20 years ago, when it wasn't on the headlines of every major paper,
    the headlines of every major technical publication. When it's a page
    12 story, guys like DFS say "It never happened".

    It's like Neo-Nazis who insist the holocaust never happened because
    there were no headlines of it in the newspapers of Nazi German
    publications.

+ Reply to Thread