IRIX freeware releases - SGI

This is a discussion on IRIX freeware releases - SGI ; In article , Ivan Rayner wrote: ...... >I really don't know how many customers would go from a PC to an Altix. >It doesn't seem like a logical progression, however if a customer did >have an SGI Intel workstation and ...

+ Reply to Thread
Page 2 of 2 FirstFirst 1 2
Results 21 to 34 of 34

Thread: IRIX freeware releases

  1. Re: IRIX freeware releases

    In article ,
    Ivan Rayner wrote:
    ......

    >I really don't know how many customers would go from a PC to an Altix.
    >It doesn't seem like a logical progression, however if a customer did
    >have an SGI Intel workstation and was happy with it, that would only
    >help our chances of selling them a high end system if they required
    >one.


    We got an Altix *despite* all the PC stuff preceeding it.
    It is attractive as being one of the few (or only) single system
    image implementations of the Itanium II. One big question is
    the developer environment. Having Intel devoting lots of resources
    to compilers makes things more attractive, but losing the irix
    compilers and tools is not a plus.

    >Still, I expect we'll sell Altixes to customers that want to upgrade
    >their old Origins, and others that are simply impressed by its
    >performance. 'Course I'm just a lowly engineer, so I know nothing
    >about sales.


    Kickbacks. Think kickbacks. :-)

    --
    Daniel Packman
    NCAR/ACD
    pack@ucar.edu

  2. Re: IRIX freeware releases

    > Still, I expect we'll sell Altixes to customers that want to upgrade
    > their old Origins, and others that are simply impressed by its
    > performance. 'Course I'm just a lowly engineer, so I know nothing
    > about sales.


    for future reference hardware or software?

    >
    >
    > Ivan
    > --
    > Ivan Rayner
    > ivanr@sgi.com



  3. Re: IRIX freeware releases

    SkyWriter wrote:
    >> Still, I expect we'll sell Altixes to customers that want to upgrade
    >> their old Origins, and others that are simply impressed by its
    >> performance. 'Course I'm just a lowly engineer, so I know nothing
    >> about sales.

    >
    > for future reference hardware or software?


    I'm a software engineer, therefore I know nothing about the following:
    - sales
    - marketing
    - hardware
    - software

    Ivan
    --
    Ivan Rayner
    ivanr@sgi.com

  4. Re: IRIX freeware releases

    Ivan Rayner wrote:

    > SkyWriter wrote:
    > >> Still, I expect we'll sell Altixes to customers that want to upgrade
    > >> their old Origins, and others that are simply impressed by its
    > >> performance. 'Course I'm just a lowly engineer, so I know nothing
    > >> about sales.

    > >
    > > for future reference hardware or software?

    >
    > I'm a software engineer, therefore I know nothing about the following:
    > - sales
    > - marketing
    > - hardware
    > - software


    ha! sounds like a hardwares guys description of a software guy, but let it
    ride!

    >
    >
    > Ivan
    > --
    > Ivan Rayner
    > ivanr@sgi.com



  5. Re: IRIX freeware releases

    "Benjamin Gawert" writes:

    > Walter Roberson wrote:
    >
    > > Software application vendors, though, made it very clear to SGI that
    > > every time SGI does a major release upgrade (e.g., 6.5 to 6.6) then
    > > that they have to requalify all their applications for IRIX, and that
    > > that's an expensive process that a number of them are not willing to
    > > undertake. These vendors -are- willing to keep up with the subrelease
    > > 6.5.X strategy because of the compatability guarantees. Break that
    > > guarantee, upset the vendors.

    >
    > IMHO that's a really sad excuse for stagnation.


    What is stagnating?

    > The point is that most other platform vendors _do_ major release
    > upgrades with the software vendors following them. I understand SGI
    > is afraid of loosing the few remaining companies developing for IRIX
    > but besides the technical side it's very bad marketing since it
    > makes people think SGI isn't interested in IRIX any more.


    So they should leave a technically sound, working solution in favor of
    unsubstantiated PR? That's a great way to piss users off.

    > > Even if it's just a name change for sales hype reasons, the
    > > vendors have to treat it as if it is a big technical change, and
    > > that makes them unhappy.

    >
    > It depends. I have the feeling it makes _SGI_ unhappy as every new release
    > of a vendors software product might be the one that doesn't support IRIX any
    > more. If software vendors are running away from Your platform of course You
    > are afraid of changes that might cause these vendors to think about their
    > IRIX efforts twice.


    As well they should be. And not just vendors. As a customer, I value
    the backwards compatibilty very much.

    > With the current situation SGI could really use some "marketing signal" that
    > shows that there is still done something. Maybe IRIX is still up to date,
    > but the outside world gets the impression that it is not. Usually the last
    > numbers are used to show the patch level, and a numbering like 6.5.22 says
    > that it's mainly a bug fix for an operating system that hasn't seen a major
    > update since 1998! Maybe there is any progress inside IRIX, but that is
    > definitely not shown by SGI. And while the competition is producing a new OS
    > Release every 1-3 years SGI will still be at 6.5 when offering the 'IRIX
    > 6.55.485m' update.


    And THAT IS A GOOD THING!! The failure here is not the numbering
    scheme but arguably one of marketing. You make a technical decision, it is
    then up to marketing to convince the customers of the benefits. When you
    change technical decisions on non-technical grounds, that's when you
    fail miserably.

    > For most platform manufacturers a new OS Release usually is a positive
    > thing, providing some marketing hype and showing the customers that their
    > platform is still alive and being improved.


    And a bad thing for customers, who end up being forced to change
    applications and methods each time.

    > IMHO SGI shouldn't be too afraid of changes...


    IMO people should argue more about what should be improved in Irix and
    less about what number is attached to it.

    *p

  6. Re: IRIX freeware releases

    SkyWriter wrote:
    >
    > thus the death knell of disruptive technology doth ring...
    >


    Perchance he for whom this bell tolls may be so ill, as that he
    knows not it tolls for him...

    --John Donne



  7. Re: IRIX freeware releases

    In article ,
    Ivan Rayner wrote:

    : I'm a software engineer, therefore I know nothing about the following:
    [...]
    : - software

    Uh oh, I think SGI is in trouble now.


    Cheers - Tony 'Nicoya' Mantler

    --
    Tony 'Nicoya' Mantler -- Master of Code-fu -- nicoya@ubb.ca
    -- http://nicoya.feline.pp.se/ -- http://www.ubb.ca/ --

  8. Re: IRIX freeware releases

    Per Ekman wrote:

    >
    > IMO people should argue more about what should be improved in Irix and
    > less about what number is attached to it.
    >

    You get my vote for the post that should end this thread.


    --
    Alexis Cousein Senior Systems Engineer
    alexis@sgi.com SGI/Silicon Graphics Brussels




  9. Re: IRIX freeware releases

    Ivan Rayner wrote:

    >> Well, IA64 is definitely not what one would understand as "comodity
    >> parts".

    >
    > That depends on your point of view. A commodity part for SGI, is
    > something that is developed by someone else that is either cheap and
    > does its job well, or has other specific advantages over SGI
    > developing
    > the part internally.


    That probably applies to 99% of all SGI hardware. Most chips inside SGI
    machines are not made by SGI but from intel, LSI or whomever. And since MIPS
    isn't part of SGI anymore it also applies to the MIPS CPUs...

    > In this case, I think the idea was to leverage Intel's market
    > dominance and probable performance dominance.


    With the Altix series, sure. Thats a really good piece of hardware.

    Benjamin



  10. Re: IRIX freeware releases

    Per Ekman wrote:

    >> IMHO that's a really sad excuse for stagnation.

    >
    > What is stagnating?


    The development of IRIX. Yes, they are fixing bugs and improving 6.5 but
    there is nothing new in sight...

    >> The point is that most other platform vendors _do_ major release
    >> upgrades with the software vendors following them. I understand SGI
    >> is afraid of loosing the few remaining companies developing for IRIX
    >> but besides the technical side it's very bad marketing since it
    >> makes people think SGI isn't interested in IRIX any more.

    >
    > So they should leave a technically sound, working solution in favor of
    > unsubstantiated PR? That's a great way to piss users off.


    Uh? Did I miss something or did SGI loose one customer after another during
    the last years?

    So why is any new development made if all the "technically sound" and
    "working" solutions from today do the job?

    I understand Your point, but since it's more marketing that decides about
    live or dead of a company than does technology it is very stupid to stagnate
    on the 6.5.x numbers.

    >> It depends. I have the feeling it makes _SGI_ unhappy as every new
    >> release of a vendors software product might be the one that doesn't
    >> support IRIX any more. If software vendors are running away from
    >> Your platform of course You are afraid of changes that might cause
    >> these vendors to think about their IRIX efforts twice.

    >
    > As well they should be. And not just vendors. As a customer, I value
    > the backwards compatibilty very much.


    Sure, but what makes You believe a much improved IRIX (say IRIX 6.6 or IRIX
    7) would brake compatibility? Most IRIX 6.2 programs still run under 6.5.x,
    and even 5.3 binaries if they are not COFF. SGI maintained backward
    compatibility over almost all of their IRIX releases, so I see no reason
    that any really new IRIX version has to brake that, at least not to a level
    that is really critical.

    >> With the current situation SGI could really use some "marketing
    >> signal" that shows that there is still done something. Maybe IRIX is
    >> still up to date, but the outside world gets the impression that it
    >> is not. Usually the last numbers are used to show the patch level,
    >> and a numbering like 6.5.22 says that it's mainly a bug fix for an
    >> operating system that hasn't seen a major update since 1998! Maybe
    >> there is any progress inside IRIX, but that is definitely not shown
    >> by SGI. And while the competition is producing a new OS Release
    >> every 1-3 years SGI will still be at 6.5 when offering the 'IRIX
    >> 6.55.485m' update.

    >
    > And THAT IS A GOOD THING!!


    No, it's not. of course over time You _might_ end up in a highly improved
    operating system (but not necessarily), but it is still the old architecture
    that You once had when You started. And at some point nail-fixing things
    gets more difficult than writing new from scratch. That's btw why most other
    system vendors do regularly updates of their operating system. And most of
    them do that without braking compatibility with older software.

    > The failure here is not the numbering
    > scheme but arguably one of marketing. You make a technical decision,
    > it is then up to marketing to convince the customers of the benefits.


    Well, it's hard to convince any customer that IRIX should have a future when
    there less apps than even for other commercial UNIXes and when the last new
    OS Release was 1998. You have to do hypnotism to convince anyone that this
    should be a _good_ sign. Maybe SGI thinks this way, but they already proofed
    in the past that they don't know sh*t about marketing.

    > When you change technical decisions on non-technical grounds, that's
    > when you
    > fail miserably.


    Right, but usually engineers are very bad sales people. And it happend more
    than once that the incompetence of engineers in marketing decisions drove
    technically valuable companies down the river...

    >> For most platform manufacturers a new OS Release usually is a
    >> positive thing, providing some marketing hype and showing the
    >> customers that their platform is still alive and being improved.

    >
    > And a bad thing for customers, who end up being forced to change
    > applications and methods each time.


    No, not necessarily. When we do OS upgrades we usually keep most
    applications, and those that make problems usually can be fixed by a
    vendor's patch. That's at least how it works for us on the platforms we are
    using. Can't say anything about SGI since we don't use it.

    >> IMHO SGI shouldn't be too afraid of changes...

    >
    > IMO people should argue more about what should be improved in Irix and
    > less about what number is attached to it.


    Well, there were a lot of discussions about what people want to see in IRIX,
    of course without any result because everyone wants something different, and
    I personally doubt that IRIX or the MIPS platform will have any future. If
    SGI is still alive in say ten years from now it probably is because of their
    Itanium/Linux efforts and not because of IRIX/MIPS.

    Benjamin



  11. Re: IRIX freeware releases

    Benjamin Gawert wrote:
    > That probably applies to 99% of all SGI hardware. Most chips inside SGI
    > machines are not made by SGI but from intel, LSI or whomever. And since MIPS
    > isn't part of SGI anymore it also applies to the MIPS CPUs...



    Bzzt. Wrong. The R1x000 MIPS line in SGI systems is still designed
    within SGI: MIPS, inc. is designing the MIPS32 and MIPS64 family of
    "embedded" chips.

    --
    Alexis Cousein Senior Systems Engineer
    alexis@sgi.com SGI/Silicon Graphics Brussels




  12. Re: IRIX freeware releases

    Alexis Cousein wrote:

    > Bzzt. Wrong. The R1x000 MIPS line in SGI systems is still designed
    > within SGI: MIPS, inc. is designing the MIPS32 and MIPS64 family of
    > "embedded" chips.


    Oh yes, I missed that. Sorry.

    Benjamin



  13. Re: IRIX freeware releases

    Benjamin Gawert wrote:

    > Ivan Rayner wrote:
    >
    > >> Well, IA64 is definitely not what one would understand as "comodity
    > >> parts".

    > >
    > > That depends on your point of view. A commodity part for SGI, is
    > > something that is developed by someone else that is either cheap and
    > > does its job well, or has other specific advantages over SGI
    > > developing
    > > the part internally.

    >
    > That probably applies to 99% of all SGI hardware. Most chips inside SGI
    > machines are not made by SGI but from intel, LSI or whomever. And since MIPS
    > isn't part of SGI anymore it also applies to the MIPS CPUs...


    uh, the LSI chips are generally gate array or standard cells ASIC's with logic
    and I/O custom designed by SGI. they are not by definition a 'commodity'.

    >
    >
    > > In this case, I think the idea was to leverage Intel's market
    > > dominance and probable performance dominance.

    >
    > With the Altix series, sure. Thats a really good piece of hardware.
    >


    for a big peecee, yeah.

    >
    > Benjamin



  14. Re: IRIX freeware releases

    "Benjamin Gawert" writes:

    > Per Ekman wrote:
    >
    > >> IMHO that's a really sad excuse for stagnation.

    > >
    > > What is stagnating?

    >
    > The development of IRIX. Yes, they are fixing bugs and improving 6.5 but
    > there is nothing new in sight...


    Tell me concretely what you want to see improved in Irix. And a
    prettier desktop or a DVD-player is not a part of the OS.

    You seem to think that (tell me if I'm wrong here) :

    A) an unchanged major version number means that the OS isn't being
    developed

    B) changing an OS in a major way is in itself a good thing

    I think both of these are wrong. I feel that the job of an OS is to
    work and work well. If the OS should be overhauled then there has to
    be a sound technical reason for it. The only reason that I can think
    of off the top of my head is an IPF-version of Irix, but that's not
    going to happen.

    > Sure, but what makes You believe a much improved IRIX (say IRIX 6.6 or IRIX
    > 7) would brake compatibility? Most IRIX 6.2 programs still run under 6.5.x,
    > and even 5.3 binaries if they are not COFF. SGI maintained backward
    > compatibility over almost all of their IRIX releases, so I see no reason
    > that any really new IRIX version has to brake that, at least not to a level
    > that is really critical.


    Binary compatibility is the "easy" part. Ivans example with xfsdump is
    a good one. If the semantics of the xfsdump return value changed we'd
    have to rewrite and retest all out Irix backup systems.

    Even Cray (admittedly SGI at the time), which have traditionally been
    _very_ good at providing backwards compatible systems, screwed up in a
    major way between Unicos 9 and 10.

    > > > <6.5 going on for ever>

    > > And THAT IS A GOOD THING!!

    >
    > No, it's not. of course over time You _might_ end up in a highly
    > improved operating system (but not necessarily), but it is still the
    > old architecture that You once had when You started. And at some
    > point nail-fixing things gets more difficult than writing new from
    > scratch. That's btw why most other system vendors do regularly
    > updates of their operating system. And most of them do that without
    > braking compatibility with older software.


    Other vendors either a) make lousy engineering choices so they have to
    rewrite their OS every year, b) change major version numbers to
    support radically new hardware, or c) change major version numbers for
    marketing reasons. Pick whatever you feel is the most likely.

    > > The failure here is not the numbering scheme but arguably one of
    > > marketing. You make a technical decision, it is then up to
    > > marketing to convince the customers of the benefits.

    >
    > Well, it's hard to convince any customer that IRIX should have a
    > future when there less apps than even for other commercial UNIXes
    > and when the last new OS Release was 1998.


    SGI releases a new OS version 4 times a year. You're just hung up on
    which of the three numbers in the version changes.

    > You have to do hypnotism to convince anyone that this should be a
    > _good_ sign.


    You've yet to convince me that "less apps" and "OS versioning scheme"
    are connected.

    > > When you change technical decisions on non-technical grounds,
    > > that's when you fail miserably.

    >
    > Right, but usually engineers are very bad sales people. And it happend more
    > than once that the incompetence of engineers in marketing decisions drove
    > technically valuable companies down the river...


    Yes, and vice versa.

    > >> For most platform manufacturers a new OS Release usually is a
    > >> positive thing, providing some marketing hype and showing the
    > >> customers that their platform is still alive and being improved.

    > >
    > > And a bad thing for customers, who end up being forced to change
    > > applications and methods each time.

    >
    > No, not necessarily. When we do OS upgrades we usually keep most
    > applications, and those that make problems usually can be fixed by a
    > vendor's patch. That's at least how it works for us on the platforms we are
    > using. Can't say anything about SGI since we don't use it.


    It comes down to wether you can tolerate the problems or not. An
    application that stops working can cause bad data and weeks of extra
    work convincing the vendor that there is a problem and to get a patch
    written, tested and released.

    > >> IMHO SGI shouldn't be too afraid of changes...

    > >
    > > IMO people should argue more about what should be improved in Irix and
    > > less about what number is attached to it.

    >
    > Well, there were a lot of discussions about what people want to see
    > in IRIX, of course without any result because everyone wants
    > something different,


    What do you want SGI to do then, change something randomly?

    *p

+ Reply to Thread
Page 2 of 2 FirstFirst 1 2