Software vs service scalability

Discussion in 'NZ Computing' started by Lawrence D¹Oliveiro, Mar 21, 2005.

  1. Issue 45 of Linux User & Developer has an interview with Jon "Maddog"
    Hall where he brings up an interesting issue to do with how you support
    millions of users all trying to run the the same software product:

    The problems we experience with software today are only going to
    get worse. When people started in this business, they hired perhaps
    a hundred engineers, and they had a thousand customers. When one
    of those customers had a problem, they called the company, and they
    would typically get the president of the company on the line, or the
    chief programmer, or someone like that. They'd explain the problem,
    and the person on the other end would know exactly what the
    problem was, because they wrote the software -- and you'd get the
    fix. But today, the same company might have 4[-]5 million customers.
    So when you call them up, who are you going to get? Certainly not
    the president! What you get is "Push button number 1 for service"
    and some muzak...
    It's not because the companies don't want to help their customers
    -- they can't. They're being pulled in too many different directions.
    When you talk to companies about this issue, they say it's difficult
    for a large company to scale in the service marketplace. It's easy to
    make product -- Microsoft makes 80% profit on every single licence
    they sell. They can do this because they're just manufacturing the
    stuff and tossing it over the wall.
    From our viewpoint, what they're doing now is not scaling --
    because we can't get the service to make the software do what
    it's supposed to do. Every time someone comes up to me and
    says the service thing doesn't scale, I say "No -- the product doesn't
    scale, and it hasn't been scaling for the last ten years. You've got
    to fix it, and the only way you can fix it is to open up your source
    code, so that people can go out and hire people to make the
    software work."
    Companies say "It's not in our best interest to do that". Pardon
    me, but what about *my* best interest? I'm your customer, and it's
    in my best interest to have your software in the native language
    I speak, to support the printer I have, to fix bugs and extend it...
     
    Lawrence D¹Oliveiro, Mar 21, 2005
    #1
    1. Advertising


  2. > Companies say "It's not in our best interest to do that". Pardon
    > me, but what about *my* best interest? I'm your customer, and it's
    > in my best interest to have your software in the native language
    > I speak, to support the printer I have, to fix bugs and extend it...


    heh.. thats been the point of Open Source all along, let the customer help
    everyone by customising it and tailoring it for things never even
    envisioned for the Original Product
    --
    I Propose a Holy war !!?!?!
     
    Shane (aka froggy), Mar 21, 2005
    #2
    1. Advertising

  3. Lawrence D¹Oliveiro

    Mercury Guest

    what a load of generalised dribble.

    It reads like this "I can't read the manual, I can't use google, I can't
    read books, I can't use best practices, I can't architect systems, *BUT* I
    can fix everything if I have Open Source". Absolute CRAP.

    It would seem that the author needs to train his staff.

    I can't be bothered detailing other arguments other than to say that of all
    the s/w vendors I and my customers have dealings with, MS fails least. 2
    software support calls in 10 years.


    "Lawrence D¹Oliveiro" <_zealand> wrote in message
    news:...
    > Issue 45 of Linux User & Developer has an interview with Jon "Maddog"
    > Hall where he brings up an interesting issue to do with how you support
    > millions of users all trying to run the the same software product:
    >
    > The problems we experience with software today are only going to
    > get worse. When people started in this business, they hired perhaps
    > a hundred engineers, and they had a thousand customers. When one
    > of those customers had a problem, they called the company, and they
    > would typically get the president of the company on the line, or the
    > chief programmer, or someone like that. They'd explain the problem,
    > and the person on the other end would know exactly what the
    > problem was, because they wrote the software -- and you'd get the
    > fix. But today, the same company might have 4[-]5 million customers.
    > So when you call them up, who are you going to get? Certainly not
    > the president! What you get is "Push button number 1 for service"
    > and some muzak...
    > It's not because the companies don't want to help their customers
    > -- they can't. They're being pulled in too many different directions.
    > When you talk to companies about this issue, they say it's difficult
    > for a large company to scale in the service marketplace. It's easy to
    > make product -- Microsoft makes 80% profit on every single licence
    > they sell. They can do this because they're just manufacturing the
    > stuff and tossing it over the wall.
    > From our viewpoint, what they're doing now is not scaling --
    > because we can't get the service to make the software do what
    > it's supposed to do. Every time someone comes up to me and
    > says the service thing doesn't scale, I say "No -- the product doesn't
    > scale, and it hasn't been scaling for the last ten years. You've got
    > to fix it, and the only way you can fix it is to open up your source
    > code, so that people can go out and hire people to make the
    > software work."
    > Companies say "It's not in our best interest to do that". Pardon
    > me, but what about *my* best interest? I'm your customer, and it's
    > in my best interest to have your software in the native language
    > I speak, to support the printer I have, to fix bugs and extend it...
     
    Mercury, Mar 21, 2005
    #3
  4. Lawrence D¹Oliveiro

    Steve Guest

    Mercury wrote:
    > what a load of generalised dribble.
    >
    > It reads like this "I can't read the manual, I can't use google, I can't
    > read books, I can't use best practices, I can't architect systems, *BUT* I
    > can fix everything if I have Open Source". Absolute CRAP.

    That's pointless because the software isn't behaving as stated, it's
    broken. But you missed that bit, obviously. And the scaling bit, which
    was the point of the article.
    >
    > It would seem that the author needs to train his staff.

    You'd better nip over and tell him, then. However, I doubt you've even
    heard of him. I'd use google before you make even more of a fool of
    yourself.
    >
    > I can't be bothered detailing other arguments other than to say that of all
    > the s/w vendors I and my customers have dealings with, MS fails least. 2
    > software support calls in 10 years.
    >

    Must be the only times you switched it on then. If you've only been
    using computers for the last 10 years, then Maddog has been in the
    software business since long before you were born.

    Steve
     
    Steve, Mar 21, 2005
    #4
  5. Lawrence D¹Oliveiro

    Mercury Guest

    "Steve" <> wrote in message
    news:d1m2a6$5i2$...
    > Mercury wrote:
    >> what a load of generalised dribble.
    >>
    >> It reads like this "I can't read the manual, I can't use google, I can't
    >> read books, I can't use best practices, I can't architect systems, *BUT*
    >> I can fix everything if I have Open Source". Absolute CRAP.

    > That's pointless because the software isn't behaving as stated,


    Thats called a bug. If he has a legal copy of the product then he should
    feel happy with the ability provided to call PSS and get the incorrect
    behaviour fixed. He *will* get a competant support engineer and he *will* -
    if it is a real bug - get some one at the end of the line promptly that is
    not just competant, but has every resource and connection needed to solve
    such supposed show stoppers. These people are good and persistent and are a
    credit to the entire industry.

    > it's broken. But you missed that bit, obviously.


    Nope, I did state I wasn't going to go into detail at that point. Remember?

    > And the scaling bit, which was the point of the article.


    He eludes to it, sure, but how do you know? He doesn't say what type of
    software, what issues (scaling of what?), what he has done to track down
    these issues, where and why MS is failing him etc. The reality is that by
    far most developers just do not know enough to read os source code
    meaningfully and productively. Now this bod may know gobs about linux, but
    get real - what real use is Windows source code going to be to him (he'd
    need a nanny for the first week to find his way around), and more inline
    with the article, what use will the source be to all the other middle tier
    developers out there? He is rather pathetically trying to arm twist (who?)
    MS into opening Windows source up! I'd suggest that this chap is using the
    wrong approach and suffering the consequences largely by his own choice to
    justify the ends he wishes to arrive at. IE "I wrote an app, don't work
    well, won't ring MS for help, so I'll ask for source code access, get turned
    down, and MOAN ABOUT IT". FFS. Doesn't sound very bright to me...

    You need to be realistic about this.

    How many millions of lines of code are there in Windows kernel? How quickly
    can a person digest enough of it for it to be beneficial? The issue is not
    about Him, it is about the target market of his comments. IE Joe Bloggs
    ordinary programmer that does not know C, does not know how OS are really
    constructed etc - THE READER. "Propaganda" springs to mind.

    He isn't being realistic.

    >>
    >> It would seem that the author needs to train his staff.

    > You'd better nip over and tell him, then. However, I doubt you've even
    > heard of him. I'd use google before you make even more of a fool of
    > yourself.


    I'd be happy to buy him a beer or take over a carton of steinies / CHCH
    water / whatever and have a chat, but I am a realist & have work to do so
    unless someone sponsors me I am not about to splash out for my airfares and
    suffer loss of income to boot. So, hows about some sponsors for a showdown?
    Put your money where your mouth is.

    How is it that he is the only one raising these issues as gripes about MS?
    The article starts out well then he tosses out credibility by indicating
    lack of scalability without any supporting evidence. So is he playing with
    32 way xeon systems and Windows 2000 Data Center, or is it Win 3.1? He
    doesn't say. If he is going to make claims, then the claims are meaningless
    unless they are qualified with adequate fact. What is he comparing to? Seems
    like just another one of these brain dead Pro Linux articles...

    Knowing how to program linux productively does not for a nano second mean
    that a person has any qualification for Windows. Have you ever hired a
    university graduate who's love in life is ANSI C in an academic environment?

    Does the author know about and is the app / system using any of the
    following? IO Completion Ports, Spin Locks, Interlocks, Overlapped IO,
    Mutexes, Semaphores, Threads, Fibres, Memory Mapped files ... to name a few
    that spring to mind.

    Better yet, rather than a fishing expedition, why didn't the author bother
    to state where the issues are, why they arise, and what he has tried to do
    to solve the problems?

    I ask of these as many apps are written to C standard without taking
    advantage of performance benefiting OS API's in windows. Some of these poor
    performing systems are large and famous - they do not scale or perform as
    they should due to laziness of the programmers or failures in the
    arhitecture process or both. They fail to take advantage of the OS as it is
    designed and intended to be used. You may be surprised to learn what is on
    that list.

    Just how is it that so many server systems on windows out perform Linux
    systems? Lack of scalability?

    >>
    >> I can't be bothered detailing other arguments other than to say that of
    >> all the s/w vendors I and my customers have dealings with, MS fails
    >> least. 2 software support calls in 10 years.
    >>

    > Must be the only times you switched it on then.


    Ahhhhh there are a lot of resources out there. Thats obviously the
    difference between you and me. If I strike a problem, I solve it using the
    resources available - that is the secret - using the resources available.

    I have had all versions of windows since 3.0 except data center server and
    perhaps some 'Advanced Server' versions.. Can the author of the article say
    the same? Is he half as familiar with Win32 API as any seasoned windows C /
    C++ programmer would be? If his apps suffer scalability problems, then it
    seems not.

    > If you've only been using computers for the last 10 years, then Maddog has
    > been in the software business since long before you were born.


    Rule number 1 of computers. NEVER ASSUME. So, no, I have not been using
    computers for only 10 years.

    That suggestion of 10 years is also totally irrelevant - the author is
    trying to be authoratative about Windows. Does he have 20 years with
    windows? 10? 5? 2? 1? 1 - maybe. I bet there are kids out there that out
    learn 99% of COSC masters graduates in a really short time. I went to school
    with a few...

    10 years is a arbitrary period of time. Would you feel better if I had said
    20 or 30 years? Tell me what a 4004 and a 2102 is and I'll give you 1 browny
    point. (hint they are both chips and no internet search engines to find
    out).


    >
    > Steve
     
    Mercury, Mar 21, 2005
    #5
  6. Lawrence D¹Oliveiro

    -=rjh=- Guest

    Mercury wrote:
    >>
    >>"Steve" <> wrote in message
    >>
    >>You'd better nip over and tell him, then. However, I doubt you've even
    >>heard of him. I'd use google before you make even more of a fool of
    >>yourself.

    >


    ROTFL :)

    >
    > I'd be happy to buy him a beer or take over a carton of steinies / CHCH
    > water / whatever and have a chat, but I am a realist & have work to do so
    > unless someone sponsors me I am not about to splash out for my airfares and
    > suffer loss of income to boot. So, hows about some sponsors for a showdown?
    > Put your money where your mouth is.


    Well - he'll very likely be in Dunedin next year. You still don't have a
    clue who he is, do you?
     
    -=rjh=-, Mar 21, 2005
    #6
  7. -=rjh=- wrote:
    >>> You'd better nip over and tell him, then. However, I doubt you've
    >>> even heard of him. I'd use google before you make even more of a fool
    >>> of yourself.


    >> I'd be happy to buy him a beer or take over a carton of steinies /
    >> CHCH water / whatever and have a chat, but I am a realist & have work
    >> to do so unless someone sponsors me I am not about to splash out for
    >> my airfares and suffer loss of income to boot. So, hows about some
    >> sponsors for a showdown? Put your money where your mouth is.


    > Well - he'll very likely be in Dunedin next year. You still don't have a
    > clue who he is, do you?


    Does it matter if he knows who he is, will he automatically cower in
    shame about the things he has said?
     
    Dave - Dave.net.nz, Mar 21, 2005
    #7
  8. Lawrence D¹Oliveiro

    -=rjh=- Guest

    Mercury wrote:
    > what a load of generalised dribble.


    I disagree - I think he makes some really good points. In a general sort
    of way :)
    >
    > It reads like this "I can't read the manual, I can't use google, I can't
    > read books, I can't use best practices, I can't architect systems, *BUT* I
    > can fix everything if I have Open Source". Absolute CRAP.


    You might be reading it like that, but that's not what I take out of it.
    What I get is that companies with large user bases just cannot cope with
    support issues. Open source is one way of helping with that. Look how
    important the role Google (as you point out) and user groups, newsgroups
    and forums play in removing support issues from the original software
    vendor. That process is probably significantly assisted if source is
    available.
    >
    > It would seem that the author needs to train his staff.
    >
    > I can't be bothered detailing other arguments other than to say that of all
    > the s/w vendors I and my customers have dealings with, MS fails least. 2
    > software support calls in 10 years.


    And the other s/w vendors were what? All open source? What does your
    statement show? The article wasn't about MS per se, but relates to
    issues that all companies with large user bases have.
     
    -=rjh=-, Mar 21, 2005
    #8
  9. Lawrence D¹Oliveiro

    -=rjh=- Guest

    Dave - Dave.net.nz wrote:
    > -=rjh=- wrote:
    >
    >>>> You'd better nip over and tell him, then. However, I doubt you've
    >>>> even heard of him. I'd use google before you make even more of a
    >>>> fool of yourself.

    >
    >
    >>> I'd be happy to buy him a beer or take over a carton of steinies /
    >>> CHCH water / whatever and have a chat, but I am a realist & have work
    >>> to do so unless someone sponsors me I am not about to splash out for
    >>> my airfares and suffer loss of income to boot. So, hows about some
    >>> sponsors for a showdown? Put your money where your mouth is.

    >
    >
    >> Well - he'll very likely be in Dunedin next year. You still don't have
    >> a clue who he is, do you?

    >
    >
    > Does it matter if he knows who he is, will he automatically cower in
    > shame about the things he has said?


    Nope, but maybe be less likely to dismiss what was said out of hand.
    Knowing something about the author puts the article in context.
     
    -=rjh=-, Mar 21, 2005
    #9
  10. Lawrence D¹Oliveiro

    Impossible Guest

    "-=rjh=-" <> wrote in message
    news:d1nkpa$7ba$...
    > Dave - Dave.net.nz wrote:
    >>
    >> Does it matter if he knows who he is, will he automatically cower
    >> in shame about the things he has said?

    >
    > Nope, but maybe be less likely to dismiss what was said out of hand.
    > Knowing something about the author puts the article in context.


    Ok, I'll bite. Who the heck is the author and what's the big mystery?
     
    Impossible, Mar 22, 2005
    #10
  11. Lawrence D¹Oliveiro

    Tim Guest

    The author is irrelevant.
    It is the content that has been questioned.
    If knowing the author affects the value of the content, then the content is
    incomplete in itself.

    "Impossible" <> wrote in message
    news:...
    > "-=rjh=-" <> wrote in message
    > news:d1nkpa$7ba$...
    >> Dave - Dave.net.nz wrote:
    >>>
    >>> Does it matter if he knows who he is, will he automatically cower
    >>> in shame about the things he has said?

    >>
    >> Nope, but maybe be less likely to dismiss what was said out of hand.
    >> Knowing something about the author puts the article in context.

    >
    > Ok, I'll bite. Who the heck is the author and what's the big mystery?
    >
    >
    >
     
    Tim, Mar 22, 2005
    #11
  12. In article <d1m8u6$irg$>, "Mercury" <>
    wrote:

    >If he has a legal copy of the product then he should
    >feel happy with the ability provided to call PSS and get the incorrect
    >behaviour fixed. He *will* get a competant support engineer and he *will* -
    >if it is a real bug - get some one at the end of the line promptly that is
    >not just compet[e]nt, but has every resource and connection needed to solve
    >such supposed show stoppers.


    That doesn't jibe with the well-known fact that many closed-source
    software vendors have absolutely atrocious customer support. With open
    source, if you don't like one vendor's support, you can go
    elsewhere--without having to rip out the software and put in something
    else.

    >How many millions of lines of code are there in Windows kernel? How quickly
    >can a person digest enough of it for it to be beneficial?


    The problem is that the Windows code was never designed to be
    peer-reviewed by an entire community of developers ever ready to suggest
    improvements, so it is certainly starting off on the back foot there.

    Imagine if Windows had a resource equivalent to what
    <http://lxr.linux.no/> provides for the Linux source code...

    >Does the author know about and is the app / system using any of the
    >following? IO Completion Ports, Spin Locks, Interlocks, Overlapped IO,
    >Mutexes, Semaphores, Threads, Fibres, Memory Mapped files ... to name a few
    >that spring to mind.


    And what are most of them but concepts, possibly with Windows-peculiar
    names, that are pretty much the same across most modern OSes? Apart from
    the Windows-peculiar concepts that aren't worth using anyway...
     
    Lawrence D¹Oliveiro, Mar 22, 2005
    #12
  13. Lawrence D¹Oliveiro

    -=rjh=- Guest

    Tim wrote:
    > The author is irrelevant.
    > It is the content that has been questioned.
    > If knowing the author affects the value of the content, then the content is
    > incomplete in itself.


    Heh, not meaning to be rude or anything, but you sound like Mr Spock :)

    In this case, the context of who the author is, is relevant. He is the
    Executive Director of Linux International, which is a "non-profit
    association of user groups, corporations, and other interested parties
    that promotes the development of Linux."

    Hardly somebody with an unbiased viewpoint. And the article was
    expressing opinion.


    >
    > "Impossible" <> wrote in message
    > news:...
    >
    >>"-=rjh=-" <> wrote in message
    >>news:d1nkpa$7ba$...
    >>
    >>>Dave - Dave.net.nz wrote:
    >>>
    >>>>Does it matter if he knows who he is, will he automatically cower
    >>>>in shame about the things he has said?
    >>>
    >>>Nope, but maybe be less likely to dismiss what was said out of hand.
    >>>Knowing something about the author puts the article in context.

    >>
    >>Ok, I'll bite. Who the heck is the author and what's the big mystery?
    >>
    >>
    >>

    >
    >
    >
     
    -=rjh=-, Mar 22, 2005
    #13
  14. In article <d1os64$opl$>,
    -=rjh=- <> wrote:

    >In this case, the context of who the author is, is relevant. He is the
    >Executive Director of Linux International, which is a "non-profit
    >association of user groups, corporations, and other interested parties
    >that promotes the development of Linux."
    >
    >Hardly somebody with an unbiased viewpoint. And the article was
    >expressing opinion.


    That's like saying that somebody from Friends of the Earth isn't
    qualified to express an opinion about the environment, because they're
    obviously biased in favour of it. :)
     
    Lawrence D¹Oliveiro, Mar 22, 2005
    #14
  15. Lawrence D¹Oliveiro

    -=rjh=- Guest

    Lawrence D¹Oliveiro wrote:
    > In article <d1os64$opl$>,
    > -=rjh=- <> wrote:
    >
    >
    >>In this case, the context of who the author is, is relevant. He is the
    >>Executive Director of Linux International, which is a "non-profit
    >>association of user groups, corporations, and other interested parties
    >>that promotes the development of Linux."
    >>
    >>Hardly somebody with an unbiased viewpoint. And the article was
    >>expressing opinion.

    >
    >
    > That's like saying that somebody from Friends of the Earth isn't
    > qualified to express an opinion about the environment, because they're
    > obviously biased in favour of it. :)


    No its not - I'm just saying that knowing the author's background is
    relevant. I wasn't saying that they weren't qualified to express an
    opinion at all. I take a lot of cues from extraneous information that
    isn't in the article; this is probably true of any article of this type.

    Because this article was written by Jon Hall, I read it as being an
    attempt to portray what he thinks about a particular subject. If it had
    been written by Bill Gates, I'd probably interpret it as satire. If it
    was written by Enderle or Gartner or somebody of that ilk I'd want to
    know who paid for it.
     
    -=rjh=-, Mar 23, 2005
    #15
    1. Advertising

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

It takes just 2 minutes to sign up (and it's free!). Just click the sign up button to choose a username and then you can ask your own questions on the forum.
Similar Threads
  1. AM
    Replies:
    1
    Views:
    380
    Walter Roberson
    Mar 7, 2005
  2. charles cheng
    Replies:
    0
    Views:
    504
    charles cheng
    Nov 19, 2003
  3. Salmonaxe
    Replies:
    0
    Views:
    525
    Salmonaxe
    Nov 2, 2006
  4. Bling Bling
    Replies:
    6
    Views:
    388
    whoisthis
    Nov 5, 2005
  5. Lawrence D'Oliveiro

    When Is A Visaster Service Pack Not A Service Pack?

    Lawrence D'Oliveiro, Jun 26, 2008, in forum: NZ Computing
    Replies:
    5
    Views:
    481
Loading...

Share This Page