How many "advocates" have me killfiled? - Linux

This is a discussion on How many "advocates" have me killfiled? - Linux ; In article , Linonut wrote: > > If Linux users are more likely to use Facebook than the general computer > > using population is, then the percentage of Linux users found among > > Facebook users would be higher ...

+ Reply to Thread
Page 4 of 9 FirstFirst ... 2 3 4 5 6 ... LastLast
Results 61 to 80 of 163

Thread: How many "advocates" have me killfiled?

  1. Re: How many "advocates" have me killfiled?

    In article ,
    Linonut wrote:
    > > If Linux users are more likely to use Facebook than the general computer
    > > using population is, then the percentage of Linux users found among
    > > Facebook users would be higher than the percentage in the general
    > > computer using population. If Linux usage is independent of Facebook
    > > usage, then the percentage of Linux users found among Facebook users
    > > would be about the same as the percentage in the general computer using
    > > population. In both of these cases, Linux use on Facebook would give us
    > > a good upper limit on general Linux use.

    >
    > No, it would give us the middle of the upper and lower limits, wouldn't
    > it?


    If Linux users are overrepresented among Facebook users, you get a
    higher percentage there than in the general computer using population.
    So, the Facebook number in that case would be an upper limit. If Linux
    users are in the Facebook population at the same percentage as they are
    in the general computer population, then the Facebook number would give
    you the Linux number. That, too, works as an upper limit, since it
    would not be lower than the correct number.

    > > You seem to doubt that we can get good upper limits from sites like
    > > this, which would mean that they have a lower use than would be expected
    > > based on Linux use in the general population.

    >
    > Why do you think it is possible to use Facebook as a "good upper limit"
    > /or/ a "good lower limit"?
    >
    > Personally, since Linux is still somewhat of a techie system, I would
    > suspect that most Facebook users are Windowsers. (Except, of course,
    > for guys like Reuven Lerner of Linux Journal, who has writtem some
    > articles about the "facebook API".)


    You seem to be assuming that techies would be less likely than
    non-techies to use Facebook, and so Linux would be underrepresented on
    Facebook. Certainly that is possible.

    But I can't think of any particular reason Linux users would not be just
    as interested in Facebook as Windows users or Mac users (since I believe
    that the stereotype of techies being friendless basement dwellers with
    no interest in social interaction is just that--a stereotype).

    --
    --Tim Smith

  2. Re: How many "advocates" have me killfiled?

    In article ,
    High Plains Thumper wrote:
    [blah blah blah]

    You know, HPT, your "I don't nym shift" claims would work a lot better
    if you hadn't posted that report about your nym shifting experiments.


    --
    --Tim Smith

  3. Re: How many "advocates" have me killfiled?

    In article ,
    Kelsey Bjarnason wrote:
    > > Then if there is no correlation between Linux and visiting facebook,

    >
    > then the usage of facebook has no correlation to the use of Linux. Yes,
    > finally, you're getting it.
    >
    > Oh, no, you're not.


    Let

    Lf = number of Linux users who use Facebook
    L = number of Linux users
    Wf = number of Windows users who use Facebook
    W = number of Windows users

    (To simplify things, we'll ignore Mac users).

    If Facebook usage isn't correlated with OS usage, then we'd expect

    Lf/L = Wf/W (1)

    The fraction of Linux users in the general computer using population is

    L / (L + W) = 1 / (1 + W/L) (2)

    and the fraction of Linux users in the Facebook population is

    Lf / (Lf + Wf) = 1 / (1 + Wf/Lf) (3).

    But from (1), we have W/L = Wf/Lf, and so (2) and (3) are equal.

    If Lf/L > Wf/W, then W/L > Wf/Lf, and (2) < (3). (Facebook has a higher
    fraction of Linux users than the general population does).

    If Lf/L < Wf/W, then W/L < Wf/Lf, and (2) > (3). (Facebook has a lower
    fraction of Linux users than the general population does).

    Do you have any reason to believe that Facebook use is correlated with
    OS use, and so Lf/L != Wf/W?

    --
    --Tim Smith

  4. Re: How many "advocates" have me killfiled?

    * William Poaster peremptorily fired off this memo:

    > I haven't noticed much of anything from googlegroups, as they just don't
    > seem to bother about the inordinate amount of spam coming through them.
    > Consequently, like others, I block them & whitelist people I want to see.
    > http://improve-usenet.org/
    >
    > In fact googlegroups is so bad, one Swedish usenet provider blocks them,
    > & announces: "No more google groups posts on this server."
    > http://www.usenet4all.se/
    >
    > (I can almost hear Quack screaming "censorship!")


    The first link notes:

    Lamers

    Now, about the rest of the Google Groups problem.

    Even ignoring the spam issue for the sake of further discussion here,
    most of the people who post to Usenet via the clunky Google Groups
    web interface are lusers or lamers. Because of their use of a clunky
    Usenet web interface (and all Usenet web interfaces suck - Usenet
    wasn't designed for webification and does not need
    webification), they have no idea what Usenet is, how it works,
    or how to use it properly. And, generally, they don't want to learn.

    Here's a truly priceless quote from a Google Grouper accused of
    spamming a newsgroup (with my elision of a couple of letters), which
    nicely illustrates this: "As you f**kwits keep reminding me, YOU are
    on usenet. The rest of the millions upon millions of us real world
    readers are on the internet." Ah! I get it -- it's FedEx or
    something like that that carries Usenet, not the Internet. Heh.

    "Real world". Remind you of any lame f**kwits?

    --
    Gary Kildall was one of the original pioneers of the PC revolution. He was a
    very creative computer scientist who did excellent work. Although we were
    competitors, I always had tremendous respect for his contributions to the PC
    industry. His untimely death was very unfortunate and he and his work will
    be missed.
    -- Bill Gates, The Computer Chronicles. "Special Edition: Gary Kildall." 1995

  5. Re: How many "advocates" have me killfiled?

    * Tim Smith peremptorily fired off this memo:

    > You seem to be assuming that techies would be less likely than
    > non-techies to use Facebook, and so Linux would be underrepresented on
    > Facebook. Certainly that is possible.
    >
    > But I can't think of any particular reason Linux users would not be just
    > as interested in Facebook as Windows users or Mac users (since I believe
    > that the stereotype of techies being friendless basement dwellers with
    > no interest in social interaction is just that--a stereotype).


    Well, apart from usenet, I don't really do much social interaction over
    a network. My view is, I quite agree, personal and probably myopic.

    And since Microsoft has a stake in Facebook, I'm even less inclined to
    use it -- I quite agree, personal and probably myopic.

    --
    Success is a lousy teacher. It seduces smart people into thinking they can't
    lose.
    -- Bill Gates

  6. Re: How many "advocates" have me killfiled?

    On Wed, 04 Jun 2008 07:22:03 -0400, Linonut wrote:

    > * William Poaster peremptorily fired off this memo:
    >
    >> I haven't noticed much of anything from googlegroups, as they just don't
    >> seem to bother about the inordinate amount of spam coming through them.
    >> Consequently, like others, I block them & whitelist people I want to
    >> see. http://improve-usenet.org/
    >>
    >> In fact googlegroups is so bad, one Swedish usenet provider blocks them,
    >> & announces: "No more google groups posts on this server."
    >> http://www.usenet4all.se/
    >>
    >> (I can almost hear Quack screaming "censorship!")

    >
    > The first link notes:
    >
    > Lamers
    >
    > Now, about the rest of the Google Groups problem.
    >
    > Even ignoring the spam issue for the sake of further discussion here,
    > most of the people who post to Usenet via the clunky Google Groups web
    > interface are lusers or lamers. Because of their use of a clunky Usenet
    > web interface (and all Usenet web interfaces suck - Usenet wasn't
    > designed for webification and does not need webification), they have no
    > idea what Usenet is, how it works, or how to use it properly. And,
    > generally, they don't want to learn.
    >
    > Here's a truly priceless quote from a Google Grouper accused of
    > spamming a newsgroup (with my elision of a couple of letters), which
    > nicely illustrates this: "As you f**kwits keep reminding me, YOU are on
    > usenet. The rest of the millions upon millions of us real world readers
    > are on the internet." Ah! I get it -- it's FedEx or something like that
    > that carries Usenet, not the Internet. Heh.


    Yeah, I saw that posted in another group (by "Blinky the Shark") in a
    discussion about googlegropers. LOL!

    > "Real world". Remind you of any lame f**kwits?


    It do! Oh, it do!

    --
    Mandriva 2008.1 64-bit.
    This message was sent from a
    computer which is guaranteed
    100% free of the M$ Windoze virus.

  7. Re: How many "advocates" have me killfiled?

    On Wed, 04 Jun 2008 03:50:21 -0700, Tim Smith wrote:

    > In article ,
    > High Plains Thumper wrote:
    > [blah blah blah]
    >
    > You know, HPT, your "I don't nym shift" claims would work a lot better
    > if you hadn't posted that report about your nym shifting experiments.


    Not to mention that he is not very good at it and often screws up and gets
    caught.

    --
    Moshe Goldfarb
    Collector of soaps from around the globe.
    Please visit The Hall of Linux Idiots:
    http://linuxidiots.blogspot.com/

  8. Re: How many "advocates" have me killfiled?

    On Wed, 04 Jun 2008 04:16:55 -0700, Tim Smith wrote:

    > In article ,
    > Kelsey Bjarnason wrote:
    >> > Then if there is no correlation between Linux and visiting facebook,

    >>
    >> then the usage of facebook has no correlation to the use of Linux.
    >> Yes, finally, you're getting it.
    >>
    >> Oh, no, you're not.

    >
    > Let


    Let me explain this to you in terms even you might be able to understand:

    The goal is to determine the number of Linux users from web stats.
    We can tell how many Linux users use the site.
    All we need now is to know how many Linux users there are, so we can
    perform the relevant math.

    I'm sorry, I realized I used terms such as "math", making this too
    complicated for you. For the rest, I'll simply point out that to make
    this work requires *already* knowing how many Linux users there are, in
    order to figure out how many Linux users there are, which is to say, the
    sort of endeavour which would make sense only to someone like Timmy.


  9. Re: How many "advocates" have me killfiled?

    On Tue, 03 Jun 2008 19:04:09 -0700, cc wrote:

    You're not though. You can snip the arguments all you want, but it
    doesn't make you look any smarter.


    Well, let's see if I can't simplify this enough for your pea brain to
    cope with.

    A site - we're using Facebook as a hypothetical example - has the
    following stats:

    100 million visitors
    90% Windows
    5% Linux
    5% Other

    How do we determine, from this, how many Linux users there are - not just
    Linux users of Facebook, but *total* Linux users (just counting
    "desktop", not server, usage)?

    Well, let's try it for Windows. We know 90% of our users run Windows.
    Thus, we need to know how popular Facebook is with Windows users - that
    is, what percentage of Windows users use facebook. Our stats won't tell
    us this, it only tells us some 90 million Windows users also use
    facebook. This could be 90 million out of 90 million, or it could be 90
    million out of 90 trillion, for all the numbers we have tell us. We have
    a value - 90 million, or 90% - but nothing to compare it to.

    Hmm. We do have certain stats put out by MS on the sales of Windows, if
    nothing else. Suppose they say there are approximately 600 million
    Windows machines out there. This means our 90 million Windows-using
    visitors represent about 15% of the Windows-using population.

    Note, however, the problem: in order to come up with that 15%, we had to
    *already know* how many Windows users there are. We had absolutely no
    way to determine, from the web stats we have, how many Windows users
    there are; we can only determine how many Windows users _we get here_
    and, of our users, what percentage use Windows.

    So let's try it now for Linux. We know 5% of our visitors, or about 5
    million visitors, use Linux. How many people _in total_ use Linux?

    To answer that, we need to know... oh, right, we need to know how many
    people use Linux. Just as we needed to know for Windows. So, in order
    to find out, from the web stats, how many people use Linux, all we need
    to know first is how many people use Linux - but if we know that, then
    our web stats become irrelevant.

    The only other way to do this is to figure out how popular Facebook is
    with Linux users, then multiply out. If 10% of Linux users also use
    Facebook, then the 5 million Linux-based Facebook users represent some 50
    million Linux users.

    And how do we tell how popular Facebook is with Linux users? Right: we
    compare the number of Linux users using Facebook to the total population
    of Linux users, meaning we need to already know the total population of
    Linux users, thus making the web stats irrelevant.

    Thus the only way we can even hope to use these stats to determine Linux
    usage is to simply assert, as a "truth" fabricated out of whole cloth,
    that Linux users use Facebook at approximtely the same rate as Windows
    users use Facebook.

    While such an assertion may seem reasonable on the face of it, it is a
    wholly unsupported assertion, one not derived from any sort of actual
    measurement or analysis, it is in fact no more valid than rolling a
    couple of ten-sided dice and asserting "that many percent" use both Linux
    and Facebook.

    Which means we're back to the original point: web stats of this sort are
    good and useful to the site, insofar as they tell Facebook or Google or
    whomever which browsers, which OSen and so forth they should focus on due
    to their particular user base, but those same stats are absolutely
    worthless in trying to figure out how many Linux users, in toto, are out
    there in the wilds.

    So, any of you boneheads who want to tell us that web stats are useful in
    determining Linux usage, you feel free. Just remember to tell us how you
    calculated the total Linux usage beforehand, and why you bothered dicking
    around with pointless web stats to arrive at an answer you already have
    to have to make your web stats mean anything relevant.


  10. Kelsey Bjarnason can't do simple algebra (was: How many "advocates" have me killfiled?)


    "Kelsey Bjarnason" wrote in message
    news:r20jh5-q9o.ln1@spankywork.localhost.net...
    > On Tue, 03 Jun 2008 19:04:09 -0700, cc wrote:
    >
    > You're not though. You can snip the arguments all you want, but it
    > doesn't make you look any smarter.
    >
    >
    > Well, let's see if I can't simplify this enough for your pea brain to
    > cope with.
    >
    > A site - we're using Facebook as a hypothetical example - has the
    > following stats:
    >
    > 100 million visitors
    > 90% Windows
    > 5% Linux
    > 5% Other
    >
    > How do we determine, from this, how many Linux users there are - not just
    > Linux users of Facebook, but *total* Linux users (just counting
    > "desktop", not server, usage)?


    For starters... why is an "absolute" number required when a "percentage"
    (relative) value is sufficient?



    > Hmm. We do have certain stats put out by MS on the sales of Windows, if
    > nothing else. Suppose they say there are approximately 600 million
    > Windows machines out there.


    Suppose you pulled this 600 million Windows machines number out of your ass.
    How exactly did you conclude there are this many Windows machines?



    > This means our 90 million Windows-using
    > visitors represent about 15% of the Windows-using population.
    >
    > Note, however, the problem: in order to come up with that 15%, we had to
    > *already know* how many Windows users there are. We had absolutely no
    > way to determine, from the web stats we have, how many Windows users
    > there are; we can only determine how many Windows users _we get here_
    > and, of our users, what percentage use Windows.
    >
    > So let's try it now for Linux. We know 5% of our visitors, or about 5
    > million visitors, use Linux. How many people _in total_ use Linux?


    Dream on with your 5% linux number. But for the sake of argument let's say
    it's true. Obviously you're a math IDIOT because this is easily calculated.

    If both Windows and linux users are just as likely to visit a site and if
    you know that there are 600-million windows users and they account for 90%
    of the visitors then it's simple algebra to solve for the number of linux
    users who account for 5% of the hits.

    When you get to 6th grade you'll learn this thing called "algebra" and you
    too will be able to solve equations with a single unknown variable.



    > To answer that, we need to know... oh, right, we need to know how many
    > people use Linux. Just as we needed to know for Windows. So, in order
    > to find out, from the web stats, how many people use Linux, all we need
    > to know first is how many people use Linux - but if we know that, then
    > our web stats become irrelevant.
    >
    > The only other way to do this is to figure out how popular Facebook is
    > with Linux users, then multiply out. If 10% of Linux users also use
    > Facebook, then the 5 million Linux-based Facebook users represent some 50
    > million Linux users.
    >
    > And how do we tell how popular Facebook is with Linux users? Right: we
    > compare the number of Linux users using Facebook to the total population
    > of Linux users, meaning we need to already know the total population of
    > Linux users, thus making the web stats irrelevant.
    >
    > Thus the only way we can even hope to use these stats to determine Linux
    > usage is to simply assert, as a "truth" fabricated out of whole cloth,
    > that Linux users use Facebook at approximtely the same rate as Windows
    > users use Facebook.
    >
    > While such an assertion may seem reasonable on the face of it, it is a
    > wholly unsupported assertion, one not derived from any sort of actual
    > measurement or analysis, it is in fact no more valid than rolling a
    > couple of ten-sided dice and asserting "that many percent" use both Linux
    > and Facebook.
    >
    > Which means we're back to the original point: web stats of this sort are
    > good and useful to the site, insofar as they tell Facebook or Google or
    > whomever which browsers, which OSen and so forth they should focus on due
    > to their particular user base, but those same stats are absolutely
    > worthless in trying to figure out how many Linux users, in toto, are out
    > there in the wilds.
    >
    > So, any of you boneheads who want to tell us that web stats are useful in
    > determining Linux usage, you feel free. Just remember to tell us how you
    > calculated the total Linux usage beforehand, and why you bothered dicking
    > around with pointless web stats to arrive at an answer you already have
    > to have to make your web stats mean anything relevant.
    >



    ** Posted from http://www.teranews.com **

  11. Re: How many "advocates" have me killfiled?

    In article <59uih5-q9o.ln1@spankywork.localhost.net>,
    Kelsey Bjarnason wrote:
    >
    > Let me explain this to you in terms even you might be able to understand:
    >
    > The goal is to determine the number of Linux users from web stats.


    No, the goal is to determine the percentage of Linux users among the
    general computer using population, not the absolute number.

    Since you seem to have overlooked this, I'll repeat it:

    Let

    Lf = number of Linux users who use Facebook
    L = number of Linux users
    Wf = number of Windows users who use Facebook
    W = number of Windows users

    (To simplify things, we'll ignore Mac users).

    If Facebook usage isn't correlated with OS usage, then we'd expect

    Lf/L = Wf/W (1)

    The fraction of Linux users in the general computer using population is

    L / (L + W) = 1 / (1 + W/L) (2)

    and the fraction of Linux users in the Facebook population is

    Lf / (Lf + Wf) = 1 / (1 + Wf/Lf) (3).

    But from (1), we have W/L = Wf/Lf, and so (2) and (3) are equal.

    If Lf/L > Wf/W, then W/L > Wf/Lf, and (2) < (3). (Facebook has a higher
    fraction of Linux users than the general population does).

    If Lf/L < Wf/W, then W/L < Wf/Lf, and (2) > (3). (Facebook has a lower
    fraction of Linux users than the general population does).

    Do you have any reason to believe that Facebook use is correlated with
    OS use, and so Lf/L != Wf/W?

    --
    --Tim Smith

  12. Re: How many "advocates" have me killfiled?

    In article ,
    Kelsey Bjarnason wrote:
    > Thus the only way we can even hope to use these stats to determine Linux
    > usage is to simply assert, as a "truth" fabricated out of whole cloth,
    > that Linux users use Facebook at approximtely the same rate as Windows
    > users use Facebook.
    >
    > While such an assertion may seem reasonable on the face of it, it is a
    > wholly unsupported assertion, one not derived from any sort of actual
    > measurement or analysis, it is in fact no more valid than rolling a
    > couple of ten-sided dice and asserting "that many percent" use both Linux
    > and Facebook.


    It's not just Facebook. It's Google, and the BBC news site, and many
    many others. When you have just one site with low Linux representation,
    there might be some unapparent reason that it attracts Linux users at a
    rate lower than it attracts Windows or Mac users. But when this is an
    across the board phenomenon, it is hard to blame that on some hidden
    bias.

    --
    --Tim Smith

  13. Re: How many "advocates" have me killfiled?

    On Jun 4, 1:45*pm, Kelsey Bjarnason wrote:
    > On Tue, 03 Jun 2008 19:04:09 -0700, cc wrote:
    >
    > You're not though. You can snip the arguments all you want, but it
    > doesn't make you look any smarter.
    >
    > Well, let's see if I can't simplify this enough for your pea brain to
    > cope with.
    >
    > A site - we're using Facebook as a hypothetical example - has the
    > following stats:
    >
    > 100 million visitors
    > 90% Windows
    > 5% Linux
    > 5% Other
    >
    > How do we determine, from this, how many Linux users there are - not just
    > Linux users of Facebook, but *total* Linux users (just counting
    > "desktop", not server, usage)?
    >



    You don't determine total Linux users. That's obvious. And that isn't
    the argument. The argument is does the 5% reflect the percentage of
    Linux users you would have out of X number of users. X can be 10, X
    can be the total number of computers in the world. X doesn't matter.
    The 5% matters. Because every day X changes. X yesterday is X +
    1000000 today. But the 5% isn't changing.

    And so the answer to the real question is yes, the 5% does reflect
    percentage of Linux users, because as you've already said, there is no
    reason why using Linux would keep you from going to facebook, and no
    reason why using windows would make you more likely to go. I
    personally don't like the facebook example, so lets use google instead
    from now on.

    > So, any of you boneheads who want to tell us that web stats are useful in
    > determining Linux usage, you feel free. *Just remember to tell us how you
    > calculated the total Linux usage beforehand, and why you bothered dicking
    > around with pointless web stats to arrive at an answer you already have
    > to have to make your web stats mean anything relevant.


    Wow. You've got it completely backwards. Web stats are relevant
    because we can't know the exact total of Linux users. There is a
    unknown total of computer users in the world, of which about 5% (in
    your hypothetical example, about 2% in the real stats) are Linux
    users. Exact total of computer users is irrelevant just like exact
    total of Linux users. I posted this before in this thread, albeit in
    less detail, and you snipped it. You're going to have to explain why
    the total number Linux users changes what the 5% means. Here's a hint,
    it doesn't.

  14. Re: How many "advocates" have me killfiled?

    * Tim Smith peremptorily fired off this memo:

    > In article ,
    > Kelsey Bjarnason wrote:
    >> Thus the only way we can even hope to use these stats to determine Linux
    >> usage is to simply assert, as a "truth" fabricated out of whole cloth,
    >> that Linux users use Facebook at approximtely the same rate as Windows
    >> users use Facebook.
    >>
    >> While such an assertion may seem reasonable on the face of it, it is a
    >> wholly unsupported assertion, one not derived from any sort of actual
    >> measurement or analysis, it is in fact no more valid than rolling a
    >> couple of ten-sided dice and asserting "that many percent" use both Linux
    >> and Facebook.

    >
    > It's not just Facebook. It's Google, and the BBC news site, and many
    > many others. When you have just one site with low Linux representation,
    > there might be some unapparent reason that it attracts Linux users at a
    > rate lower than it attracts Windows or Mac users. But when this is an
    > across the board phenomenon, it is hard to blame that on some hidden
    > bias.


    Of course, you don't get one consistent number, you get a fairly wide
    range. Only by making an assumption can you pin down the true "linux
    percentage" based on a collection of web stats. One such assumption is
    "all OS users visit sites without regard to the operating system they
    use."

    This, of course, is somewhat false, since many sites discourage visits
    by non-Windows systems.

    --

  15. Re: How many "advocates" have me killfiled?

    Linonut wrote:


    >
    > This, of course, is somewhat false, since many sites discourage visits
    > by non-Windows systems.




    Let's see a long list of URLs. I'll start:

    windowsupdate.microsoft.com


    Your turn.


  16. Re: How many "advocates" have me killfiled?


    "Frank" wrote in message
    news:g26uhk$6c3$1@registered.motzarella.org...
    > Linonut wrote:
    >
    >
    >>
    >> This, of course, is somewhat false, since many sites discourage visits
    >> by non-Windows systems.

    >
    >
    >
    > Let's see a long list of URLs. I'll start:
    >
    > windowsupdate.microsoft.com
    >
    >
    > Your turn.
    >



    That's pretty much the extent of it. It's another one of those canned
    excuses that these guys use every time linux comes up short. Other canned
    excuses ready to be deployed at a moments notice are:

    Monopoly effects
    Heard mentality
    Microsoft thugs force XYZ to do ABC
    Bribery and corruption
    You can't measure linux
    People change their browser strings to spoof IE


    These are very lame excuses.





    ** Posted from http://www.teranews.com **

  17. Re: How many "advocates" have me killfiled?

    In comp.os.linux.advocacy, Linonut

    wrote
    on Wed, 4 Jun 2008 15:57:16 -0400
    :
    > * Tim Smith peremptorily fired off this memo:
    >
    >> In article ,
    >> Kelsey Bjarnason wrote:
    >>> Thus the only way we can even hope to use these stats to determine Linux
    >>> usage is to simply assert, as a "truth" fabricated out of whole cloth,
    >>> that Linux users use Facebook at approximtely the same rate as Windows
    >>> users use Facebook.
    >>>
    >>> While such an assertion may seem reasonable on the face of it, it is a
    >>> wholly unsupported assertion, one not derived from any sort of actual
    >>> measurement or analysis, it is in fact no more valid than rolling a
    >>> couple of ten-sided dice and asserting "that many percent" use both Linux
    >>> and Facebook.

    >>
    >> It's not just Facebook. It's Google, and the BBC news site, and many
    >> many others. When you have just one site with low Linux representation,
    >> there might be some unapparent reason that it attracts Linux users at a
    >> rate lower than it attracts Windows or Mac users. But when this is an
    >> across the board phenomenon, it is hard to blame that on some hidden
    >> bias.

    >
    > Of course, you don't get one consistent number, you get a fairly wide
    > range. Only by making an assumption can you pin down the true "linux
    > percentage" based on a collection of web stats. One such assumption is
    > "all OS users visit sites without regard to the operating system they
    > use."
    >
    > This, of course, is somewhat false, since many sites discourage visits
    > by non-Windows systems.
    >


    Not only that, but there's the usual issues regarding
    multiboots. I'm setting up my laptop now, in fact, to boot
    a number of distros (not all of them Linux-related), mostly
    because I want to. (Regrettably, ReactOS won't be one
    of them; it's not ready to install on extended partitions
    yet for some reason. I can install ReactOS on a phantom
    machine, and it gives me a rather basic desktop. I don't
    know about its networking yet.)

    So, if I were to natively install Gentoo (which I've
    already done), Fedora, openSuSE, Sabayon, Debian, Ubuntu,
    Debian/HURD or Gentoo/HURD, FreeDOS, XP (which came with
    the machine), ReactOS, LinuxFromScratch, and FreeBSD,
    either natively or in QEMU disk images on this box, how
    much market share do I influence? ;-)

    --
    #191, ewill3@earthlink.net
    Useless C++ Programming Idea #12995733:
    bool f(bool g, bool h) { if(g) h = true; else h = false; return h;}
    ** Posted from http://www.teranews.com **

  18. Re: How many "advocates" have me killfiled?

    On Jun 4, 5:06*pm, The Ghost In The Machine
    wrote:
    > In comp.os.linux.advocacy, Linonut
    >
    > *wrote
    > on Wed, 4 Jun 2008 15:57:16 -0400
    > :
    >
    >
    >
    >
    >
    > > * Tim Smith peremptorily fired off this memo:

    >
    > >> In article ,
    > >> *Kelsey Bjarnason wrote:
    > >>> Thus the only way we can even hope to use these stats to determine Linux
    > >>> usage is to simply assert, as a "truth" fabricated out of whole cloth,
    > >>> that Linux users use Facebook at approximtely the same rate as Windows
    > >>> users use Facebook.

    >
    > >>> While such an assertion may seem reasonable on the face of it, it is a
    > >>> wholly unsupported assertion, one not derived from any sort of actual
    > >>> measurement or analysis, it is in fact no more valid than rolling a
    > >>> couple of ten-sided dice and asserting "that many percent" use both Linux
    > >>> and Facebook.

    >
    > >> It's not just Facebook. *It's Google, and the BBC news site, and many
    > >> many others. *When you have just one site with low Linux representation,
    > >> there might be some unapparent reason that it attracts Linux users at a
    > >> rate lower than it attracts Windows or Mac users. *But when this is an
    > >> across the board phenomenon, it is hard to blame that on some hidden
    > >> bias.

    >
    > > Of course, you don't get one consistent number, you get a fairly wide
    > > range. *Only by making an assumption can you pin down the true "linux
    > > percentage" based on a collection of web stats. *One such assumption is
    > > "all OS users visit sites without regard to the operating system they
    > > use."

    >
    > > This, of course, is somewhat false, since many sites discourage visits
    > > by non-Windows systems.

    >
    > Not only that, but there's the usual issues regarding
    > multiboots. *I'm setting up my laptop now, in fact, to boot
    > a number of distros (not all of them Linux-related), mostly
    > because I want to. *(Regrettably, ReactOS won't be one
    > of them; it's not ready to install on extended partitions
    > yet for some reason. *I can install ReactOS on a phantom
    > machine, and it gives me a rather basic desktop. *I don't
    > know about its networking yet.)
    >
    > So, if I were to natively install Gentoo (which I've
    > already done), Fedora, openSuSE, Sabayon, Debian, Ubuntu,
    > Debian/HURD or Gentoo/HURD, FreeDOS, XP (which came with
    > the machine), ReactOS, LinuxFromScratch, and FreeBSD,
    > either natively or in QEMU disk images on this box, how
    > much market share do I influence? ;-)
    >


    I'm assuming you do some web browsing from all of them, so why
    wouldn't you show up for all of them? You'd count more in Linux's
    favor, but you're using more than one Linux distro. So you'd influence
    exactly how much market share you'd think you would. As in, count one
    per distro. What's so hard about that situation?

  19. Re: How many "advocates" have me killfiled?

    On Wed, 04 Jun 2008 11:52:04 -0700, Tim Smith wrote:

    > In article <59uih5-q9o.ln1@spankywork.localhost.net>,
    > Kelsey Bjarnason wrote:
    >>
    >> Let me explain this to you in terms even you might be able to
    >> understand:
    >>
    >> The goal is to determine the number of Linux users from web stats.

    >
    > No, the goal is to determine the percentage of Linux users among the
    > general computer using population, not the absolute number.


    It's the same ****ing number, you moron.

    If Linux is used by 15% of a user population of 100 million people, it is
    used by 15 million users.

    So what is the total Linux user population? 15 million.

    Try to plug your brain in before posting.

    BTW, I snipped your "math" which *still* rests on the unfounded
    assumption that a given site will be as popular with Linux users as with
    Windows users. See if you can check the stats on the Windows update site
    and figure out why your assumption is almost as stupid as you are.


  20. Re: How many "advocates" have me killfiled?

    [snips]

    On Tue, 03 Jun 2008 21:59:22 -0400, Linonut wrote:

    > Personally, since Linux is still somewhat of a techie system, I would
    > suspect that most Facebook users are Windowsers.


    Bingo. This is the point the boneheads persist in overlooking.

    One analysis works as follows:

    There are an estimated 600 million Windows machines.
    There are some 100 million Facebook users.
    90% of those facebook users run Windows.

    Thus, some 15% of Windows users, total, use Facebook.

    Using this value in conjunction with the (hypothetical) fact that 5% of
    Facebook users run Linux, we crank the numbers:

    5% of facebook users = 5 million people using both Facebook and Linux.
    15% of total users also use Facebook, therefore there are some 33 million
    Linux users.

    Which all makes perfect sense until you stop and realize it is based
    entirely on one completely unfounded assumption: that Linux users use the
    same sites, share the same interests, in about the same proportion as
    Windows users do.

    The most glaringly obvious refutation of this is to point out that if the
    interests of Linux users matched those of Windows users, they wouldn't
    *BE* Linux users.

    Simply choosing to use Linux already means you have a divergent interest
    set. How divergent cannot be determined short of polling the users.

    So how does one determine Linux usage from such information? You can't.
    You can determine, to some accuracy, how many Linux users _use that
    site_. You haven't a snowball's chance in Hell of using this to
    determine general Linux usage, as you have no way to figure what
    percentage of Linux users this number represents.

    For some reason, this doesn't sink in with them. Nor does the next
    obvious point: that the only way to determine that percentage is to start
    by figuring out how many people are using Linux, at which point you no
    longer need the web stats to tell you this.


+ Reply to Thread
Page 4 of 9 FirstFirst ... 2 3 4 5 6 ... LastLast