[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

InfoWorld Cooks Data so Microsoft will win

  If you check out the following URL:
  You will find that InfoWorld magazine recently ran benchmarks for an Applix
  Java app on various platforms. The OS's tested were Windows NT, Windows 95,
  Windows 3.1, OS/2, Sun Solaris, MacOS, and Caldera OpenLinux. If one just
  looks at the bottom line, NT won, with a speed of :37, squeaking past
  Linux, buried at the end of the list, with :39. But if you look at the
  details, which the executives won't, you'll find that NT was tested on a
  Pentium 180 with 64 Megs of RAM against Linux on a Pentium 166 with 48
  Megs. Given how neck-and-neck the test came out, it is almost certain that
  Linux would have won in a legitimate benchmark, i.e., one performed on the
  same hardware.
  InfoWorld may point out that some of the OS's they tested, such as Mac, had
  to be run on different hardware, and that others, such as Solaris, were
  best tested elsewhere, even though they are available for Intel.  Neither
  of these comments applies to Linux, which, like Windows, was developed and
  is still most widely deployed on Intel. InfoWorld may also protest that it
  was testing Java portability, not benchmarking OS's. This objection is
  belied by the fact that the introduction to the article chiefly compares
  the performance of the OS's. Personally, I think a public issue should be
  made of this, as it can undermine InfoWorld's credibility as an objective
  source of technical information, which they can ill afford.  More
  importantly, if this article gets attention, it can serve to cause greater
  questioning of trade press biases generally, and bring Linux forward in the
  public mind as an alternative to Windows.