|
Authored by: Gringo_ on Monday, September 17 2012 @ 08:02 PM EDT |
Without the source, it is impossible for anybody to assess the accuracy
of the study you refer to. I find it hard to believe anybody outside of
microsoft could make such a determination, because where would they
get accurate Microsoft data? I doubt I would believe any such report
emanating from Microsoft itself. Remember their "Get the facts" fud?[ Reply to This | Parent | # ]
|
|
Authored by: tknarr on Monday, September 17 2012 @ 09:08 PM EDT |
Part of it's that a lot of the "bugs" in Microsoft software aren't
technically bugs, they're deliberate design features. Eg., IE ignoring the
declared content type of a file and examining it's contents to determine how to
handle it, which allows malware to bypass normal checks and slip through. Or IE
having system privileges that need protected from access by malware.
It's
like the telnet protocol. The fact that an eavesdropper can sniff your
unencrypted password out of the data stream is a major problem, sufficient to
justify disabling the protocol entirely, but it's not a bug because the
protocol's working exactly as designed. Or the rsh protocol, which depends on
easily-forged UIDs to establish trust and authentication. In both cases the
problem isn't a bug in the implementation, it's a fundamental flaw in the design
that can only be fixed by throwing out the entire design and choosing a
different one (which is what they did to give us the ssh protocol that's
replaced rsh and telnet for most purposes). [ Reply to This | Parent | # ]
|
|
Authored by: ailuromancy on Tuesday, September 18 2012 @ 01:53 AM EDT |
Windows requires a pile of third party device drivers
but Linux device
drivers are usually included in the
kernel source. Windows runs on AMD64 X86
and nearly one
variant of ARM. Linux has code for many more
architectures. For
a better comparison, select a
common architecture and set of device drivers,
then
count the number of lines of code that actually get
compiled.
At a
wild guess, there will be more lines of code
in Windows. Assuming that but code
sets have about the
same number of bugs per thousand lines of code, that
implies more bugs in Windows.
[ Reply to This | Parent | # ]
|
|
|
|
|