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

Re: Is information on the number of downloads available?



On Fri, Feb 11, 2005 at 09:26:47PM +0100, Matthias Kilian wrote:
| On Fri, Feb 11, 2005 at 08:41:11PM +0100, Marco Feenstra wrote:
| > Maybe a better count can be created by scanning the internet for active
| > OpenBSD systems every so often, maybe www.netcraft.com can be used to
| > find the amount of installed OpenBSD systems
|
| Netcraft doesn't differentiate NetBSD from OpenBSD, so you'll have to
| estimate the Net/OpenBSD ratio.
|
| Another problem is that you can't estimate the number of installed
| systems from the number of running *servers*. Lots of people seem
| to use BSD on firewalls and web servers but Linux on the desktop.
|
| And if it comes to bug detection (a topic on which the OP seems to
| investigate), desktop usage will lead to a completely other bug
| type pattern and detection ratio than web servers and firewalls.

I think the basic idea is that with a growing number of downloads/CD
purchases/CVS checkouts/typewriter-copies-sent-by-snailmail/whatever
there is a growing number of uses.

So we do not know what the ratio is between number of downloads and
number of users (or number of installed systems), but we can assume
that this ratio is the same from release to release.

Whether this is a valid assumption is not the point. You make a model
(simplified version of the truth) of something and for this model you
make assumptions. As long as these assumptions are documented
(preferably with an (estimated) errormargin) you're doing something
that may resemble valid statistics.

Cheers,

Paul 'WEiRD' de Weerd

--
>++++++++[<++++++++++>-]<+++++++.>+++[<------>-]<.>+++[<+
+++++++++++>-]<.>++[<------------>-]<+.--------------.[-]
                 http://www.weirdnet.nl/

[demime 1.01d removed an attachment of type application/pgp-signature]