On Wed, Oct 14, 2009 at 6:22 PM, Christoph Wickert
<christoph.wickert@googlemail.com> wrote:
Am Mittwoch, den 14.10.2009, 17:35 +0200 schrieb Gregory Zysk:
> There is a problem with measuring downloads though, it does not give
> us a realistic number of how many people actually use Fedora.
We do have smolt data for that:
http://smolt.fedoraproject.org/static/stats/stats.html
And you think distrowatch is more representative? I don't think so.
>>> It is representative for downloads, or at least some of them.
Distrowatch is counting visitors by user agent strings. Ask yourself:
Who visits distrowatch?
>>>I did!! That is how I cam to using Fedora. That is how everyone I know has referred me to and the same way they came to using FOSS.
Likely somebody who is looking for a new distro
to try. So are these people really happy with their distro or even
convinced of it?
>>>This is accomplished by separate surveys or sampling techniques.
This is something the numbers cannot show.
>>>Once a methodology and research design is set in place. These numbers will show exactly what you have created them to be.
By counting uniqe visits, you wont get numbers of OS installations or
users, because you don't know how often people visit that site. There
are people who say the number of Fedora installs is actually larger than
the number of Ubuntu installations, but quite a lot of them are servers
running in data centers [1]. They will never show up at distrowatch
because servers usually don't visit that site. ;)
>>>This from an objective point of view is just opinion. You ask an Ubuntu user and they will say they have the most. You ask a person who works for microsoft who has the best OS? Don't be surprised if they say: microsoft.
Regards,
Christoph
[1] Just look at the high number of virtual machines in smolt.
--