Regarding Get Fedora page
duffy at fedoraproject.org
Tue Jun 1 01:39:51 UTC 2010
On Tue, 2010-06-01 at 09:19 +0800, Ed Greshko wrote:
> I don't buy into your survey..and the fact that you have concluded that
> only 19% of current users make use of torrents. It is true that 19% of
> the people that knew about your survey, and that chose to respond, use
> torrents. Nothing other than that can be inferred.
Surveys are not 100% reflective of reality nor does any reasonable
person claim them to be so. That being said, they certainly have the
potential to be a lot more objective a measure of reality than any
single person's opinion or guess.
It's a nice logical tactic to disprove the utility of surveys completely
and to attempt to discount their results completely. However, arguing
over whether or not surveys are a valid object measure is really not an
appropriate topic for this list. We must make the assumption here that
the many years of research, money, and effort that have gone into
constructing and developing tools, processes, and methods of devising
surveys, running surveys, and analysing the results of surveys were not
for naught. If you need any single example of the inherently flawed
nature of surveys still somehow rendering results that reflect reality,
look at the US presidential vote exit surveys over the past few
presidential elections and how accurately they reflected the actual
80-something people is enough participants in a survey to assume a
reasonable margin of error so don't go making excuses that enough people
weren't surveyed either.
So let's make the assumption that surveys are a useful tool and leave it
at that. If you are uncomfortable with that assumption then we can't
engage in this discussion together, I'm sorry.
> Are you saying that 19% is an insignificant portion of the community?
> What percentage must it be to garner consideration? What is the
> threshold? What is the margin of error in the survey?
If I told you the margin of error of the survey (not difficult to
calculate) would you even understand how to interpret it?
Did you read the survey analysis and see that the percentages were
additive? Did you notice that the numbers in the piechart don't add up
to 100%? This is because users were able to select more than one
download method. So if someone used both torrents and mirrors, they were
counted in the 19% of torrent users.
> While this is purely anecdotal..... My dad is 84 years old. He had
> been a long time MS-Windows user. I got tired of trying to fix his
> Windows problems via long distance. Last year we talked about it and he
> agree to move to Linux. One of the first questions he asked me was "can
> I download it with bittorrent?". That is probably due to his addiction
> to "The Pirate Bay". The point is, torrents are not only for the highly
Anecdotal evidence about your 84-year-old father is not an appropriate
response to quantitative and objective data. If you are willing to step
outside of your immediate surroundings and gather some actual data to
inform the design we would obviously be more than welcome to consider it
in our design processes in the future.
> But, isn't it time to let this go? The Get Fedora page has now been
> modified. One can now find the torrents from that page. It may be a
> bit more obscured than I would have it...but it is there.
You're the one who continued to post to the thread...
More information about the users