Re: Survey answers part 3: systemd is not portable and what this means for our ports
On Fri, Jul 19, 2013 at 8:48 PM, John Paul Adrian Glaubitz
<glaubitz@physik.fu-berlin.de> wrote:
> On 07/19/2013 07:47 PM, Scott Kitterman wrote:
>> John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> wrote:
>>>
>>> On 07/19/2013 06:57 PM, Scott Kitterman wrote:
>>>>
>>>> sysvinit 148865 99.83%
>>>
>>>
>>> The reason might be that systemd does not conflict with sysvinit :).
>>
>>
>> So are we playing word games now or trying to solve a problem? According
>> to the popcon data, neither systemd nor upstart have enough deployment to be
>> considered anything other than essentially zero.
>
>
> I don't understand why anyone would assume that the popcon data in this
> context is not accurate.
>
> Again, sysvinit is essential, systemd is not and it doesn't have any
> reverse dependencies. Thus, every counted systemd installation was
> actually actively performed.
Oh, really? Because its the second time you claim that?
apt-cache rdepends systemd --important --no-conflicts --no-breaks -o
APT::Cache::ShowDependencyType=1
Granted, the rdependencies in stable/testing/unstable are all alternatives
in the second position which can be easily solved otherwise.
But we have also gnome-settings-daemon in experimental which depends without
an alternative on it. Now, if you look really close on the popcon data for
systemd you see that in March 2013 there is a plateau reached for systemd,
which "suddenly" at the end of the month is followed by the constant up –
which neatly alines with the upload of the previously mentioned
gnome-settings-daemon to experimental with a systemd dependency.
So, what do you think: More people wanting to testdrive GNOME 3.8 or systemd?
Oh, and while you have the graph open: Do you see the bump in the graph at
the end of May? Lets guess when the systemd survey was… so we can assume
that "a lot" people installed systemd to test it for the survey and it
worked so "great" that they not only not continued to use it, but have also
actively deinstalled it again – even though you don't have to, like for
upstart which doesn't seem to have obvious bumps (beside the ubuntu
misreporting) so everyone installing it sticked with it…
I will leave the obvious conclusion as an exercise for the reader.
Of course, both analysis are obviously flawed as this popcon data can't
really be interpreted that way as its an apple to banana comparison and
way too few datapoints, but everyone likes misinterpret statistics as
"proven" by this thread – and statistics say that I am a pro-faker!
"I only believe in statistics that I doctored myself."
-- Winston Churchill
Best regards
David Kalnischkies
P.S.: Everyone who is now trying to disprove my "facts" has missed the point.
Reply to: