On Thu, 28 Sep 2006, Oliver Fromme wrote:
Marc G. Fournier wrote:
> Oliver Fromme wrote:
> > By the way, I've got a small question. Does the database
> > throw all entries away at the end of each month, and start
> > all over again with zero entries? Or is each entry expired
> > after a certain time has elapsed (31 days or whatever)?
>
> Neither ... the month that the report was submitted for has one entry for
> host ... we'll be able to graph stuff like growht in # of reporting hosts
> and such ...
I'm not sure that will work well ... Lets see what happens
when October begins. At the beginning of September, the
statistics were all reset to zero.
That is correct ... this is one of my hosts:
id | unique_key | country_code | first_connect
-----+-----------------------------------+--------------+----------------------------
235 | cdd6a5011ac543400bcaafb413ae577d | PA | 2006-08-15
06:53:50.077533
And this is what its reported for so far:
id | operating_system | release | architecture | report_month
-----+------------------+-------------+--------------+----------------------------
235 | FreeBSD | 4.11-STABLE | i386 | 2006-08-15
07:08:50.694954
235 | FreeBSD | 4.11-STABLE | i386 | 2006-09-01
08:30:06.198988
(2 rows)
One for August, one for September ... start of October, a third entry will
be added ...
Sorry, I was a bit unclear ... I didn't mean to say that you shouldn't
collect the numbers for each OS sub-variant (or lets call it
"distribution") separately. But I think it would make sense to group
them together for the "Big 4" on the front page (main page) at
bsdstats.org.
Except those that are working hard on the various 'sub-variants' are proud
to see the fact that the work they are doing is being used, plus, it helps
to advertise those sub-variants, so that ppl know they are out there ...
By the way, I think you cannot tell much from those numbers, because
they don't really show any real-world usage of the various BSD variants.
Not now, and not in a year from now. The reason for that is that the
different BSD projects have very different policies for enabling the
bsdstats script by default during installation or during update.
The thing is, we're not looking at producing a "this is all the BSD hosts
that are out there" sort of number ... we are trying to produce a "see,
there *is* a BSD market" ... what does some place like Adaptec consider a
"market", that I don't know ... 100k hosts? 500k hosts? I don't know
that, each manufacturer (both software and hardware) will have different
thresholds ... the point is that we now have *some* marketing numbers that
aren't "purely guesswork" ...
Hell, even looking at the numbers now ... 10,328 hosts, 42.2% of which are
OpenBSD ... 4 362 hosts doesn't sound like alot, but those are 4 362 hosts
that will *never* see an Adaptec controller because of Adaptec's
closed-doc policy ... what's the average price of an Adaptec controller
nowadays? Looking at there site, their SAS RAID controller is SRP: $995
... that is $4 340 190 in potential revenu *if* everyone bought that card
... even if average price was $100, that is $436 200 in potential revenue
that can't be tap'd ... and that number is probably not even 1/10th of the
actual # of hosts out there ...
And ya, I know, not everyone would by Adaptec even if they had open docs
... that isn't the point ... the point is that Adaptec is getting *zero*
right now from the OpenBSD market, since they are closed docs ...
By the way (apropos default policies): Guess why OpenBSD has gotten
ahead of FreeBSD in the statistics? It has been growing at a much higer
rate all the time, and will continue to do so. Soon the statistics will
"prove" that OpenBSD's user base is ten times larger than FreeBSD's,
because we won't have a bsdstats option in sysinstall in 6.2-Release.
I'd be willing to submit a patch (I'm somewhat familiar with the
sysinstall code), but I assume it's too late because we're already in
code freeze, and sysinstall is a particularly critical piece of code.
Apart from that, such a patch will probably be shredded to pieces by
bike shed discussions.
Of course, not submitting the patch ASAP will ensure that not only do it
not getting into 6.2-RELEASE, but it won't get into subsequent releases,
or -CURRENT, or ... :)
*sigh* I'm sorry, what I wrote isn't really constructive, but rather
bellyaching about the whole situation. Maybe I should better shut up
now. :-)
I got tired of bellyaching, and created bsdstats.org ... *shrug*
----
Marc G. Fournier Hub.Org Networking Services (http://www.hub.org)
Email . [EMAIL PROTECTED] MSN . [EMAIL PROTECTED]
Yahoo . yscrappy Skype: hub.org ICQ . 7615664
_______________________________________________
freebsd-advocacy@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-advocacy
To unsubscribe, send any mail to "[EMAIL PROTECTED]"