Re: Are popcon stats per package and arch possible?
On Sat, May 05, 2007 at 10:40:34AM +0200, Petter Reinholdtsen wrote: > > [Michael Hanke] > > I wonder whether it is possible to get information about which > > package is installed/used on a particular architecture. > > It is possible, but it isn't done at the moment. We do not store the > data needed to generate such reports, but it could be done with the > data set we have at every given point in time. > > But I would be reluctant to do it for the less used architectures, for > privacy reasons. The information we get from such analysis would be > too easy to associate with individual users (like the kfreebsd-amd64 > porters for that arch. :). I fully agree. However, this should not prevent interested Debian developers can still log on gluck and look in /org/popcon.debian.org/popcon-mail/popcon-entries/ and make theirs own stats and maybe publish a summary as long as they keep privacy in mind. But publishing the data daily through popcon.debian.org is a totally different issue privacy-wise. Cheers, -- Bill. <[EMAIL PROTECTED]> Imagine a large red swirl here. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Are popcon stats per package and arch possible?
Steve Langasek <[EMAIL PROTECTED]> writes: > On Sun, May 06, 2007 at 10:15:40PM +0200, Goswin von Brederlow wrote: >> Steve Langasek <[EMAIL PROTECTED]> writes: > >> > On Sat, May 05, 2007 at 06:23:36PM +0200, Petter Reinholdtsen wrote: > >> >> [Michael Hanke] >> >> > To me it looks like stats for the major architectures up to (and >> >> > including) powerpc are ok wrt privacy concerns. Do you agree? > >> >> I'm not sure if that would be the correct cutoff point, or if only >> >> amd64 and i386 have enough submissions to ignore the privacy issue. > >> > Well, at that point what use is a per-arch stat anyway? > >> So porters can see what package are most usefull for that >> architecture. > > Ok, WTF. How have four different people missed the point here that *being > able to get per-arch per-package statistics only for i386 and amd64 does not > give meaningful per-arch information*? If you have i386, amd64, ppc and others that would still be somewhat usefull. The others would be important to keep. Otherwise I totaly agree with you. I expect i386 and amd64 to be nearly identical to each other and the total. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Are popcon stats per package and arch possible?
On Sun, May 06, 2007 at 04:10:39PM +0200, Frans Pop wrote: > > I must note the coolness of how many more 'secondary' architecture > > installations we are attaining with etch. > > Wouldn't that just be because with Etch we again offer to install popcon > during new installations? We did not do that with Sarge... Still, people who had those machines on sarge would have simply upgraded, they generally wouldn't be running the installer on the same machines all over again. -- 2. That which causes joy or happiness. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Are popcon stats per package and arch possible?
On Sun, May 06, 2007 at 10:15:40PM +0200, Goswin von Brederlow wrote: > Steve Langasek <[EMAIL PROTECTED]> writes: > > On Sat, May 05, 2007 at 06:23:36PM +0200, Petter Reinholdtsen wrote: > >> [Michael Hanke] > >> > To me it looks like stats for the major architectures up to (and > >> > including) powerpc are ok wrt privacy concerns. Do you agree? > >> I'm not sure if that would be the correct cutoff point, or if only > >> amd64 and i386 have enough submissions to ignore the privacy issue. > > Well, at that point what use is a per-arch stat anyway? > So porters can see what package are most usefull for that > architecture. Ok, WTF. How have four different people missed the point here that *being able to get per-arch per-package statistics only for i386 and amd64 does not give meaningful per-arch information*? -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Are popcon stats per package and arch possible?
Steve Langasek <[EMAIL PROTECTED]> writes: > On Sat, May 05, 2007 at 06:23:36PM +0200, Petter Reinholdtsen wrote: > >> [Michael Hanke] >> > To me it looks like stats for the major architectures up to (and >> > including) powerpc are ok wrt privacy concerns. Do you agree? > >> I'm not sure if that would be the correct cutoff point, or if only >> amd64 and i386 have enough submissions to ignore the privacy issue. > > Well, at that point what use is a per-arch stat anyway? So porters can see what package are most usefull for that architecture. The sorting of binary packages onto CDs or DVDs could be arch specific. For example on m68k that would probably remove kde/gnome from cd/dvd1 and put other stuff there that is more usefull. It could also help the struggeling ports deciding to drop some package. For example axiom is totaly useless on m68k and I bet never had a single user. As for the privacy concerns. So what if you know that 95% of all arm users run foobar. That still does not tell you who they are. There is no connection between the actual person and the data. MfG Goswin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Are popcon stats per package and arch possible?
On Sunday 06 May 2007 05:07:48 Michael Hanke wrote: > But sometimes upstream does not agree. > > Nevertheless, when they say, 'we provide binaries for Linux', they always > mean i386 Linux with everything linked statically to a huge binary blob. > > I'd really like to be able to provide some hard numbers about users of > similar packages (same field or a direct competitor) running it on > arches different from i386. Worse, it's usually *called* a "static" binary in their download list, but running ldd on it shows it actually depends on specific versions of twelve different libraries only found in an old development build of Fedora. ;) -- Wesley J. Landaker <[EMAIL PROTECTED]> OpenPGP FP: 4135 2A3B 4726 ACC5 9094 0097 F0A9 8A4C 4CD6 E3D2 pgpAPu2DIzgem.pgp Description: PGP signature
Re: Are popcon stats per package and arch possible?
On Sunday 06 May 2007 15:55, Josip Rodin wrote: > I must note the coolness of how many more 'secondary' architecture > installations we are attaining with etch. Wouldn't that just be because with Etch we again offer to install popcon during new installations? We did not do that with Sarge... pgpcd7Lp4Xq74.pgp Description: PGP signature
Re: Are popcon stats per package and arch possible?
On Sat, May 05, 2007 at 10:40:34AM +0200, Petter Reinholdtsen wrote: > On Thu, Apr 12, 2007 at 11:14:32AM +0200, Petter Reinholdtsen wrote: > > [...] > >44 0.13% hppa > >52 0.15% alpha > >53 0.15% mipsel > > 171 0.49% sparc > > 448 1.27% powerpc > > 615 1.75% arm > > [...] > > And for those interested, here are the latest numbers: > > [...] >57 0.12% hppa >65 0.14% mipsel >68 0.14% alpha > 223 0.47% sparc > 598 1.25% powerpc > 676 1.42% arm > [...] I must note the coolness of how many more 'secondary' architecture installations we are attaining with etch. Obviously their total percentages don't show as much improvement as the absolute numbers because the numbers for the two x86 architecture installations are galloping, but still. :) -- 2. That which causes joy or happiness. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Are popcon stats per package and arch possible?
On Sun, 6 May 2007 13:07:48 +0200 Michael Hanke <[EMAIL PROTECTED]> wrote: > So far, I only know the general fraction of non-i386 users. But this > fraction is most likely very different for particular fields (e.g. > office suite on ARM machines or embedded sutff on AMD64). (Which is then skewed by people like me who cross-build ARM on AMD64 and run amd64 versions of embedded applications like gpe-* for testing purposes.) ;-) -- Neil Williams = http://www.data-freedom.org/ http://www.nosoftwarepatents.com/ http://www.linux.codehelp.co.uk/ pgpFMYd2a7LYo.pgp Description: PGP signature
Re: Are popcon stats per package and arch possible?
On Sat, May 05, 2007 at 11:23:54AM -0700, Steve Langasek wrote: > On Sat, May 05, 2007 at 06:23:36PM +0200, Petter Reinholdtsen wrote: > > > [Michael Hanke] > > > To me it looks like stats for the major architectures up to (and > > > including) powerpc are ok wrt privacy concerns. Do you agree? > > > I'm not sure if that would be the correct cutoff point, or if only > > amd64 and i386 have enough submissions to ignore the privacy issue. > > Well, at that point what use is a per-arch stat anyway? When I first contact some upstream authors of a software I want to package, I often faced the argument: 'We already provide binaries that should run on Linux systems and therefore we do not see why we should support your packaging attempt.' My usual answer is to describe all the nice features a Debian package provides. After a lengthy discussion they normally agree that from the user perspective a Debian package offers a much higher convenience level. But sometimes upstream does not agree. Nevertheless, when they say, 'we provide binaries for Linux', they always mean i386 Linux with everything linked statically to a huge binary blob. I'd really like to be able to provide some hard numbers about users of similar packages (same field or a direct competitor) running it on arches different from i386. So far, I only know the general fraction of non-i386 users. But this fraction is most likely very different for particular fields (e.g. office suite on ARM machines or embedded sutff on AMD64). Cheers, Michael -- GPG key: 1024D/3144BE0F Michael Hanke http://apsy.gse.uni-magdeburg.de/hanke ICQ: 48230050 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Are popcon stats per package and arch possible?
On Sat, May 05, 2007 at 11:23:54AM -0700, Steve Langasek wrote: > On Sat, May 05, 2007 at 06:23:36PM +0200, Petter Reinholdtsen wrote: > > > [Michael Hanke] > > > To me it looks like stats for the major architectures up to (and > > > including) powerpc are ok wrt privacy concerns. Do you agree? > > > I'm not sure if that would be the correct cutoff point, or if only > > amd64 and i386 have enough submissions to ignore the privacy issue. > > Well, at that point what use is a per-arch stat anyway? not to mention amd64 i386non-x86 = === . . . might be interesting :-) and might make a comfortable starting point for exploring whether there is anything interesting to be found in there. Regards, Paddy -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Are popcon stats per package and arch possible?
On Sat, May 05, 2007 at 11:23:54AM -0700, Steve Langasek wrote: > On Sat, May 05, 2007 at 06:23:36PM +0200, Petter Reinholdtsen wrote: > > > [Michael Hanke] > > > To me it looks like stats for the major architectures up to (and > > > including) powerpc are ok wrt privacy concerns. Do you agree? > > > I'm not sure if that would be the correct cutoff point, or if only > > amd64 and i386 have enough submissions to ignore the privacy issue. > > Well, at that point what use is a per-arch stat anyway? To illustrate that there are no arches any more ? it brings out an interesting detail. presumably the boxes of amd64 users are faster on average. maybe users of faster hardware prefer different software. but at that point, arch is a very rough proxy for 'fast', for which there could be more interesting numbers available. I would be surprised if there weren't some users who wouldn't happily provide more detailed info about the host hardware, and similarly I would imagine that different users would have different thresholds for how that information might be used. A really good trick would be to come up with a more general system by which people could publish information about themselves in confidence expressing constraints on how that information could be used to create statistics that would then be made public. Although I am not aware of it, I would be surprised if work had not already been done in this area (and, no, I don't mean DRM). You probably also want policy on the other side. FWIW, I would think you could get interesting numbers out down at least as far as the sparc group, if not further. Techniques might include suppressing particularly sensitive packages (social-outcast-1.1, security-risk-5.2) suppressing small numbers that might identify particular individuals. publishing finished comparisons, rather than rawer data. For example, "MTA(s) installed". Regards, Paddy -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Are popcon stats per package and arch possible?
On Sat, May 05, 2007 at 06:23:36PM +0200, Petter Reinholdtsen wrote: > [Michael Hanke] > > To me it looks like stats for the major architectures up to (and > > including) powerpc are ok wrt privacy concerns. Do you agree? > I'm not sure if that would be the correct cutoff point, or if only > amd64 and i386 have enough submissions to ignore the privacy issue. Well, at that point what use is a per-arch stat anyway? -- Steve Langasek Give me a lever long enough and a Free OS Debian Developer to set it on, and I can move the world. [EMAIL PROTECTED] http://www.debian.org/ -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Are popcon stats per package and arch possible?
[Michael Hanke] > To me it looks like stats for the major architectures up to (and > including) powerpc are ok wrt privacy concerns. Do you agree? I'm not sure if that would be the correct cutoff point, or if only amd64 and i386 have enough submissions to ignore the privacy issue. I'm not quite sure how to make that judgment either. > I'd be very interested in this information and would be glad if it > could be made available. I am afraid someone else would have to implement support for it. The server part of popularity-contest is in the popularity-contest package, in /usr/share/doc/popularity-contest/examples/. It would probably be a good idea to make the cutoff point configurable. :) Friendly, -- Petter Reinholdtsen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Are popcon stats per package and arch possible?
On Sat, May 05, 2007 at 10:40:34AM +0200, Petter Reinholdtsen wrote: > [Michael Hanke] > > I wonder whether it is possible to get information about which > > package is installed/used on a particular architecture. > > It is possible, but it isn't done at the moment. We do not store the > data needed to generate such reports, but it could be done with the > data set we have at every given point in time. > > But I would be reluctant to do it for the less used architectures, for > privacy reasons. The information we get from such analysis would be > too easy to associate with individual users (like the kfreebsd-amd64 > porters for that arch. :). Right, thanks for pointing this out. > And for those interested, here are the latest numbers: > > 2 0.00% i486 > 2 0.00% kfreebsd-amd64 > 3 0.01% hurd-i386 > 3 0.01% ppc64 > 8 0.02% armel > 9 0.02% armeb > 9 0.02% s390 >11 0.02% m68k >11 0.02% kfreebsd-i386 >20 0.04% mips >42 0.09% ia64 >57 0.12% hppa >65 0.14% mipsel >68 0.14% alpha > 223 0.47% sparc > 598 1.25% powerpc > 676 1.42% arm > 5625 11.79% amd64 > 40283 84.42% i386 > 47715 100.00% total (ignored 229 without arch info) To me it looks like stats for the major architectures up to (and including) powerpc are ok wrt privacy concerns. Do you agree? I'd be very interested in this information and would be glad if it could be made available. Thanks, Michael -- GPG key: 1024D/3144BE0F Michael Hanke http://apsy.gse.uni-magdeburg.de/hanke ICQ: 48230050 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Are popcon stats per package and arch possible?
On Sat, 2007-05-05 at 10:40 +0200, Petter Reinholdtsen wrote: > [Michael Hanke] > > I wonder whether it is possible to get information about which > > package is installed/used on a particular architecture. > > It is possible, but it isn't done at the moment. We do not store the > data needed to generate such reports, but it could be done with the > data set we have at every given point in time. > > But I would be reluctant to do it for the less used architectures, for > privacy reasons. The information we get from such analysis would be > too easy to associate with individual users (like the kfreebsd-amd64 > porters for that arch. :). I agree about "privacy reasons". Thanks for keeping this aspect in mind. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Are popcon stats per package and arch possible?
[Michael Hanke] > I wonder whether it is possible to get information about which > package is installed/used on a particular architecture. It is possible, but it isn't done at the moment. We do not store the data needed to generate such reports, but it could be done with the data set we have at every given point in time. But I would be reluctant to do it for the less used architectures, for privacy reasons. The information we get from such analysis would be too easy to associate with individual users (like the kfreebsd-amd64 porters for that arch. :). And for those interested, here are the latest numbers: 2 0.00% i486 2 0.00% kfreebsd-amd64 3 0.01% hurd-i386 3 0.01% ppc64 8 0.02% armel 9 0.02% armeb 9 0.02% s390 11 0.02% m68k 11 0.02% kfreebsd-i386 20 0.04% mips 42 0.09% ia64 57 0.12% hppa 65 0.14% mipsel 68 0.14% alpha 223 0.47% sparc 598 1.25% powerpc 676 1.42% arm 5625 11.79% amd64 40283 84.42% i386 47715 100.00% total (ignored 229 without arch info) Friendly, -- Petter Reinholdtsen -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]