Re: [gentoo-dev] Gentoostats, SoC 2011
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 On 25-08-2011 14:35, Alec Warner wrote: > On Thu, Aug 25, 2011 at 5:20 AM, Rich Freeman > wrote: >> The big issue with opt-out is privacy law - especially in Europe >> (that's leaving aside just being up-front with users). We'd end >> up having to have EULAs or such and perhaps a number of other >> legal controls, and I don't think that is a direction that we want >> to go in. I'm just not seeing the upside - better to just figure >> out good ways to use data that is easy and safe to obtain first. >> >> Earlier somebody suggested that this decision wasn't really in the >> domain of the Council/Trustees. I'm not sure I agree here - any >> kind of opt-out data collection is something that has potential >> legal ramifications as well as huge reputation concerns for the >> distro (the software is distributed from Foundation-owned hardware >> utilizing a Foundation-owned domain name and the data goes back to >> Foundation-owned hardware - I'm sure any lawyer could make a case >> for this). Just because there isn't a policy written down >> somewhere doesn't mean that we can't use common sense. Devs >> certainly don't need to run everything past the Council, but if you >> want to do something high-profile post it on -dev, and if there is >> an uproar look for an official second opinion before doing it. > > We did post to -dev, hence this thread. The point is that we don't > need any 'official opinion' to do anything; and I don't want to set > that precedent. If you have specific concerns about actions we plan > to take (which by the way, we are not planning an opt-out solution. > If we plan to do an opt-out solution, we will again have a thread on > -dev) then let us know. If you have specific legal concerns about > the application, data retention, encryption, logs, backups, onerous > european privacy laws, and other such questions you should raise > those concerns now. I've picked this message as I want to address one point in this thread that was focused on this sub-thread. I disagree with the idea that adding an application to the Gentoo tree that collects data from users and sends it to a central (or distributed) system is the same as adding any other application to the tree. Having the ability to add ebuilds to the tree is part of what you gain by getting gentoo-x86 access. Issues with significant users privacy concerns and substantial changes like adding packages to the tree that collect data from users and compile it, should not be at the discretion of individual developers but be subject of global policies that should take into account the legal ramifications (trustees) and reflect the developers desire and goals (council). - -- Regards, Jorge Vicetto (jmbsvicetto) - jmbsvicetto at gentoo dot org Gentoo- forums / Userrel / Devrel / KDE / Elections / RelEng -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.18 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iQIcBAEBAgAGBQJOVxCXAAoJEC8ZTXQF1qEP7KAQAJBwDHp4aS+5l8gahHUrsWYI 0gUpO+qtsFODsKToQa4ZZ9jTZhFvN0iscyApXvgO8FBOnPzFCMiq+LblI/j/cnFK OwVYJ4/tvcc1C1fE1lQecd1kNVlnVLCEvR8NbeKA184ty4kS7cJy2FqAiWbzGGno /zNsQI+iDUg6ZCamCz29EZ5FJgfUzXzG+Ipbh61T0c/Ukugq5xHA8c5zTzoRre2u /fSRMM9qPakmgaHJoV8t+8B0ejJccW/+MquKIyFdDnUDvQH5U/RnXl3D5oe7+0vb Eak3VB5iUrkZifqhpOQMEeAtuNColigPy4oPr6BsQz7t0uiC2M0MHei4cigbN8kn yp4U+RZE4PhJ/+b/U/jnaiidGu8IF+Kdl3DPgCR130N4vbpO8u7KjyphdoL7QZx5 hnc3A5ZxQxraQolKtFnl8Be8P5NvuKdiP192wYmACuCw3W95XVNDtUhc63n++fqo 0K9WTEudO+JZN7JYZFSU6OJo5hvujHcQvvIO2sG30Q56x7EfvCRFCzMUsRC8mU0L uSKW+YFHVp1+yCJ9BbnTWp9afPUVQ56/1YtCxLDsqEi0lI7otm0TpuJFIC/fDJ1F Hf9Kqaap9kZzc1WBKuMY0Rvvf8CKf/9bd9QTxT5Fz/tpiNGkU9MTMFPHghDFUP8h 773YR/NFapQVLHyqemla =G4Y6 -END PGP SIGNATURE-
Re: [gentoo-dev] mesa r600 gallium news item
Francisco Blas Izquierdo Riera (klondike) schrieb: > You can point the reader to > http://www.x.org/wiki/RadeonFeature#Decoderringforengineeringvsmarketingnames > in order to know wether his ATI card is or not an r600 ATI card. > Specially since the reference to HD2000 includes some r500 cards. You are right, but that page is slightly confusing and still not 100% accurate. I think I'll change it to "HD 2400 and later", which should cover exactly the r600 driven cards. Best regards, Chí-Thanh Christopher Nguyen
Re: [gentoo-dev] mesa r600 gallium news item
El 25/08/11 23:25, Matt Turner escribió: > 2011/8/25 Chí-Thanh Christopher Nguyễn : >> Hello, >> >> Please see the attached news item for review. The news item should be >> published before mesa-7.11 goes stable. >> Corresponding bug report: https://bugs.gentoo.org/show_bug.cgi?id=377349 >> >> >> Best regards, >> Chí-Thanh Christopher Nguyễn > Looks good to me. > > Matt You can point the reader to http://www.x.org/wiki/RadeonFeature#Decoderringforengineeringvsmarketingnames in order to know wether his ATI card is or not an r600 ATI card. Specially since the reference to HD2000 includes some r500 cards. signature.asc Description: OpenPGP digital signature
Re: [gentoo-dev] mesa r600 gallium news item
2011/8/25 Chí-Thanh Christopher Nguyễn : > Hello, > > Please see the attached news item for review. The news item should be > published before mesa-7.11 goes stable. > Corresponding bug report: https://bugs.gentoo.org/show_bug.cgi?id=377349 > > > Best regards, > Chí-Thanh Christopher Nguyễn Looks good to me. Matt
[gentoo-dev] mesa r600 gallium news item
Hello, Please see the attached news item for review. The news item should be published before mesa-7.11 goes stable. Corresponding bug report: https://bugs.gentoo.org/show_bug.cgi?id=377349 Best regards, Chí-Thanh Christopher Nguyễn Title: Mesa r600 driver now defaults to gallium Author: Chí-Thanh Christopher Nguyễn Content-Type: text/plain Posted: 2011-08-28 Revision: 1 News-Item-Format: 1.0 Display-If-Installed: http://www.gentoo.org/doc/en/xorg-config.xml
[gentoo-dev] Last rites: sys-boot/arcboot
# Matt Turner (25 Aug 2011) # Masked for removal in 30 days. Use arcload instead. sys-boot/arcboot
Re: [gentoo-dev] Gentoostats, SoC 2011
On Thu, Aug 25, 2011 at 10:35 AM, Alec Warner wrote: > We did post to -dev, hence this thread. My post was intended to be general in applicability, and not critical of the particular instance of this issue being discussed. I would generally suggest that implementing this as a package and not as a function built-into portage would tend to make more sense to me (do we really want portage to do EVERYTHING?). However, I don't think that anybody needs anybody's blessing in particular to take one course or the other there. And, in the Gentoo tradition of everybody-does-whatever-they-want-to, there is nothing wrong with one set of devs doing it one way and another set doing it another way so that we end up with two data repositories with somewhat redundant data so that we can start another discussion on -dev about what the differences in the datasets mean. That is, until eventually devs get bored and after enough bugs pile up one or both of the collection mechanisms gets treecleaned. Then in five years somebody can build a new one. :) If I had strong concerns with anything that seemed likely to get adopted I'd voice them. Rich
Re: [gentoo-dev] Re: Gentoostats, SoC 2011
On 25-08-2011 22:15:22 +0800, Patrick Nagel wrote: > The prompt should offer three options: > > [s]end the data directly > s[h]ow me the data* > s[k]ip > > You can disable this prompt by having either 'SEND_STATS="yes"' (to always > send) or 'SEND_STATS="no" (to never send) in your /etc/make.conf. > > *) And in the next step, after showing the data set(s): Send? [y/n] > > (why do all those words have to start with an 's'??) send display/view later :) -- Fabian Groffen Gentoo on a different level
Re: [gentoo-dev] Gentoostats, SoC 2011
On Thu, Aug 25, 2011 at 5:20 AM, Rich Freeman wrote: > On Thu, Aug 25, 2011 at 6:48 AM, Roy Bamford wrote: >> It has to be opt-in as opt out would be a dangerous precendent to set. >> >> I don't see any harm is a gentle reminder message from emerge, provided >> that the reminder can be turned off too, if the user really does not >> want to opt in. Thats no worse than being nagged about unread news. > > I tend to agree, the more I think about it. > > The simplest solution (which doesn't require any portage mods/etc), is > to simply make this a package that installs the appropriate logic in > cron.daily, and we send out a news item encouraging users to install > it voluntarily. If the user does nothing, they don't get the package. > > If somebody can come up with really good reason that we should be more > aggressive in promoting it, then we can promote it more aggressively. > That /might/ go as far as a forced opt-in/out decision. However, the > more I think about it the more I'm concerned with pure opt-out by > default. Why is the thread bikeshedding an out-opt that we aren't even considering doing right now? > > The big issue with opt-out is privacy law - especially in Europe > (that's leaving aside just being up-front with users). We'd end up > having to have EULAs or such and perhaps a number of other legal > controls, and I don't think that is a direction that we want to go in. > I'm just not seeing the upside - better to just figure out good ways > to use data that is easy and safe to obtain first. > > Earlier somebody suggested that this decision wasn't really in the > domain of the Council/Trustees. I'm not sure I agree here - any kind > of opt-out data collection is something that has potential legal > ramifications as well as huge reputation concerns for the distro (the > software is distributed from Foundation-owned hardware utilizing a > Foundation-owned domain name and the data goes back to > Foundation-owned hardware - I'm sure any lawyer could make a case for > this). Just because there isn't a policy written down somewhere > doesn't mean that we can't use common sense. Devs certainly don't > need to run everything past the Council, but if you want to do > something high-profile post it on -dev, and if there is an uproar look > for an official second opinion before doing it. We did post to -dev, hence this thread. The point is that we don't need any 'official opinion' to do anything; and I don't want to set that precedent. If you have specific concerns about actions we plan to take (which by the way, we are not planning an opt-out solution. If we plan to do an opt-out solution, we will again have a thread on -dev) then let us know. If you have specific legal concerns about the application, data retention, encryption, logs, backups, onerous european privacy laws, and other such questions you should raise those concerns now. > > Rich > >
Re: [gentoo-dev] Re: Gentoostats, SoC 2011
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi, On 2011-08-25 20:43, Markos Chandras wrote: > On 25/08/2011 11:42 ??, Nikos Chantziaras wrote: >> On 08/24/2011 01:48 PM, Patrick Lauer wrote: >>> [...] If you sneakily add something to cron.daily by default you can >>> get pretty nice coverage. But I guess anyone trying that in >>> Gentooland will meet some rather unpleasant resistance :) > >> emerge always asks me after a world update whether I want to "auto >> clean packages" with a yes/no prompt. I wouldn't be bad if once a >> month or whatever it would ask me whether I want to upload my stats. >> Gentoostats should probably become a runtime dep of Portage itself by >> default, but not used automatically. > > I like your idea and people seem to like making things complicated. > Simple solution: > > opt-in > > How: Display a warning after an emerge -u{DNav} world. Let user disable > this warning by using a special variable in make.conf > > STATS_ENABLE="no". > > By default, this variable will be "yes" on base/ profiles That sounds perfect to me. The prompt should offer three options: [s]end the data directly s[h]ow me the data* s[k]ip You can disable this prompt by having either 'SEND_STATS="yes"' (to always send) or 'SEND_STATS="no" (to never send) in your /etc/make.conf. *) And in the next step, after showing the data set(s): Send? [y/n] (why do all those words have to start with an 's'??) Cheers, Patrick. - -- Key ID: 0x86E346D4http://patrick-nagel.net/key.asc Fingerprint: 7745 E1BE FA8B FBAD 76AB 2BFC C981 E686 86E3 46D4 -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.18 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk5WWPoACgkQyYHmhobjRtSwewCgyJzgsLLvjfZpX5vg8XcxkNMb tg8AoIkHz1z6b9DxTrnJxe3YyTDMOYsr =ZKYx -END PGP SIGNATURE-
Re: [gentoo-dev] Re: Gentoostats, SoC 2011
-BEGIN PGP SIGNED MESSAGE- Hash: SHA512 On 25/08/2011 11:42 ??, Nikos Chantziaras wrote: > On 08/24/2011 01:48 PM, Patrick Lauer wrote: >> [...] If you sneakily add something to cron.daily by default you >> can get pretty nice coverage. But I guess anyone trying that in >> Gentooland will meet some rather unpleasant resistance :) > > emerge always asks me after a world update whether I want to "auto > clean packages" with a yes/no prompt. I wouldn't be bad if once a > month or whatever it would ask me whether I want to upload my > stats. Gentoostats should probably become a runtime dep of Portage > itself by default, but not used automatically. > > I like your idea and people seem to like making things complicated. Simple solution: opt-in How: Display a warning after an emerge -u{DNav} world. Let user disable this warning by using a special variable in make.conf STATS_ENABLE="no". By default, this variable will be "yes" on base/ profiles - -- Regards, Markos Chandras / Gentoo Linux Developer / Key ID: B4AFF2C2 -BEGIN PGP SIGNATURE- Version: GnuPG v2.0.18 (GNU/Linux) iQIcBAEBCgAGBQJOVkNbAAoJEPqDWhW0r/LCWFEP+wSHLeGuqIGIV1OcHzlrgMJn qBusQcvWb2SbP58G+3hndPUHWsnSIRDk2zws21wZH1DLaCr3Hyc5V9s8WM7HMWeX WHgLVN71KDSmhREvz9fJWGA08rLJHa/Xs3XdNBUp5mMGsogPmwgEmNgSyfpKe6wC nsBC6pk/R6kvAdlinyLp2uybBDI00GtiUSN3fyvvuoE3XMlVdyYLYTKTguVmJB4A aR97bIR2otZlWTmSGMvr899vht35Mwoe0/UHDV2SJHtWk9zzIYzWHEW9mMUoWAAw 1/2bYAafE+rp0/MPTv0tJ8WtoB6NhpJBRZGa5MuQM0V8jQqbgweq1YBO4zO8VAO+ tn6/thWk+heSklzTR1Uom1F+pYMxJBY0nBQUQ6ppZLjOM43WTT5TmPkc/rYwEB3M 8FywLV+Mryax+nQesLkgMQLLKUSuQygy+4zxoCUHkY5DwB08C5dfYua/Y+nLbb0l iCmCNpYSKUJF2cbHm6kDfOonsyfre//+JNvycxuz5J932XqFwkv7dM8ywSwP7Srw FfWWspLfHxyoXSHlljwXj0UxfQcjNbg3MyUt8s0pSys9GlUzdSv1vx4xGgS8dtzA zTSvTE1xO3fUzO1Au+KaSs0GC49BO7Gf1fWFdIhr8ZcJD82+jDaooompzPqFm0lB XZQAVwslAMQbSaJAEfCY =V+cS -END PGP SIGNATURE-
Re: [gentoo-dev] Gentoostats, SoC 2011
On Thu, Aug 25, 2011 at 6:48 AM, Roy Bamford wrote: > It has to be opt-in as opt out would be a dangerous precendent to set. > > I don't see any harm is a gentle reminder message from emerge, provided > that the reminder can be turned off too, if the user really does not > want to opt in. Thats no worse than being nagged about unread news. I tend to agree, the more I think about it. The simplest solution (which doesn't require any portage mods/etc), is to simply make this a package that installs the appropriate logic in cron.daily, and we send out a news item encouraging users to install it voluntarily. If the user does nothing, they don't get the package. If somebody can come up with really good reason that we should be more aggressive in promoting it, then we can promote it more aggressively. That /might/ go as far as a forced opt-in/out decision. However, the more I think about it the more I'm concerned with pure opt-out by default. The big issue with opt-out is privacy law - especially in Europe (that's leaving aside just being up-front with users). We'd end up having to have EULAs or such and perhaps a number of other legal controls, and I don't think that is a direction that we want to go in. I'm just not seeing the upside - better to just figure out good ways to use data that is easy and safe to obtain first. Earlier somebody suggested that this decision wasn't really in the domain of the Council/Trustees. I'm not sure I agree here - any kind of opt-out data collection is something that has potential legal ramifications as well as huge reputation concerns for the distro (the software is distributed from Foundation-owned hardware utilizing a Foundation-owned domain name and the data goes back to Foundation-owned hardware - I'm sure any lawyer could make a case for this). Just because there isn't a policy written down somewhere doesn't mean that we can't use common sense. Devs certainly don't need to run everything past the Council, but if you want to do something high-profile post it on -dev, and if there is an uproar look for an official second opinion before doing it. Rich
Re: [gentoo-dev] Gentoostats, SoC 2011
On 2011.08.24 11:48, Patrick Lauer wrote: [snip] > > If you sneakily add something to cron.daily by default you can get > pretty nice coverage. But I guess anyone trying that in Gentooland > will > meet some rather unpleasant resistance :) > > > > This app and if its opt in or opt out will set a precedence for any future apps that want automatic user feedback in Gentoo It has to be opt-in as opt out would be a dangerous precendent to set. I don't see any harm is a gentle reminder message from emerge, provided that the reminder can be turned off too, if the user really does not want to opt in. Thats no worse than being nagged about unread news. -- Regards, Roy Bamford (Neddyseagoon) a member of elections gentoo-ops forum-mods trustees pgpz8BkPEPndt.pgp Description: PGP signature
[gentoo-dev] Re: Gentoostats, SoC 2011
On 08/24/2011 01:48 PM, Patrick Lauer wrote: [...] If you sneakily add something to cron.daily by default you can get pretty nice coverage. But I guess anyone trying that in Gentooland will meet some rather unpleasant resistance :) emerge always asks me after a world update whether I want to "auto clean packages" with a yes/no prompt. I wouldn't be bad if once a month or whatever it would ask me whether I want to upload my stats. Gentoostats should probably become a runtime dep of Portage itself by default, but not used automatically.
Re: [gentoo-dev] Gentoostats, SoC 2011
On Wed, 24 Aug 2011 13:03:44 +0200 "Andreas K. Huettel" wrote: > Am Mittwoch 24 August 2011, 12:48:35 schrieb Patrick Lauer: > > > > If you sneakily add something to cron.daily by default you can get > > pretty nice coverage. But I guess anyone trying that in Gentooland > > will meet some rather unpleasant resistance :) > > > > Of course, we could place it in some blatantly obvious way into a > default configuration, together with a big fat message what it does > and how to quickly disable it. > > We'd get better coverage in an opt-out system than in an opt-in > system. And a larger number of angry users which missed the warning and now have to pay for additional GPRS transfer or so. And when people use GPRS rarely, they usually don't think about random apps that use the connection in background. > (First idea- package is pulled in by a default-on useflag and > installs itself into cron.daily. BEFORE it runs the first time it > outputs said message and asks for permission to proceed (which cannot > be done in the cron job obviously but we'd find a way).) And what if it can't ask for that? Assuming you're talking about 'opt-out', I guess the fallback would be to 'yes'. We don't want to end up like Windows, where you get AFK for five minutes and then discover the system has rebooted. -- Best regards, Michał Górny signature.asc Description: PGP signature