- The following discussion is closed. Please do not modify it. Subsequent comments should be made in a new section.
We welcome comments in all languages, and would like to ask anyone willing to help out to help facilitate communication by translating comments (even in summary). Thank you! --Maggie Dennis (WMF) (talk) 18:43, 19 June 2013 (UTC)
What info is collected about readers and editors
I would like to see the policy updated to include what specific information is collected about readers and also editors. I believe some of the information may be anonymized before others have access to it, so I would like to see that clarified as well. 18.104.22.168 05:33, 19 June 2013 (UTC)
- Likewise, and for searchers, and how long is each category retained? I've been trying to follow the discussion about transitioning readership log retention from 1/1000 to 100%, without much luck. Are the questions about that (e.g., "For how long is the access log kept for readers and is there an access log for all readers or just a 1/1000 sample?") going to be answered? EllenCT (talk) 21:02, 19 June 2013 (UTC)
I've broken this out into a series of specific questions below. –SJ talk 21:52, 20 June 2013 (UTC)
What is logged, for how long?
- What logs are kept?
- How long are logs kept?
- Which logs are kept for all users, for all readers, for a random subset?
- Which logs and data are available to developers? To anyone with a feed?
- What sorts of secondary, composite or processed logs are produced? (e.g., wikistats)
- What data is stripped out or anonymized from each (secondary) log?
One key thing is just to make sure everything is specific. Saying "a limited period of time" just doesn't cut it. Copyright lasts a limited period of time. Even in the year 384,572,250, the copyright on "Happy Birthday" will still be less than 20 years from expiring. Let's hope that's not also the case for today's Wikipedia records. Wnt (talk) 21:58, 19 June 2013 (UTC)
- Hi Wnt! Thank you for your input. While we hope that the Wikimedia projects are still thriving in the year 384,572,250, our data retention practices will hopefully not mimic the unfortunate copyright on the "Happy Birthday" song. =) Mpaulson (WMF) (talk) 22:35, 25 June 2013 (UTC)
Borrado de las páginas de usuario
Aunque sea una propuesta totalmente contraria a la política general sobre edición de páginas, me parece deseable que las páginas de usuario (no su discusión) pudieran borrarse de forma absoluta y por el propio usuario sin necesidad siquiera de "llamar la atención" sobre ello pidiéndolo en los tablones públicos.
Muchos usuarios comienzan en Wikipedia aportando datos sensibles sobre sí mismos (lugar de residencia, ideas políticas o religiosas, identidad sexual, etc.) que desearían, mas tarde y por razones que nadie debería juzgar, que fuesen retiradas no sólo de la página visible, sino incluso de historiales que pudieran ser rastreados. Por supuesto, esto puede ser de inmensa importancia si, en algún momento, tales editores comenzasen a trabajar en temas que fuesen susceptibles de algún tipo de revancha en el mundo real (conflictos políticos, corrupción económica, criminalidad, etc), aunque con más frecuencia podría evitar conflictos en el mundo laboral, en el que hoy en día es cada vez más frecuente que las empresas rastreen el historial en Internet de sus candidatos. Un saludo. --Fremen (talk) 09:37, 19 June 2013 (UTC)
- Translation: The Anonymouse (talk) 23:15, 20 June 2013 (UTC)
Deletion of user pages
Although it is a completely against the general policy on editing pages, it seems desirable that the user pages (not discussion pages) could be deleted by the user without even needing to "get attention" by asking for on public boards.
Many users that start at Wikipedia provide sensitive information about themselves (where they live, political or religious beliefs, sexual identity, etc.). Later, they would like (and for reasons no one should judge them) for it to be removed not only from the visible page, but even of records that could be tracked [the history]. Of course, this can be very important if, at some point, such editors begin working on issues that were subject to some kind of revenge in the real world (political conflicts, economic corruption, crime, etc.), but more often for avoiding conflicts in the workplace, as it is now increasingly common for companies to track the internet history of their candidates. Best regards. --Fremen (talk) 09:37, 19 June 2013 (UTC)
- If every user could delete their user page, then there would be many users who move their own talk page there and delete those discussions or vandals who move other pages onto their user page to delete them there. If they moved more than one page after another there and delete it there, then the version histories of different pages would be merged together and admins would have to do a big job to get those pages fixed again. Maybe, it wouldn’t be possible, if then there would be to many versions and stewards would have to do this. Or it would be too difficult at all, so noone would fix those things. As long as a user page is being treated the same way as any other page and everyone can edit or move it anywhere or move other pages there, I strongly oppose this. Then there would first be a need for a software change, so that only bureaucrats and stewards (ok, every admin/sysop is also ok, since they can delete and restore pages anyway, and often users create articles on their user page which then should be moved by admins to a user subpage) would be able to move user pages or to move pages there. But in no way should every user (or any user who wasn’t elected as an admin) be technically able to delete any pages they want.
- I have also seen an admin having moved different own user pages which weren’t needed anymore (but where also other users had made contribs to) to one place in the user namespace (called trash or something like that) and deleted them there after another. I think that later there were versions restored which was a mix of different version histories and really confusing. This way, it shouldn’t be done. But if admins do such kind of things, what will normal or new, unexperienced users or even vandals do with this possibility then?
- Any user can e-mail an admin for a request to delete their user page, (s)he doesn’t have to request this onwiki.
- And I hope you only mean the main user page and not also user subpages (where also sometimes private data have been revealed), cause very often articles are moved from main article namespace to the creator’s user namespace, if they aren’t ready, but different users then already have taken part in that article in the main namespace. Such articles should never be deleted by normal users without admin rights, there should always be an admin who takes a look upon it before deleting it. --– Geitost diskusjon 21:06, 23 June 2013 (UTC)
- Probably someone could implement a filter that stops users from moving pages to their user page? What is the benefit of moving an article to one´s user page if it is outside the user namespace? --22.214.171.124 16:56, 25 June 2013 (UTC)
- I don’t think that there’s any benefit with that, except perhaps, if someone has moved his user page onto a subpage and then later wants to move it back again with the version history. But that could also be solved, if only admins would have the right to move a page onto someone’s (main) user page, cause they may delete and restore pages. Then you would have to ask an admin to move the page there, that wouldn’t be such a problem.
- I have also once seen a user moving an article of a new user onto that user’s main user page, cause the article wasn’t good enough. I then choose to move the article further onto one of the user’s sub pages, where it then a bit later has been deleted. Thinking about that now, I think it is good that this article wasn’t deleted on the user’s main page, where the user never had placed it himself, and maybe at some time in the future together with other deleted normal user main page versions. So, it’s no good idea that this possibility isn’t blocked by the software (MediaWiki) itself, that’s no task for filters, but for the software. There’s no need for filter logs, it just should be forbidden for all normal users which aren’t admins.
- Also the fact, that users can move their user’s main page onto another (not-)existing user’s main page, perhaps because they want to rename themselves (which happens from time to time), would be blocked then automatically, by the way. That would also be very useful. Should also be the case if a user wants to move a user’s main talk page into another (not-)existing user’s namespace (onto a user’s main talk page, user’s shouldn’t be able to move pages onto a user’s main page or main talk page either; it doesn’t differ, if it’s their own page or one of another person). The second one with the user’s talk page move should only be done by bureaucrats with renaming users and – after August – that’s a task for the stewards, but nevertheless, bureaucrats could still do that, if there would be any need for it after August. I’ve never seen any useful move of a page onto a user’s main (talk) page except for bureaucrats renaming users. --Geitost diskusjon 00:55, 26 June 2013 (UTC)
- Hi All. The ability to delete user pages is certainly a sensitive and controversial topic. While it does touch on privacy concerns, it strikes the heart of a deeper question -- when should something be struck from the public records of a Wikimedia project? This question is one that belongs with the community. If there was or ever is a great call from the community for the ability to delete user pages, we would look into the feasibility of helping answer that call. However, such a change would need substantial community support to be considered. (Side note - time permitting, would someone be kind enough to translate this thread into Spanish for Fremen? Thank you!) Mpaulson (WMF) (talk) 22:49, 25 June 2013 (UTC)
- Traducción: Lguipontes (talk) 17:27, 26 June 2013 (UTC) (eh, yo hablo portugués, disculpa si mi castellano de escuela brasileña es demasiado malo; tengo certeza que hay portuñol como las bananas de Brasil aquí, para dar y vender :3)
Re: Borrado de las páginas de usuario
Si cada usuario podríamos deletar nuestras páginas, entonces hubería muchos usuarios que mueven sus proprias conversas allí para deletar aquellas discusiones, o vándalos que mueven otras páginas en sus proprias para poder deletarlas. Si [estos vándalos] moviesen después de otra y deletasen todas allí, entonces los históricos de ediciones pudrían ser misturados y los administradores de Wikipédia terían que hacer un trabajo demasiado grande para fijarlas de nuevo. Quizás, eso no sería posible, si entonces allí hubieran muchas versiones y stewards (no sé como se llaman en español :S) terían que hacerlo. O podría ser que eso fuera tan difícil que no sería posible de ninguna manera, entonces nadie fijaria estas páginas. Con tal que una página de usuário es tratada en la misma manera que cualquier otra página y todos podríamos editarla o moverla en cualquer lugar, yo firmemente me opongo a eso. Entonces allí sería la primera necesidad para un canbio de software, teríamos que ter solamente burocratas y stewards (ok, cada admin es también ok, ya que ellos pueden deletar y restorar páginas de cualquer manera, y muchas veces usuarios crían artículos en sus proprias páginas de usuario, los cuales los admin tienen que mover a subpáginas) como capaces de mover páginas de usuario o de mover las paginas además allí. Pero de ninguna manera debería todo usuario (o cualquier usuario que no fuera elegido admin) ser tecnicamente capaz de deletar cualquier páginas quisiéramos.
También he visto un/una admin mover diferentes páginas de usuario de si mism@ que no le necesitaba más (pero donde también otros usuarios contribuyeron) a un lugar en las páginas individuales de usuario (llamado basura/papelera o algo del género) y deletaron ellas allí una después de otra. Creo que después hubieron versiones restoradas las quales había una mistura de diferentes históricos de versiones y aquello era notablemente confuso. En esta manera, nosotros no deberíamos tener eso. Pero si administradores hacen estos tipos de cosa, que harían los usuarios normales o principiantes o mismo vándalos con esa posibilidade?
Cualquier usuario puede se comunicar por e-mail con un admin para un pedido de deletar su página, ell@ no necesita hacer eso abiertamente a la visión de otros wikipedistas. Y tengo esperanza que usted solamente propone [en su sugestión] las páginas principales y no también las otras páginas "userspace" como las conversaciones (donde también hay informaciones privadas a seren reveladas), porque muchas veces artículos son movidos del espacio principal para o userspace del creador, si aún no están prontos, pero diferentes usuarios pueden tener ya contribuido para aquel artículo en la area principal de Wikipédia. Estos artículos no deberían jamás ser deletados por usuarios normales sin derechos de administración, debería siempre haber un admin que le mire antes de deletarlo. --– Geitost diskusjon 21:06, 23 de junio de 2013 (UTC)
- ¿Probablemente alguien podría implementar un filtro que evita que los usuarios "muevan" páginas de la area principal de Wikipédia para sus respectivos userspaces? ¿Cuál es el beneficio de mover un artículo a la página de un usuario si esa está fuera del userspace? --126.96.36.199 16:56, 25 de junio de 2013 (UTC)
- No creo que existiría algun beneficio con eso, con la excepción tal vez, de alguién ter movido su página de usuario a una subpágina y entonces después la mueve para el otro lugar con el histórico de versiones. Pero eso podría ser resolvido, con solamente algunos admins tengan el poder de mover una página en la página de usuario principal de alguién, porque ellos pueden deletar y restaurar las páginas. Por tanto uestes terían que preguntar un admin para mover las páginas allí, que no sería problemático.
- También he visto una vez un usuario mover un artículo creado por un novato a página de usuario principal del mismo, porque el artículo no era bueno suficientemente. Entonces elegí mover o artículo más, para las subpáginas en el userspace, donde aquello fue poco después deletado. Pensando sobre eso ahora, creo que es bueno que el artículo no fue deletado en la página principal del usuario, donde el usuario nunca le puesto el mismo, y tal vez en algun tiempo en el futuro junto a otras versiones históricas de la página normal de dominio principal. De esta manera, no es una buena idea que esa posibilidad no sea bloqueada por el software (MediaWiki) por si mesmo, esta no es tarea para filtros, pero para el software. No hay necesidad de filtros de registro, a penas debería ser proibido para todos los usuarios normales que no los administradores.
- También hay el facto de que usuarios pueden mover sus páginas de usuario principales en otras (no-)existentes páginas del mismo tipo, tal vez porque ellos puedan querer se renomear (una cosa que acontece de tiempos en tiempos), pode ser bloqueado automaticamente así, a propósito. Eso puede ser muy útil. Debe también ser el caso si un usuario pueda querer mover la conversación principal de un usuario en otra página de userspace (sea cual fuera; no importa, si es el caso de su propria página o de otra persona). La secunda opción con una página de conversación movendo debe ser hecha solamente por burocratas con renomeamento de usuarios y - después de agosto - sería una tarea para stewards, pero de cualquier manera, burocratas también podría hacerlo, caso haiga alguna necesidad para eso después de agosto. Nunca he visto alguna movida útil de una página para dentro de una página de conversación o usuario princial con la excepción de burocratas renomeando usuarios. Geitost diskusjon 00:55, 26 de junio de 2013 (UTC)
Hola a todos. La capacidad de eliminar las páginas de usuario es sin duda un tema delicado y controvertido. A pesar de que hace contacto en cuestiones de privacidad, golpea el corazón de una cuestión más profunda – cuando debe suprimirse algo de los registros públicos de un proyecto de Wikimedia? Esta pregunta es una que pertenece a la comunidad. Si hubiera o de cualquier manera hay un grande llamado de la comunidad para la habilidade de deletar páginas de usuario, podríamos mirar en la viabilidad de ayudar a responder a ese llamado. Sin embargo, un cambio de este tipo necesitaría un importante apoyo comunitario para ser considerado. (Nota al margen - si el tiempo lo permite, sería alguien cariñoso el suficiente para traducir este hilo en español para Fremen? Gracias!) Mpaulson (WMF) (conviersa) 22:49, 25 de junio de 2013 (UTC)
Location of draft
Thanks for this
Michelle and the legal team:
- Similarly, EU privacy law is considerably more private in some regards than US privacy law. We should consider what the appropriate movement-wide rules are, and not simply adopt the minimum level of privacy mandated by the country holding our servers. –SJ talk 21:56, 20 June 2013 (UTC)
- We really appreciate you taking time out of your day to give us your input on these subjects. Mpaulson (WMF) (talk) 23:31, 25 June 2013 (UTC)
Principles in responding to government requests for user data
- The Foundation should require a warrant supported by probable cause in order to disclose content of communications. (Because the Foundation's projects have so few means of communicating privately, this may have limited applicability.)
- The Foundation should promise to tell users when the government seeks their data unless prohibited by law. Notice is the only way that users can attempt to mount a legal challenge to such a request.
- The Foundation should publish statistics on how often they provide user data to the government. As we learned recently, the public cannot raise concerns about government requests for user data if they are not informed (at least in aggregate numbers) of the existence of such requests.
- The Foundation should adopt and publish policies or guidelines explaining how it responds to data demands from the government, such as guides for law enforcement. The Foundation operates in many jurisdictions. Providing guidelines to law enforcement in their native languages, taking into account the legal systems operating in those places can help guide law enforcement into making narrower, more reasonable requests, and sets expectations appropriately.
- The Foundation should promise users that if, in its judgment, a government demand for access to user content is overbroad, then the Foundation will challenge that demand in court. The reality is that the average user does not have the know-how or resources to mount a court challenge to a request for their data. The Foundation is better situated to challenge overbroad requests.
- The Foundation should adopt a policy that seeks the modernization of electronic privacy laws to defend users in the digital age. Specifically, as to the U.S., the Foundation should join the Digital Due Process Coalition. This might entail some advocacy in legislatures around the world, but the sum of all human knowledge being online helps one a lot less if one's government leaves you terrified to utilize it. Brianwc (talk) 21:17, 19 June 2013 (UTC)
- Surely national security letters from the US grovernment makes all of this moot while the foundation is bound by secret US laws? Is the only way to run the sites ethically to move the foundation to the EU? -- Jeandré, 2013-06-28t10:49z
- Moving it to Latin America and Asia-Pacific is actually much safer, according to that recent data leak. The United States is obviously interested in spying Europe. For example, the worse thing Brazilians do online is creating 2 million webpages (not websites) and making 400000 yearly content downloads related to neo-Nazism (as SOME people from Kurirama love to rant about how better-off and whiter they are and people from the Southeast and Center-West blame the fact that they aren't like the "colonizing hillbillies" on minorities and conspirations because being stupid to a pathologic level is what people in this country do best) what is a crime here but not there, we aren't terrorists, we wait 21 years of government abuse until we make a decent street manifestation, most people discussing things on the internet aren't concerned with serious business, we don't have decent intelligentsia (or at least not one with global/important plans), and we speak "foreign", so the American government got nothing to do here with individual citizens or organizations, the stealing of natural resources that is the only concern their corporates may have about us won't end so soon. Same story with about all other Latin American neutral countries. It will be even more politically correct as mostly of our energy comes from hydro-electric generation (though it will consume more conditioner air, and I tell you, light bills here are expensive). Lguipontes (talk) 22:33, 28 June 2013 (UTC)
Hi there -
- Make the policy as simple as possible (to encourage people to read it, and in terms of our movement, to make translation easier)
- To manage expectations around where the law requires policies to commit to certain things - when drafting ours I was very clear all along that I wanted wide input, but that ultimately certain elements were hidebound by EU and UK privacy laws.
Anyone who has questions more broadly or is interested on working on the UK's policy framework here please ping me on my talk page separately! :-) Katherine Bavage (WMUK) (talk) 09:16, 20 June 2013 (UTC)
As best I know the way Wikimedia handles IP-addresses is at odds with Dutch law (itself inspired on a European directive). As I understand it, to store IP addresses you may need to be registered, and to publicly show them you need a) permission from the user in question and b) an explicit statement of your internal policy on this. And you cannot keep these addresses longer than six months? - Brya (talk) 05:02, 21 June 2013 (UTC)
Auf welcher Rechtsgrundlage werden CU durchgeführt? --Liesbeth Lasst (talk) 12:31, 21 June 2013 (UTC)
Google translate renders the question as: "On what legal basis are checkusers performed?"
Data retention term
- Hi Kwj2772! I'd like to refer you to a similar question and answer concerning data retention above. It explains what we think might help alleviate this lack of clarity in our general data retention practices. If you have additional ideas about what should be included in our future data retention guidelines, we would love to hear from you. As to the process for identifying particular functionaries, this answer might address your concerns. If we do retain functionary identification in the future, the length of retention will be explained in the data retention guidelines. Mpaulson (WMF) (talk) 00:44, 26 June 2013 (UTC)
- Related: bugzilla:37626. --MZMcBride (talk) 15:22, 24 June 2013 (UTC)
A simple thing, clearly indicate:
- What datas are collected?
- How are they collected?
- How long are they kept?
- What coockies are placed?
- What datas are affected?
- How long are they actived?
Best regards -- • Hamelin [ de Guettelet ] • 01:16, 1 July 2013 (UTC)
Is there a gravy train for this one ?
- The Hong-Kong budget for the Affiliations Committee was US$40,000, and they plan to spend it on flying 9 people there. (Just sayin'.) odder (talk) 09:42, 22 June 2013 (UTC)
- I agree that there are many Wikimedia boondoggles. I'm not sure a single meeting of the Ombudsman Commission qualifies (though it would be great if Philippe and others provided a public rationale explaining why wikis, IRC, Google Hangout, Skype, Etherpad, mailing lists, private e-mail, and every other technology on the Internet were insufficient).
- I agree that there has been a recurring criticism that the Commission is slow and ineffective. This is probably outside the scope of this page, however, and should be discussed under a separate RFC. --MZMcBride (talk) 16:09, 23 June 2013 (UTC)
- Slow ? are you sure? have you heard rumours that they are actually doing something slowly ? As far as all evidence I have ever seen shows, (and I've researched a lot) they do precisely nothing whatsoever except spend money. So 'nothing' would be a more appropriate word, unless you mean spending tens of thousands of dollars, where 'slowly' wouldn't really be the word either. Thanks odder for correcting me, $40,000 is such an incomprehensible amount to so many people my memory obviously underestimated.
- My question still hasn't been answered, so I repeat, will this policy be the excuse needed to spend incomprehensible amounts of donated funds to no visible effect ? Penyulap (talk) 16:47, 5 July 2013 (UTC)
Generation of editor profiles
- Tools that allow profiling of individual user's activity (beyond what can easily be achieved directly on the public wiki sites) must only be applied with the respective user's consent (opt-in).
The relaxation of this point lead to this RFC at Meta where we find a proposal to enable more profiled statistics and graphs for X!'s edit counter which before would have been in conflict with the toolserver's policy. From the RFC:
- This opt-in was set up because because of a law in Germany, where the toolserver is located. Since we are migrating everything from the toolserver to Wikimedia's labs (in the U.S.), this law isn't relevant anymore.
- Freedom of speech and access to information are core Wikimedia values. These values can be compromised by surveillance: editors and readers understandably are less willing to write and inform themselves as honestly and freely. Put simply, "rights of privacy are necessary for intellectual freedom."
This is exactly the point, profiles of editors that are easily accessible and which reveal working periods do not serve the need to contribute to WMF projects but may have the impact that editors are less likely to contribute. --AFBorchert (talk) 11:22, 23 June 2013 (UTC)
- I strongly support this statement. WMF should not offer software that allows the systematic and automatic generation of editor profiles without explicit consent of each person. --Martina Nolte (talk) 12:37, 23 June 2013 (UTC)/12:52, 23 June 2013 (UTC)
- Support. Und jetzt warten wir auf den ersten Beitrag mit dem Vorwurf "Am deutschen Wesen soll die Welt genesen" bzw. auf den fälligen "canvassing"-Thread. -- smial (talk) 13:08, 23 June 2013 (UTC)
- +1. Indeed, using Wikimedia user data for generating profiles with tools that are not implemented in MediaWiki already should not be allowed. Opt-in solutions are a standard within the European Union.--Aschmidt (talk) 13:29, 23 June 2013 (UTC)
- Related: bugzilla:48667. --MZMcBride (talk) 16:03, 23 June 2013 (UTC)
- +1 Support --– Geitost diskusjon 20:33, 23 June 2013 (UTC) Ich widerspreche hiermit vorsorglich der Bereitstellung von aggregierten Benutzerdaten über mich, sobald sie über die bisherigen Beitragszähler hinausgehen und Profile erstellen und ich nicht explizit der Veröffentlichung der aggretierten Daten zugestimmt habe (opt-in). Dies stellt ein vorsorgliches Opt-out für derartige Daten dar und soll dementsprechend jederzeit in jede derartige Datensammlung ohne Opt-in eingefügt werden, falls es mal solche geben sollte.
- Die Auflistung der Gesamtbeiträge inkl. und exklusive gelöschten Beiträgen sowie aufgesplittet nach Namensräumen (bzw. ebenso bei den Logbucheinträgen) sollten die einzigen Datenübersichten bleiben (neben den normal üblichen Beitragslisten), die mit Unterstützung der WMF bezogen auf jeden einzelnen Benutzer frei verfügbar bereitgestellt werden. Aber keine täglich nach Uhrzeiten, wöchentlichen nach Tagen oder monatlichen Übersichten ohne Opt-in und auch keine Übersichten über die am meisten bearbeiteten Seiten in den Namensräumen oder Ähnliches ohne Opt-in. --– Geitost diskusjon 20:33, 23 June 2013 (UTC)
- Support NNW (talk) 08:35, 24 June 2013 (UTC)
- Support The rules at the Toolserver were there for a reason. Because the US is a developing country in terms of data-privacy we need at least a WMF-policy about it. --DaB. (talk) 00:36, 26 June 2013 (UTC)
- Support --Kellerkind (talk) 08:33, 26 June 2013 (UTC)
- Support — Racconish Tk 21:20, 30 June 2013 (UTC)
- Support Per Martina Nolte, Aschmidt and DaB., above. Keith Roth (talk) 12:14, 3 July 2013 (UTC)
- Support The reasons haven been stated above. --Peter Putzer (talk) 21:54, 20 July 2013 (UTC)
- Comment: You do realize that it literally impossible to prevent this kind of aggregation of editing statistics? This isn't just something the WMF or people with Toolserver database access can do. Anyone can aggregate these statistics about an individual account because the data is public, just like page histories are public. It's something that could be aggregated from the live site, the monthly XML dumps of the wikis, and other methods. If you don't want people knowing what you edit, then don't edit. There is no way to make your edit history private. Steven Walling (WMF) • talk 23:00, 20 July 2013 (UTC)
- Just because someone else can do it there is no need for Wikimedia to provide aggregated data. NNW (talk) 13:57, 21 July 2013 (UTC)
- Steven, we know this very well. But it still makes a difference if these datamining tools are hosted on WMF resources and linked to from various pages (contributions page etc). In addition, it makes a difference according to EU legislation. --AFBorchert (talk) 13:25, 22 July 2013 (UTC)
- Support per above. --Don-kun (talk) 15:30, 12 September 2013 (UTC)
- Thank you for your suggestion, MZ! We will take that into consideration. Mpaulson (WMF) (talk) 00:50, 26 June 2013 (UTC)
CUs normally never reveal the IP used by an editor. However, it is unclear whether they may or may not reveal the link between an IP having made abusive edits and an account.
Wikis have different views on this: metawiki and stewards never reveal IPs; enwiki reveals the link implicitly (by choosing to block or not to block an IP involved in an RCU); frwiki reveals the link only with IPs that have been mentioned in an RCU; nlwiki and svwiki seem to work this way too (more or less)...
It would be great if the "Policy on Release of Data" could be updated so that CheckUsers would not make assumptions on the policy.
Elfix 07:24, 1 July 2013 (UTC)
- Hi, You raise an important issue. I agree and by the way, I am very much against this policy whether it allows implicit or a fortiori explicit revealing of this link. I don’t know how to do without it but it would be much better in terms of privacy. Anyway, you are right that CheckUser policy should be updated and made clear. Keith Roth (talk) 12:22, 3 July 2013 (UTC)
Since August 2011 the default value of MediaWiki is 180 days, see mw:Manual:$wgCookieExpiration. Change was made with rev:94430: Bumped $wgCookieExpiration to 180 days, we were one of the sites with shortest cookie lifetime for too long. This was discussed at lengths about a year ago on wikitech-l, however nobody cared to implement the suggested alternative solutions. Probably because they were overcomplicated and solved only parts of the problem.
Q1: As far as I know, we use SecurePoll on Board elections. I think voters would wonder which informations are transmitted by the extension. Does the extension process the list of voters and substance of the vote (i.e. "Support" or "Oppose") independently so that anyone could not know who voted for? I hope the answer is "yes".
- Unfortunately the answer to that is no. It is encrypted in the database (with no one but an outside party having the decrypt key during the vote and a very limited set of people after) but IF you have access both to the full database AND the decryption key you could link everything with effort. You would need both and could not do it accidentally but you could do it with the current system. We're not very happy with it to be honest (for many reasons) and so I think looking at other systems is certainly a possibility. The reason for the linking in the end is because we need to know what votes to strike and so the information (which includes that is looked at by the election committee to look for fraud or other concerns (and does not link to any specific vote info through the interface) does have some behind the scenes database connections to the separate table holding encrypted votes. If we have the requirement I to be able to strike votes I don't know of a reasonable way to not have that linking somewhere (which inevitably has a few specially trusted engineers with access) but others know that work much better then I do. Jalexander (talk) 00:53, 19 July 2013 (UTC)
Q2: Foundation processes sensitive private informations in identification required to candidates for checkusers/oversighters/stewards. Does the foundation have any plan to provide PGP public key to encrypt the data? This would reduce the risk in case that email servers are compromised, email service providers participated in PRISM project, or email are unintentionally intercepted by someone another. Best regards. – Kwj2772 (msg) 19:29, 6 July 2013 (UTC)
- PGP is currently available on request. I suppose we could expose that more clearly. We're also in the process of revamping how that whole process is handled, such that it will no longer involve email. Philippe (WMF) (talk) 01:03, 19 July 2013 (UTC)
Wording change to "User contributions"
I don't think this section of the policy clearly conveys that your page editing history is public information. Quoting:
User contributions are also aggregated and publicly available. User contributions are aggregated according to their registration and login status. Data on user contributions, such as the times at which users edited and the number of edits they have made, are publicly available via user contributions lists, and in aggregated forms published by other users.
The language makes it sound like only mundane data is collected. But the most salient fact is that there is a permanent, public record of what pages you edit. This record can reveal a focus of interests, implying likely facts about your politics, sexuality, passion for Justin Bieber, or other things you might not want in the public record.
In addition, the repeated use of the word 'aggregated' could cause people to misread and not realize contributions are also connected to individuals. This is especially relevant to users who choose to reveal their real life identities: they need to understand it's not just a matter of having your name on your user page, but that everything they do on-wiki will be traceable to their identity.
Let's remember that things like "page history" and "contribution history" are abstract ideas that take some time for new users to conceptualize. I pick this stuff up pretty fast but I certainly made many edits before understanding what was happening.
Suggestion: in the first sentence change 'aggregated' to 'logged'. In the third sentence change it to "the times at which users edited a particular page". Fletcher (talk) 02:40, 12 July 2013 (UTC)
Technical and legal coercion aspects
- (It should be "Technical and oversight" but anyway...)
- Summary of legal/technical landscape: Recent events show that laws may exist that coerce (any) organization to 1/ secretly change or allow customizing of their hardware and software to allow certain kinds of privacy breach, with 2/ coerced mandatory silence over such changes. A simple solution would be a WMF policy of mandatory swearing under perjury that no such activity has occurred or is occurring, by all senior WMF staff, but 3/ (as shown by other laws such as SOPA) it's quite possible for a law or court to rule that such staff must swear untruthfully and will have legal immunity if they do, so perjury law isn't likely to protect us. Also 4/ data centers overseas could be affected by coercive laws without US staff knowledge, and 5/ there is considerable interest in 'hacking' as a way to gain information covertly (recent news indicates that both China and US governments are willing to hack into reputable private organizations; others may as well). Finally 5/ a risk of silent interception (eg tapping telecoms cables) wouldn't be known and is beyond WMF ability to prevent.
I'm not technical enough to know what's "enough", but I can list the kinds of ideas I would like to see at a technical and policy level. As stated they are extreme, but it's hard to know what is needed. Perhaps while these are extreme someone else can use them to spark discussion of less extreme but equally effective policies:
- Enforced systems transparency - Some kind of systems approach whereby any modifications to our running systems is impossible without causing a log entry or difference to be visible, and where it's designed so that 'silent tampering' with the logging or change management system would not be possible. Alternatively some kind of hypervisor that runs on all systems, with an in-house module that monitors and integrity checks any code running, and which cannot be silently tampered or bypassed.
- Public eyeballs on core systems - Some kind of user group whereby anyone worldwide who wants, can monitor and assure themselves the above is operating correctly. This would be radically transparent, effectively it means allowing anyone worldwide who wishes it, to have absolute and total read-only oversight of core systems and running code (but not data pages or data files).
- Segregation of privacy data from everything else onto separate systems and servers - Redesign of code so that most systems do not handle actual private data - they handle tokens, or the like. The private data is then held separately and only stored and processed in a very few systems, which are hardened so that manipulation is difficult and private data can only leave the systems in a logged manner; remote managed by WMF but with overview of any remote management activity, and located/replicated in countries that do not allow for silent legal coercion. Examples -
- Perhaps core systems don't hold and cannot send out private data, and this can only be done by custom servers upon validation of a legitimate request using multiple replicated systems in two or more countries (or a consensus of replicated systems). This would mean that a "private data handling server" won't allow access to private data unless 2 or more servers in different countries validate (or agree) the legitimacy of the request (hard to tamper with servers outside any given country), and any send is then irrevocably logged before being sent directly to the requestor.
- Perhaps all SQUIDS (WMF proxy servers) whose task is fairly simple and uniform, might be modded to strip any private data and 'do something else with it' so that it cannot subsequently be leaked.
- Tampering while offline or during change - What policy would mitigate against tampering or physical privacy breaching approaches at our datacenters? Should it be policy that any time a core system is taken down or modified (in any country), at least two technical volunteers from a different country should be physically present and perform any core code change, and that any changed systems code is integrity checked against an "official" version of code that's internationally available? Should technical staff in residence where a datacenter exists, always include volunteers in residence from an overseas country, some of whom also act as privacy representatives for an annual term, who must be present for any hardware or systems changes, or offline actions? (The rationale being they are productively active in a technical capacity, but will not be bound to silence by that country's law once back overseas after their year).
- Tamper resistant hardware/firmware - What mitigation exists against bringing in tampered hardware or hardware with tampered on-board modules, for example? Should WMF have a policy of only using TPM hardware for its systems, that enforces the (known) approved firmware, OS and software only? Can we mitigate the scope (under coercion or otherwise) of installing servers with modified on-board cards or firmware to bypass systems controls? (A logical, predictable, and not too difficult 'next step' in how privacy breach might occur)
- Move to enforced HTTPS (possibly with 'allow http' as a user preference) - almost no browsers and no ISPs or systems today, lack support for encrypted connections. Has the time come to enforce https as our sole protocol, or have "allow http" as a user preference that is disabled by default? That would be an immensely powerful statement and technically quite feasible; it's also the only protection we can offer our users against optical/data cable tapping.
- (On the same point, does SSL technically allow us to notify users if SSL privacy is dubious, by checking the certificate chain our end, as users won't otherwise know if MITM interception may be occurring? Or provide transparent SSL proxy servers as entrypoints across a number of countries whose certificate chain is more directly verifiable.)
- Staff terms of service/resignation policy? - Assuming that tampering would involve consent (however reluctant) of senior/board staff, should WMF adopt a policy that any member of staff should resign if their role would require them to implement unauthorized systems? It makes it a bit harder to tamper with systems if anyone asked to assist would by agreement resign or request a role change/removal of authority.
- Privacy committee to focus on this as a living issue and improve approaches over time - Can we have an independent privacy committee, which is not from one country or continent, and where persons in countries with known government coercion capacity do not have "the final say" (making it harder to silently coerce decisions), and whose remit is to proactively consider how to improve the tamper resistant and privacy breach resistant nature of all WMF systems, and the transparent/verifiable nature of its core systems and software? (We have no commercial IP secrets so enforcing systems transparency and tamper resistance at the heart, is a good idea)
Last, if we are seen to put the work in to do it (and we communicate this project impactfully in public media), perhaps it will have wider effect. It may encourage change-oriented debate and be more persuasive, that this is important and action is possible. If our example prompts wider awareness and change, then it will bear fruit beyond just our sites and on the global web, in ways that would be unlikely via mere protest. FT2 (Talk | email) 03:38, 17 July 2013 (UTC)
I want to have cookies that last for more than 30 days, because I want to stay logged in for longer than 30 days. WhatamIdoing (talk) 22:26, 19 July 2013 (UTC)
- Out of curiosity, why is the Wikimedia expiration date (30 days) different from the MediaWiki default (180 days)? --Michaeldsuarez (talk) 23:21, 21 July 2013 (UTC)
- The answer to this question is available on this very page. odder (talk) 23:23, 21 July 2013 (UTC)
- Oh, I see now. Thanks. --Michaeldsuarez (talk) 23:30, 21 July 2013 (UTC)
- The above discussion is preserved as an archive. Please do not modify it. Subsequent comments should be made in a new section.