[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#900552: [PATCH-ish] Add privacy policy to the Debian website



Package: www.debian.org

Hi.

This is not a complete patch as I originally intended to provide, but in
the interest of getting something up sooner rather than later, and
accepting there will be changes either way, I'm submitting it now.

We need a privacy policy on our website, describing the personal data we
hold for people and what we do with it. This has become pressing due to
the EU GDPR, but is something we generally should have anyway. I've
attached a start (in wml, which I hope is helpful, though it's basically
HTML and I originally wrote it in Markdown). I do expect there to be
updates as service owners flesh things out, but I'm hoping once the
initial page and structure is in place that can be done as proper
patches.

One thing I couldn't figure out is /where/ this should live. It should
definitely be linked from our front page. I thought under /legal/ might
be a good place, but nothing else there is linked from the front page
that I can see. I leave it to the wisdom of the web team to work out
where might be best.

(There's hopefully a data protection team about to be created which will
need some more updates to the site, but I'll send those once it actually
exists.)

J.

(Well done, sigmonster.)

-- 
... Why do I get the feeling I'm going to regret this?
#use wml::debian::template title="Privacy Policy" NOCOMMENTS="yes"

<p>The <a href="https://www.debian.org/";>Debian Project</a> is a volunteer
association of individuals who have made common cause to create a free
operating system, referred to as Debian. There is no requirement for anyone who
wishes to use Debian to provide the project with any personal information; it
is freely downloadable without registration or other form of identification
from both official mirrors run by the project and numerous third parties.</p>

<p>Various other aspects of interacting with the Debian Project will, however,
involve the collection of personal information. This is primarily in the form
of names and email addresses in emails received by the project; all Debian
mailing lists are publicly archived, as are all interactions with the bug
tracking system. This is in keeping with our <a
href="https://www.debian.org/social_contract";>Social Contract</a>, in
particular our statement that we will give back to the free software community
(#2), and that we will not hide our problems (#3). We do not perform further
processing on any of the information we hold, but there are instances where it
is automatically shared with third parties (such as emails to lists, or
interactions with the bug tracking system).</p>

<p>The list below categorises the various services run by the project, the
information used by those services and the reasons it is required.</p>

<p>Please note that hosts and services under the debian.net domain are not part
of the official Debian project; they are run by individuals who have an
association with the project rather than the project themselves. Questions
about exactly what data those services hold should be directed at the service
owners rather than the Debian Project itself.</p>

<h2>Contributors (<a href="https://contributors.debian.org/";>contributors.debian.org</a>)</h1>

<p>The Debian Contributors site provides an aggregation of data about where
someone has contributed to the Debian Project, whether that's through filing a
bug report, making an upload to the archive, posting to a mailing list or
various other interactions with the Project. It receives its information from
the services in question (details about an identifier such as login name and
time of last contribution) and provides a single reference point to see where
the Project is storing information about an individual.</p>

<h2>The Archive (<a href="https://ftp.debian.org/debian/";>ftp.debian.org</a>)</h1>

<p>The primary distribution method of Debian is via its public archive network.
The archive consists of all of the binary packages and their associated source
code, which will include personal information in the form of names and email
addresses stored as part of changelogs, copyright information, and general
documentation. The majority of this information is provided via the upstream
software authors distributed source code, with Debian adding additional
information to track authorship and copyright to ensure that licenses are being
correctly documents and the Debian Free Software Guidelines adhered to.</p>

<h2>Bug Tracking System (<a href="https://bugs.debian.org/";>bugs.debian.org</a>)</h1>

<p>The bug tracking system is interacted with via email, and stores all emails
received in relation to a bug as part of that bug's history. In order that the
project can effectively deal with issues found in the distribution, and to
enable users to see details about those issues and whether a fix or workaround
is available, the entirety of the bug tracking system is openly accessible.
Therefore any information, including names and email addresses as part of email
headers, sent to the BTS will be archived and publicly available.</p>

<h2>DebConf (<a href="https://www.debconf.org/";>debconf.org</a>)</h1>

<p>The DebConf registration structure stores the details of conference
attendees. These are required to determine eligibility to bursarys, association
to the project, and to contact attendees with appropriate details. They may
also be shared with suppliers to the conference, e.g. attendees staying in the
conference provided accommodation will have their name and attendance date
shared with the accommodation provider.</p>

<h2>Developers LDAP (<a href="https://db.debian.org";>db.debian.org</a>)</h1>

<p>Project contributors (developers and others with guest accounts) who have
account access to machines within the Debian infrastructure have their details
stored within the project's LDAP infrastructure. This primarily stores name,
username and authentication information. However it also has the optional
facility for contributors to provide additional information such as address or
phone details. These are only shared with other individuals who have account
access to the Debian infrastructure and is intended to provide a centralised
location for project members to exchange such contact information. It is not
explicitly collected at any point and can always be removed by logging into the
db.debian.org web interface or sending signed email to the email interface.</p>

<h2>Gitlab (<a href="https://salsa.debian.org/";>salsa.debian.org</a>)</h1>

<p>salsa.debian.org provides an instance of the <a
href="https://about.gitlab.com/";>GitLab</a> DevOps lifecycle management tool.
It is primarily used by the Project to allow Project contributors to host
software repositories using Git and encourage collaboration between
contributors. As a result it requires various pieces of personal information to
manage accounts. For Project members this is tied to the central Debian LDAP
system, but guests may also register for an account and will have to provide
name and email details in order to facilitate the setup and use of that
account.</p>

<p>Due to the technical nature of git contributions to the git repositories
held on salsa will contain the name and email address recorded within those git
commits. The chained nature of the git system means that any modification to
these commit details once they are incorporated into the repository is
extremely disruptive and in some cases (such as when signed commits are in use)
impossible.</p>

<h2>Gobby (<a href="https://gobby.debian.org/";>gobby.debian.org</a>)</h1>

<p>Gobby is a collaborative online text editor, which tracks contributions and
changes against connected users. No authentication is required to connect to
the system and users may choose any username they wish. However while no
attempt is made by the service to track who owns usernames it should be
understand that it may prove possible to map usernames back to individuals
based upon common use of that username or the content they post to a
collaborative document within the system.</p>

<h2>Mailing Lists (<a href="https://lists.debian.org/";>lists.debian.org</a>)</h1>

<p>Mailing lists are the primary communication mechanism of the Debian Project.
Almost all of the mailing lists related to the project are open, and thus
available for anyone to read and/or post to. All lists are also archived; for
public lists this means in a web accessible manner. This fulfils the project
commitment to transparency, and aids with helping our users and developers
understand what is happening in the project, or understand the historical
reasons for certain aspects of the project. Due to the nature of email these
archives will therefore potentially hold personal information, such as names
and email addresses.</p>

<h2>New Members site (<a href="https://nm.debian.org/";>nm.debian.org</a>)</h1>

<p>Contributors to the Debian Project who wish to formalise their involvement
may choose to apply to the New Members process. This allows them to gain the
ability to upload their own packages (via Debian Maintainership) or to become
full voting members of the Project with account rights (Debian Developers, in
uploading and non-uploading variants). As part of this process various personal
details are collected, starting with name, email address and
encryption/signature key details. Full Project applications also involve the
applicant engaging with an Application Manager who will undertake an email
conversation to ensure the New Member understands the principles behind Debian
and has the appropriate skills to interact with the Project infrastructure.
This email conversation is archived and available to the applicant and
Application Managers via the nm.debian.org interface. Additionally details of
outstanding applicants are publicly visible on the site, allowing anyone to see
the state of New Member processing within the Project to ensure an appropriate
level of transparency.</p>

<h2>Popularity Contest (<a href="https://popcon.debian.org/";>popcon.debian.org</a>)</h1>

<p>"popcon" tracks which packages are installed on a Debian system, to enable
the gathering of statistics about which packages are widely used and which are
no longer in use. It uses the optional "popularity-contest" package to collect
this information, requiring explicit opt-in to do so. This provides useful
guidance about where to devote developer resources, for example when migrating
to newer library versions and having to spend effort on porting older
applications. Each popcon instance generates a random 128 bit unique ID which
is used to track submissions from the same host. No attempt is made to map this
to an individual about submissions are made via email or HTTP and it is thus
possible for personal information to leak in the form of the IP address used
for access or email headers. This information is only available to the Debian
System Administrators and popcon admins; all such meta-data is removed before
submissions are made accessible to the project as a whole. However users should
be aware that unique signatures of packages (such as locally created packages
or packages with very low install counts) may make machines deducible as
belonging to particular individuals.</p>

<p>Raw submissions are stored for 24 hours, to allow replaying in the event of
issues with the processing mechanisms. Anonymized submissions are kept for at
most 20 days. Summary reports, which contain no personally identifiable
information, are kept indefinitely.</p>

<h2>snapshot (<a href="http://snapshot.debian.org/";>snapshot.debian.org</a>)</h1>

<p>The snapshot archive provides a historical view of the Debian archive
(ftp.debian.org above), allowing access to old packages based on dates and
version numbers. It carries no additional information over the main archive
(and can thus contain personal information in the form of names + email address
within changelogs, copyright statements and other documentation), but can
contain packages that are no longer part of shipping Debian releases. This
provides a useful resource to developers and users when tracking down
regressions in software packages, or providing a specific environment to run a
particular application.</p>

<h2>Votes (<a href="https://vote.debian.org/";>vote.debian.org</a>)</h1>

<p>The vote tracking system (devotee) tracks the status of ongoing General
Resolutions and the results of previous votes. In the majority of cases this
means that once the voting period is over details of who voted (usernames +
name mapping) and how they voted becomes publicly viewable. Only Project
members are valid voters for the purposes of devotee, and only valid votes are
tracked by the system.</p>

<h2>Wiki (<a href="https://wiki.debian.org/";>wiki.debian.org</a>)</h1>

<p>The Debian Wiki provides a support and documentation resource for the
Project which is editable by everyone. As part of that contributions are
tracked over time and associated with user accounts on the wiki; each
modification to a page is tracked to allow for errant edits to be reverted and
updated information to be easily examined. This tracking provides details of
the user responsible for the change, which can be used to prevent abuse by
blocking abusive users or IP addresses from making edits. User accounts also
allow users to subscribe to pages to watch for changes, or see details of
changes throughout the entire wiki since they last checked. In general user
accounts are named after the name of the user, but no validation is performed
of the account names and a user may choose any free account name. An email
address is required for the purposes of providing a mechanism for account
password reset, and notifying the user of any changes on pages they are
subscribed to.</p>

<h2>Echelon</h1>

<p>Echelon is a system used by the Project to track member activity; in
particular it watches the mailing list and archive infrastructures, looking for
posts and uploads to record that a Debian member is active. Only the most
recent activity is stored, in the member's LDAP record. It is thus limited to
only tracking details of individuals who have accounts within the Debian
infrastructure. This information is used when determining if a project member
is inactive or missing and thus that there might be an operational requirement
to lock their account or otherwise reduce their access permissions to ensure
Debian systems are kept secure.</p>

<h2>Service related logging</h1>

<p>In addition to the explicitly listed services above the Debian
infrastructure logs details about system accesses for the purposes of ensuring
service availability and reliability, and to enable debugging and diagnosis of
issues when they arise. This logging includes details of mails sent/received
through Debian infrastructure, web page access requests sent to Debian
infrastructure, and login information for Debian systems (such as SSH logins to
project machines). None of this information is used for any purposes other than
operational requirements and it is only stored for 15 days in the case of web
server logs, 10 days in the case of mail log and 4 weeks in the case of
authentication/ssh logs.</p>

Attachment: signature.asc
Description: PGP signature


Reply to: