Bits from past DPL

From	Andreas Tille <tille@debian.org>
Newsgroups	linux.debian.announce.devel
Subject	Bits from past DPL
Date	2026-05-06 10:30 +0200
Message-ID	<MRFNM-2YGO-3@gated-at.bofh.it> (permalink)
Organization	linux.* mail to news gateway

Show all headers | View raw

[Multipart message — attachments visible in raw view] - view raw

Dear Debian community,

This is Bits from a past DPL.

1. Congratulations to Sruthi
============================

I'm glad to see that so many Debian Developers have placed their trust
in Sruthi. I share the view that she will do an excellent job in this
role.

Dear Sruthi, I wish you all the best for your term as Debian Project
Leader.

Fun fact: Indian media reported on the new DPL with a delay of about two
days[c01]. By coincidence, I saw this while a German TV crew was
recording my screen--over two years after I had started as DPL, and after
my term had already ended.

[c01] https://www.youtube.com/watch?v=D2079IJSx6g


2. Looking back
===============

Different people will remember my tenure for different reasons. Some may
associate it with the creation of the DFSG team and the introduction of
an annual review of delegations[b01]. Others gave feedback they
appreciated how certain difficult situations were handled along the
way[b02].

On a more personal note, I'm pleased that during this time the publicity
team decided to stop posting on X[b03]. That felt like a step aligned with
Debian's values.

There were also several important developments in Debian during this
period, such as the adoption of tag2upload and the release of Trixie.
These achievements were driven by the work of many contributors across
the project. My role in them was limited, but I want to acknowledge and
thank everyone involved.

One topic I cared about was improving diversity in Debian, particularly
regarding gender and geographic representation. This is not an area
where results can easily be measured, and change tends to happen slowly.

Seeing a woman from India as the next DPL is an encouraging development.
I have confidence that this perspective will help keep the topic on the
agenda.

Not everything I had hoped for could be completed. In particular, I
would have liked to see progress on improving the MIA process.

And while it did not quite happen during my tenure, I still look forward
to the day when we might announce Debian-powered robots on the
moon.[b04]

After looking back, you might wonder what my future involvement in
Debian will look like. That question is easy to answer: I have promised
the Debian Med team that I will return to working with them--and that is
exactly what I intend to do.

[b01] https://lists.debian.org/debian-devel/2026/04/msg00235.html
[b02] https://lists.debian.org/debian-devel/2026/01/msg00392.html
[b03] https://lists.debian.org/debian-devel-announce/2025/02/msg00001.html
[b04] https://www.federico.io/pdf/DeLaCroix.Rossi.ea.AERO24.pdf


3. Some statistics
==================

3.1. Bug of the Day
-------------------

One initiative during my tenure was the "Bug of the Day" project[s01].
The goal was to provide approachable entry points for newcomers, and I
have received some indications that this worked in individual cases.

At the same time, the effort also served as a QA and integration
activity. Since 2024-08-01, more than 600 packages have been processed.
These packages were migrated from various version control systems--many
still referencing Alioth, but also other VCS or none at all--to Salsa.

For each package, at least one bug was closed. In addition, we evaluated
whether packages are still needed, or whether removal would be a better
use of limited maintainer time.


3.2. Salsa Migration of Teams
-----------------------------

In January 2025, it was observed that a significant number of
team-maintained packages were still not (visibly) hosted on Salsa[s02].
At that time, more than 600 packages still referenced Alioth in their
Vcs fields, despite Alioth having been shut down years earlier. This did
not even include packages with missing Vcs fields, which likely added
another ~100 cases.

I worked with the respective teams to improve this situation. Today, the
picture looks very different.

For illustration, the following query (expanded to also include missing
Vcs fields and excluding known exceptions) now returns only very few
remaining cases:

SELECT maintainer_name, COUNT(*)
FROM ( SELECT DISTINCT ON (source) * FROM sources WHERE release = 'sid' ORDER BY source, version DESC ) s
WHERE
    (vcs_url ~* '/(git|svn|alioth|anonscm).debian.org' or vcs_url is null or vcs_url = '')
    AND maintainer ~* '(team|group|lists|force|maintainers)'
    AND extra_source_only IS NOT True
    AND maintainer_name not in ('PCP Development Team', -- maintain Debian packaging in the upstream PCP git repo
                                'Linaro Packagers', -- Uploader declared to prefer staying outside Salsa
                                'GRUB Maintainers', 'Debian EFI team', -- auto generated source
                                'XFS Development Team' -- maintained by upstream at git.kernel.org
                               )
    AND s.source NOT IN ( -- No packages scheduled for removal
              SELECT affected_sources FROM bugs WHERE bugs.source = 'ftp.debian.org' AND title LIKE 'RM:%' AND done = '' -- do not include packages with removals pending
    )
GROUP BY maintainer_name
ORDER BY count DESC;
        maintainer_name         | count 
--------------------------------+-------
 Debian LibreOffice Maintainers |     3
 Debian GCC Maintainers         |     2
 Debian Mirror Team             |     1
 Debian VSquare Team            |     1
 Debian Wine Party              |     1
 Debian Mactel                  |     1
 Ayatana Packagers              |     1
(7 rows)

The result currently lists only seven teams with a total of ten
affected packages. This suggests that the issue has been largely
resolved.  This reduces friction for collaboration and lowers the
barrier for contributions across teams.


3.3. More numbers
-----------------

Some of the goals I set at the beginning of my term can be illustrated
with figures based on numbers from trends.debian.net.

The number of packages not hosted on Salsa has decreased significantly:

 * 2024-03-01: 4801
 * 2026-03-01: 2083

At the time of writing, this number has dropped below 2000, which was
the target I had set for the end of my term.

Looking at version control more broadly, the situation has also
improved. Today, fewer than ten packages are maintained in a version
control system other than Git, and roughly 1000 packages still lack any
VCS. I intend to continue checking whether maintainers actively prefer
alternative approaches or whether support for migration would be helpful
to enable easier collaboration.

Beyond hosting, there has also been progress in simplifying and unifying
packaging tooling. For example, the use of cdbs has effectively been
phased out (from 1571 packages on 2024-03-01 to none today). Similarly,
packages using old long debhelper have been reduced from 891 to around
500.

Moving towards a smaller set of well-understood and actively maintained
tools helps reduce complexity and makes it easier to adapt Debian
packaging to future needs.


[s01] https://salsa.debian.org/qa/tiny_qa_tools/-/wikis/Tiny-QA-tasks#bug-of-the-day
[s02] https://lists.debian.org/debian-devel/2025/01/msg00148.html



4. Conclusions
==============

The past two years have been a valuable experience for me, both within
Debian and beyond (confirmed by my wife ;-) ).

One thing I learned is that it is worth tackling tasks even if they
initially seem too large. Progress may be gradual, but it is often
possible with persistence and support from others. In that regard, I
have also learned that I can rely on my fellow Debian Developers. The
willingness to collaborate, to step in, and to move things forward
together remains one of Debian’s strongest qualities.

On a personal level, I learned that changes in role can reveal a lot
about how we relate to each other. I came to particularly value those
who treat others consistently, independent of any position or title.

A reassuring takeaway for me is seeing how projects continue to thrive
beyond individual involvement. Debian Med, which has been close to me
for most of my time in Debian, continues to progress without my direct
participation.  The same applies to work in the R packaging team.
Stepping back can create space for others to contribute, and that is a
healthy and necessary dynamic in a volunteer-driven project.

Serving as DPL also reinforced that the role requires both dedication
and a clear vision. Equally important is communicating that vision in a
way that others can engage with and build upon.

Thank you to everyone who contributed, collaborated, disagreed
constructively, and kept Debian moving forward. These two years have
enriched my life, and a major reason for that has been the cooperation
within this community.


Kind regards
    Andreas.

-- 
Everything must change so that everything can stay the same.
	-- Giuseppe Tomasi di Lampedusa (The Leopard)

Back to linux.debian.announce.devel | Previous | Next | Find similar

Thread

Bits from past DPL Andreas Tille <tille@debian.org> - 2026-05-06 10:30 +0200

csiph-web