• Creating a Debian Spending proposals and discussion mailing list

    From Phil Morrell@21:1/5 to Philip Hands on Fri Apr 2 23:30:01 2021
    Hi all, moving off -vote and heavily trimming, original thread here:

    https://lists.debian.org/debian-vote/2021/03/msg00007.html

    On Thu, 18 Mar 2021, Philip Hands wrote:
    On Thu, 18 Mar 2021, Jonathan Carter wrote:
    I think as things stand now, every DD pretty much already has the entire Debian budget available at their disposal if they can think of a way to spend it that benefits the project.

    I was reflecting on the way the Peertube funding was achieved.
    There were enough people keen on that happening that if we'd each had an earmarked e.g. 1k budget to allocate, we could have just agreed it
    amongst ourselves, and done it

    On Sun, 21 Mar 2021, Raphael Hertzog wrote:
    On Thu, 18 Mar 2021, Philip Hands wrote:
    On Thu, 18 Mar 2021, Raphael Hertzog wrote:
    I'm considering paying someone to identify useful
    projects. That person could talk to various teams, make proposals based on
    their own experience, and even run a poll among Debian developers.

    I've been pondering how it might be possible to spend more of Debian's money, and it occurred to me that we could allocate a budget to each DD which they could spend on pretty-much anything

    Encouraging people to pool their budgets to fund bigger things would hopefully result in them forming teams of mentors to oversee the work.

    I really like your idea! I wonder if there would be some infrastructure
    that would make it easy to describe projects and track how much money
    has been allocated by the various DD.

    But if we find something usable, we could have a volunteer in charge of entering the "votes" of the DD by adjusting a Debian pledge in a open
    system (and have some associated ledger where the DD allocations are tracked).

    This whole idea inspired me, so if just a couple of DDs agree with the
    below, I'm hereby volunteering to do the admin work involved.

    I've thought about what such a system could look like, perhaps signed
    commits to a salsa project or a simple site like mentors. I came to the conclusion that there's already a working system in place for counting
    DD support of suggestions. debian-vote has proposals, low-bureaucracy seconders, and the Project Secretary validating signatures.

    I propose creating an experimental debian-spending mailing list based on
    the same rules to test this idea. The equivalent of the GR here would be
    a filled out project proposal for consideration by the DPL or Freexian.
    I'd have to do more research on the process first, since I haven't
    interacted with the voting system myself. Taking the peertube example:

    Alice: Proposal: Video team needs peertube v3 to support streaming
    Bob: Amendment: Donate €20,000 to the public fundraising campaign
    Carol: Seconded: Bob
    Dan?: Seconded: Bob
    Erin: Amendment: Donate €10,000 to the public fundraising campaign
    emorrp1: Dan you forgot to sign
    Frank: Seconded: Erin
    Dan: Seconded: Bob
    ... etc.

    I would then send an email to the DPL (or fill out the Freexian
    template) with the project details and the list of signatories. The
    outcome could be absolutely anything of course - the DPL could see that
    30 people supported 10k and 10 people 20k, but they know that the budget
    can handle 15k and do that, the point is to have a concrete proposal.

    Another example is that debian-android-tools has all the DD availability
    for sponsoring Kotlin uploads, and most initial work done by GSoC
    students, but no-one had free time to work on it betwen Oct and Mar. Now
    it turns out that no less than 3 other teams are depending/awaiting a
    kotlin upload: Java, Freedombox for Jitsi, Games for Mindustry. That's a
    lot of potential DDs who could have seconded a tender for a third-party contractor to bid on say a week to a month's work.

    I think this will lower the barrier for proposals. I looked up what the
    current process is and it's literally "email the DPL and convince them",
    which could be offputting without knowing how many other people support
    your idea. Similarly I would expect a lot of teams to know their own
    problem areas but be unaware of the level of support outside the team
    through multiple levels of reverse dependencies.

    Hopefully that's enough to convince you, and it's clear I'm only
    offering to do the administration, not make any actual decisions. I'd
    also like to help the list get started in a personal capacity by digging
    up old BudgetIdeas and writing up proposals say once a week.

    This doesn't block implementing the idea of a per-DD budget later. Say, €100pm and the Seconders include how many months of their allocation
    they wish to spend on a project. I could include the total in the email
    to DPL, or feedback to the list that a project has reached 50% funding
    and needs more Seconders, or if they've exceeded the budget just include
    them in the signatories with a sum of 0.
    --
    Phil Morrell (emorrp1)

    -----BEGIN PGP SIGNATURE-----

    iHUEABYKAB0WIQSBP39/Unco6Ai78+TbymUJHySObAUCYGeHvgAKCRDbymUJHySO bNoqAQCoJolOY2ymKdPkyyurvPfFmrgx8r92VPH/4JrcVRGa8gEA1l8SPYsIs1Fe jXXhoXim5CY1BTIdIWcTA378AHsigwk=
    =wiEU
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Raphael Hertzog@21:1/5 to Phil Morrell on Sun Apr 4 12:10:02 2021
    Hi,

    On Fri, 02 Apr 2021, Phil Morrell wrote:
    I've thought about what such a system could look like, perhaps signed
    commits to a salsa project or a simple site like mentors. I came to the conclusion that there's already a working system in place for counting
    DD support of suggestions. debian-vote has proposals, low-bureaucracy seconders, and the Project Secretary validating signatures.

    I propose creating an experimental debian-spending mailing list based on
    the same rules to test this idea. The equivalent of the GR here would be
    a filled out project proposal for consideration by the DPL or Freexian.

    That's a simple way to ensure some basic level of support to an idea
    or some way to gauge whether a need is shared by multiple persons.

    But it doesn't really help to turn those ideas into actionable projects
    that can be funded.

    It's OK when the project is as simple as "give 5 KEUR to peertube" but definitely not enough when the idea is "build a web interface to manage
    Debian reimbursement request" (taking an example that I just got on https://salsa.debian.org/freexian-team/project-funding/-/merge_requests/5).

    I would be saddened if this system turned only into a way to give our
    money to other free software projects instead of using that money to help
    us towards our common mission.

    Another example is that debian-android-tools has all the DD availability
    for sponsoring Kotlin uploads, and most initial work done by GSoC
    students, but no-one had free time to work on it betwen Oct and Mar. Now
    it turns out that no less than 3 other teams are depending/awaiting a
    kotlin upload: Java, Freedombox for Jitsi, Games for Mindustry. That's a
    lot of potential DDs who could have seconded a tender for a third-party contractor to bid on say a week to a month's work.

    Ack.

    I think this will lower the barrier for proposals. I looked up what the current process is and it's literally "email the DPL and convince them", which could be offputting without knowing how many other people support
    your idea. Similarly I would expect a lot of teams to know their own
    problem areas but be unaware of the level of support outside the team
    through multiple levels of reverse dependencies.

    It's certainly good to lower the barrier for proposals but for your Kotlin example, the issue is more "who will be paid to to the work"? Someone has to select a winning bid and having that kind of responsibility within Debian
    is the historical friction point related to use of money in Debian.

    I tried to solve this by defining up-front for each project who will be
    the reviewer and thus the person who will have to select the winning bid
    (in cooperation with the person/organization who will pay the bill). Or by encouraging developers to propose projects that they want to implement themselves... so that there's no choice of person to make, it's more a "do we want this project at this time" and it can be answered collectively
    (in Freexian's case by the paid LTS contributors since it's money saved
    from the offering where they are involved).

    Hopefully that's enough to convince you, and it's clear I'm only
    offering to do the administration, not make any actual decisions. I'd
    also like to help the list get started in a personal capacity by digging
    up old BudgetIdeas and writing up proposals say once a week.

    IMO the bulk of the work is not in finding ideas, but on transforming
    them into actionable projects, and on selecting which project can have the largest impact on Debian.

    That said I would welcome if we could query the Debian developers at
    large to find out what kind of ideas/changes would be most beneficial
    to Debian.

    Cheers,
    --
    ⢀⣴⠾⠻⢶⣦⠀ Raphaël Hertzog <hertzog@debian.org>
    ⣾⠁⢠⠒⠀⣿⡁
    ⢿⡄⠘⠷⠚⠋ The Debian Handbook: https://debian-handbook.info/get/
    ⠈⠳⣄⠀⠀⠀⠀ Debian Long Term Support: https://deb.li/LTS

    -----BEGIN PGP SIGNATURE-----
    Comment: Signed by Raphael Hertzog

    iQEzBAABCgAdFiEE1823g1EQnhJ1LsbSA4gdq+vCmrkFAmBpjmEACgkQA4gdq+vC mrmMgAgAuUBxLPqarTnKgHHt0l53VqUwT6JRgVpVu+NOgJrYxP74umrfzop/i4PB wTDEIF3Yu+6h/CvAAXHaJP+rij8VkdmnkO4HKVic5/R7JaJwOR69qxmPUtMCOYOV StYKgRu3lQuQp3HYqz3a1fmJ31dr6OwMWgXnxLeLpluPYfNjgHhrFAzFPiD+kTg7 LysQqXdWV28b/Bs6bvIVMULViJntDMabzr9DDrvcLc7TbN1jPcEC2e6hEMv3LDTc 9ymCoUHmn0DO+a8tWBClRXwJUWrEbZM8tvA2sc+3SRyeAjVK00gkc0yjGW1b3vFm OFYd+NN3TZ6ZfA+LAUgw1ISNgf76Ng==
    =Stes
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Phil Morrell@21:1/5 to Raphael Hertzog on Sun Apr 4 17:00:01 2021
    Hi Raphal, your feedback in particular is very much appreciated.

    Please keep in mind that I'm proposing this list purely as a practical experiment, it does nothing that can't already be done elsewhere, and if
    it doesn't work out after say 6 months, then so be it. All I'm looking
    for is an indication that it would not be a complete waste of my time to
    set up, that doing so has the potential to help Debian, and that some
    DDs would be willing to review and Second proposals.

    On Sun, Apr 04, 2021 at 12:01:05PM +0200, Raphael Hertzog wrote:
    I would be saddened if this system turned only into a way to give our
    money to other free software projects instead of using that money to help
    us towards our common mission.

    I'm not so sure about that, there's a lot of overlap involved (e.g. reproducible builds or calamares) especially since the project resists
    funding direct packaging work like QA. Even assuming the initiative accidentally ends up solely donating to other existing projects,
    provided they benefit Debian I'd count that as successful - a contra
    policy guidance or GR could also be proposed if it becomes excessive.

    On Fri, 02 Apr 2021, Phil Morrell wrote:
    I think this will lower the barrier for proposals. I looked up what the current process is and it's literally "email the DPL and convince them", which could be offputting without knowing how many other people support your idea. Similarly I would expect a lot of teams to know their own problem areas but be unaware of the level of support outside the team through multiple levels of reverse dependencies.

    It's certainly good to lower the barrier for proposals but for your Kotlin example, the issue is more "who will be paid to to the work"? Someone has to select a winning bid and having that kind of responsibility within Debian
    is the historical friction point related to use of money in Debian.

    Isn't that the same issue you have for Freexian? Presumably the Proposal
    would be either an Executor or (implied by default) Reviewer by your terminology, so then the Seconds are agreeing who will review the work.

    https://salsa.debian.org/freexian-team/project-funding/-/blob/master/Rules-LTS.md#who-can-make-bids

    IMO the bulk of the work is not in finding ideas, but on transforming
    them into actionable projects

    Exactly my reason for proposing this new list to facilitate fleshing out
    of ideas collectively. Unlike -vote, I'm hoping most Amendments would be accepted by the proposer, so perhaps my Peertube example was unhelpful
    here. There are a lot of current suggestions that the money could be put towards some goal or category, but as you say they're not actionable.

    https://wiki.debian.org/BudgetIdeas

    and on selecting which project can have the largest impact on Debian.

    I think Jonathan's comment (and indeed platform) shows that we're not at
    a point where that is a concern. For now I am happy giving more
    visibility to actionable projects with *any* reasonable impact on
    Debian. This also reminds me of multiple DebConfs sentiments on the bank balance to the effect of: Whenever Debian successfully manages to spend
    money, we're burdened with increased funding to compensate.

    -----BEGIN PGP SIGNATURE-----

    iHUEABYKAB0WIQSBP39/Unco6Ai78+TbymUJHySObAUCYGnDbgAKCRDbymUJHySO bPmcAP93Zbqph//v+wt7sh7xvNMrwDMHZmU4S+lYUPhKkDtHwgD/R7rh6/qXq9tF LgKtHRD4LAigYJj/tHXP0XKnF4WlQgE=
    =jrU9
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Raphael Hertzog@21:1/5 to Phil Morrell on Tue Apr 6 22:20:02 2021
    Hi,

    On Sun, 04 Apr 2021, Phil Morrell wrote:
    Please keep in mind that I'm proposing this list purely as a practical experiment, it does nothing that can't already be done elsewhere, and if
    it doesn't work out after say 6 months, then so be it. All I'm looking
    for is an indication that it would not be a complete waste of my time to
    set up, that doing so has the potential to help Debian, and that some
    DDs would be willing to review and Second proposals.

    I think it's important to experiment in this direction but for a low-key experimentation I'd rather go with a gitlab project where you file ideas
    as issues, people vote up and down various ideas with the usual +1 -1
    buttons (gitlab can then show you a sorted list by popularity). You can
    have draft projects in text files and people can collaborate with MR on enhancement to those drafts.

    We could have a "debian/spending-ideas" if you want so that all DD have
    write access by default. We could restrict access to issues for project
    members (that automatically includes all DD + selected non-DD directly
    added to the project).

    It's certainly good to lower the barrier for proposals but for your Kotlin example, the issue is more "who will be paid to to the work"? Someone has to
    select a winning bid and having that kind of responsibility within Debian is the historical friction point related to use of money in Debian.

    Isn't that the same issue you have for Freexian?

    Well, no. Freexian is not Debian. I said _within Debian_.

    Presumably the Proposal
    would be either an Executor or (implied by default) Reviewer by your terminology, so then the Seconds are agreeing who will review the work.

    If freexian decides to trust a Debian contributor to select a winning bid,
    or trust a Debian contributor to implement a project, it's not the same
    as if the money was given by Debian directly to a Debian contributor.

    Obviously that decision can still have an impact on the community and
    that's why I aim to have a clear and transparent process.

    a point where that is a concern. For now I am happy giving more
    visibility to actionable projects with *any* reasonable impact on
    Debian.

    Ack.

    Cheers,
    --
    ⢀⣴⠾⠻⢶⣦⠀ Raphaël Hertzog <hertzog@debian.org>
    ⣾⠁⢠⠒⠀⣿⡁
    ⢿⡄⠘⠷⠚⠋ The Debian Handbook: https://debian-handbook.info/get/
    ⠈⠳⣄⠀⠀⠀⠀ Debian Long Term Support: https://deb.li/LTS

    -----BEGIN PGP SIGNATURE-----
    Comment: Signed by Raphael Hertzog

    iQEzBAABCgAdFiEE1823g1EQnhJ1LsbSA4gdq+vCmrkFAmBswTkACgkQA4gdq+vC mrnaDQgAoblZXtTEvNG7PuuMEQudaBfHTrJllw9UWYCUIBbDoevbU0zVlbZg4sc/ AQcuPdZwkz42oLhQvz8EkawEhmGw1+vQS1EPj4oqdrDaTjMemXgUSJ6DrtxtWggp XqARhOKkWSdgWzm7BHy6q6dMZerUfD+tEg1Goiq48hntGfYraXVsPEozVoXT9WU7 lY2/wb7I6zAuzkiQwpZrmW8idy0jIUSLNgM2J1R1QZKHBH62loUOBtil8dXf/Di3 QXv788UpwC+o0kwYzKa1KpCnazTEY7bi8rqeQKpSErYIIEgNkwC2Y65rNr8lOCpR N4uJuf7c5KrC9Vm1Kg4njphbIGlTjQ==
    =7DpE
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Phil Morrell@21:1/5 to Raphael Hertzog on Mon Apr 19 19:20:01 2021
    Now that the DPL voting is over, I'd like to ask Jonathan directly what
    you think of this idea in the context of your plans for an Expenditure
    policy? Could this fit alongside, help feed into it or is likely to be
    made obsolete? The thread starts here:

    https://lists.debian.org/debian-project/2021/04/msg00001.html

    On Tue, Apr 06, 2021 at 10:14:50PM +0200, Raphael Hertzog wrote:
    On Sun, 04 Apr 2021, Phil Morrell wrote:
    Please keep in mind that I'm proposing this list purely as a practical experiment, it does nothing that can't already be done elsewhere, and if
    it doesn't work out after say 6 months, then so be it. All I'm looking
    for is an indication that it would not be a complete waste of my time to set up, that doing so has the potential to help Debian, and that some
    DDs would be willing to review and Second proposals.

    I think it's important to experiment in this direction but for a low-key experimentation I'd rather go with a gitlab project where you file ideas
    as issues, people vote up and down various ideas with the usual +1 -1
    buttons (gitlab can then show you a sorted list by popularity). You can
    have draft projects in text files and people can collaborate with MR on enhancement to those drafts.

    We could have a "debian/spending-ideas" if you want so that all DD have
    write access by default. We could restrict access to issues for project members (that automatically includes all DD + selected non-DD directly
    added to the project).

    Understood, I'm happy to organise it that way too if folks would prefer.
    It just *seems to me* that the email workflow of seconds and inline
    quoting is less structured and very familiar to DDs for fleshing out an
    idea, perhaps augmented with an ad-hoc jitsi call.
    --
    Phil Morrell (emorrp1)

    -----BEGIN PGP SIGNATURE-----

    iHUEABYKAB0WIQSBP39/Unco6Ai78+TbymUJHySObAUCYH2tGQAKCRDbymUJHySO bDKZAQC60z2nwGBdAdKnK+IjwEkMbIqBWGOEPsHB6LLB5ivCywEAni74tCZb4OQZ VcHbudzwdLYovLbrd0uiK71VDA7RowU=
    =+LNs
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Raphael Hertzog@21:1/5 to Phil Morrell on Mon Apr 19 21:10:03 2021
    On Mon, 19 Apr 2021, Phil Morrell wrote:
    On Tue, Apr 06, 2021 at 10:14:50PM +0200, Raphael Hertzog wrote:
    We could have a "debian/spending-ideas" if you want so that all DD have write access by default. We could restrict access to issues for project members (that automatically includes all DD + selected non-DD directly added to the project).

    Understood, I'm happy to organise it that way too if folks would prefer.

    Kentaro Hayashi already started something in this direction after having
    read this discussion: https://salsa.debian.org/debian/grow-your-ideas

    You might want to work together. I saw his announce on planet Debian but
    AFAIK he never mentioned it in this thread: https://kenhys.hatenablog.jp/entry/2021/04/10/221301

    It should be more widely announced, at least with a "Developer News": https://wiki.debian.org/DeveloperNews

    It just *seems to me* that the email workflow of seconds and inline
    quoting is less structured and very familiar to DDs for fleshing out an
    idea, perhaps augmented with an ad-hoc jitsi call.

    I don't think that a gitlab issue is much more structured. It's still
    quite readable and works well by email as well.

    As far as I am concerned, I think we are way too much email centric and we
    are missing out on opportunities to retain new contributors due to this
    but I digress...

    Cheers,
    --
    ⢀⣴⠾⠻⢶⣦⠀ Raphaël Hertzog <hertzog@debian.org>
    ⣾⠁⢠⠒⠀⣿⡁
    ⢿⡄⠘⠷⠚⠋ The Debian Handbook: https://debian-handbook.info/get/
    ⠈⠳⣄⠀⠀⠀⠀ Debian Long Term Support: https://deb.li/LTS

    -----BEGIN PGP SIGNATURE-----
    Comment: Signed by Raphael Hertzog

    iQEzBAABCgAdFiEE1823g1EQnhJ1LsbSA4gdq+vCmrkFAmB9zv0ACgkQA4gdq+vC mrmGewgAgjpsMUxLbQAJ0ict7GduyjqcAlZKOZR/UAyhiixg03bAre32JS4ttlT1 +8ONUv6UZfqp/kEKlnW5I9oYH0RclW5tjsZmrsuGKMEW3S/2sHMIIaVZws/MjFdf 7ZiHGkNPhxIKuR6nExkbgdtlPo+Isyee5bLdzdT2oKabcAYMStQQWFvZXmnOBjGC jDQTrOQu+3nMu9WsDB0q/a0PRUbKrySyE4bexGfKp4DVIteoTBxDjKtarGnIMEsR aQt8tjT2Zgw2LbPkWeP5RExuK3udGM+Kh0HuZaoS3dysekDkkl+JKjedHJm91H/A ChRQAdQcnuEQFMqxGaOW/+RESIucjQ==
    =WAk5
    -----END PGP SIGNATURE-----

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)