 So, welcome everybody. We're going to have a booth together, so we mainly discuss, but I prepared the first slides of the introduction. The topic today is using Debian money to fund Debian projects. It comes from two, well, we've seen that Debian has a lot of money, so on one side we want to be, to look how we can use all this money to further Debian schools, maybe through funding Debian projects. And on the other side, I've been handling Debian LTS, so which is using money, not necessarily Debian's money, depending on one's view, but we are funding Debian contributors and well, the community is still alive and the world didn't end, so there are things that can be done and we're going together to try to find more things that could be funded in some way. So, Benjamin McCoy had a nice, wrote a nice text on the dangers of funding volunteers, or dangerous funding people in volunteer projects. So, here's a list of the known problems, so at least the problem of jealousy between members, which are paid, those who are unpaid. There's a problem of volunteers stepping down from work that paid contributors are doing, sometimes for philosophical reasons, sometimes just because, well, if someone else is already doing, then I can do something more interesting or tackle another area where there is nobody working on it. There's a risk of paid contributors trying to lock themselves in their positions because they want to pay again and again for the same work. So, either they do it voluntarily or they can also do it unconsciously because out of laziness, it's easier to well coordinate with yourself and not document anything. And if you want to let other people join, you have to make a report of documenting things. So, it's a risk when you pay people. And the last point is obviously that when you're paid by someone, you listen carefully to what this person or company has to say and you may be less willing to consider the global problems that all users have, those who have money and those who don't have any. Benjamin also found not solutions but things that limit the problems. So, he has a few advice for us. The first one is that we should fund work that volunteers do not do. The LTS project is in this space obviously not many people are interested in doing secretive dates and outdated software. But we can also consider in this area all the administrative work, bookkeeping, accounting sometimes. You should fund limited projects both in scope and in duration. The Google Summary of Scots basically is in this category because it lasts only a few months and the projects are rather well defined. It's not something that is recurring. It's a time limited thing. You can also fund widely desired features that you've been waiting for for a long time. There are quite a few features that when we discuss between us, we always say well, it would be nice if we had that but nobody did anything in years. Or maybe we could use money here to get something going forward. Another important point is more of a process point is that when you handle money you have to do it in a fair and transparent way. I've given LTS talks this morning and I tried to explain what rules we have set up to guarantee this transparency and this fairness. I invite you to look it up if you have the time. Maybe we'll come again to it later during the discussion. On the last point it's possible it's not really a solution. It's rather a workaround. Given the implication of money it's sometimes safer to externalize the handling of money to an external organization. This is also something that we did for LTS not really out of choice but because I believe that it was the safest option to start with because we had history with downtown and everything and we did not want to go through Debian directly because it would have been difficult to get things done in time but maybe know that things are rolling, working well. It might be time to discuss the possibility to bring back LTS support in Debian if that's what we want. I mean the handling of money related to LTS. I also want to share a bit of experience on other free software projects that have been funded. I know at least of the Django project their foundation pays at Django Fellow which is in charge of administrative and community management tasks. He publishes a weekly report and he has in charge tasks that would not be fundable on the Debian side but still they have very good results in particular in terms of timely release. Django always struggled a bit and thanks to their fellow they are able to release timely or rather timely. Reproducible build has two contributors funded through the core infrastructure initiative round. I wanted to give you some detail but unfortunately Olga did not come and I wanted to question him to be able to say a bit more. I asked him over mail but he was unable to respond in time with anything useful. But what I know is that when they add the grant they discussed internally who was interested and basically all the people who were interested had the opportunity to work on a padway. So it came afterwards but it was relatively well handled within the community and nobody seems to have stepped away due to the money or due to the grant. I know also that Taze is paying people. They usually do it for administrative and boring tasks and what is interesting is that fundraising itself is included in that. People are doing fundraising uppaid by Taze. Sometimes after the fact when they get the grant they have the means to pay for the time that they invested in getting the grant. They also pay development projects but all those which are funded are directly related to their roadmap and their roadmap is elaborated in a collaborative way beforehand. They are rather small and they have not a very formalized process to select projects. You are free to discuss with Intrigueri or Solveg and others involved in the Taze projects who are here if you want to learn more. Now we are going to switch to the Gobi document. There is a Gobi document fund raising, funding deviant projects in the DevCon set both directory. I invite you to open it and to take notes because those are the questions that we would like to start with. By the way if you have early questions on what I presented, now it's... What is this? Any questions so far how the LTS program works? How many is 100 for the project? I'll just reiterate the comments that I already shared with both of you in email prior to the conference for the benefit of the room which is that I don't like the current situation between LTS and Debian and your company despite the good work that is being done and the administrative work that is a burden that you are doing. I appreciate that doing it through SPI would be painful given how few resources they can bring to bear to accounting challenges. But I would prefer LTS to either be the work that you are doing through your company to be further removed from Debian so that you are not using the Debian trademark and logos and so that we are not referencing each other with the Debian Wiki and other pages or bring it closer meaning that the money is handled through one of the trusted organizations. Even if that means your own company becomes a trusted organization. This current middle ground is not... I don't like it. And sorry my second comment which is related but not particularly pointed at LTS is that prior to Debconf we had a couple of pints down at the UCT club and we chatted about the perverse incentives that funding people to do work brings to bear. You mentioned it a little bit in some of your risks where people become entrenched in the work they are doing or they are becoming paid to do repetitive work because they don't ever complete the task, whatever the case might be but it brings some perverse incentives to bear. So I know of those challenges. I'm not sure we want to start in this discussion right now because I would like to go through the question first. Still there are two sides. Some people would like to go closer. Some prefer in the external. So I think we will have to make two distant proposals and maybe vote on it too because I don't know how to bring back both sides together. But the idea of making friction a trusted organization is worth exploring I guess because it's a middle ground possibly between... And the thing is you're mentioning that we use a trademark. I made an official request to the trademark team and basically as we refer to services which is long-term support related to Debian they believe that we do not need any special authorization. Anyway let's start with the first question which is besides Debian LTS security work are there other tasks that we should do in Debian that we are not doing because not enough persons are motivated or interested by them? Do you know of any? Please tell and fill the dots. I could also just write it in but I think in general I would love to hear from the auditors, the accounting guys because I know that's a lot of work and a lot of difficult work where it's easy to make errors and I've said this before in the mailing list possibly beneficial if we would outsource this and by that I mean create an interface that forces us to be more correct about what we're doing. I share it but my opinion is that as a project we don't do... It should work but I don't know why. Did you hear me? As a Debian project we don't do any accounting. It's on the TOs. The TOs have to be more active and doing things on that but we as a project we don't do accounting and we probably never need to do any other work. Yes we have a couple of TOs and I think it's great that this is working in a distributed fashion but if this buff is about funding work in Debian in a coordinated fashion and early on we were talking also about the Debian partners program and potentially increasing the revenue or whatever we do to get money that we can then spend on projects then I don't think you're going to be able to do that without consolidation at the top at the project level for accounting and that's hard work and probably not something. I know Martin has been doing a lot of it over there but I'm sure that he can do other things with his time. There's also two different types of work we're talking about now because the long term support work we're assuming that it would be people who are already involved in Debian and normal project members on who would be doing it. If we were saying we want to pay someone to do accounting it's not clear whether we would necessarily choose someone who's a regular project member or just a regular accountant outside so it is a different category. We're planning to put you both. I don't think in the line of what Mehdi said, I don't think project is not doing any doesn't have any incomes and doesn't have any outcomes. This is pipeline through trusted organizations. So we have more than a couple trusted organizations and those are the ones that are doing the accounting. So if there's a partners program or anything it's just the trusted organizations with the ones that are doing this accounting at that level. Yes, I don't think that's true because the trusted organizations at least for SPI might be different for the other ones which actually are more Debian specific but for SPI, SPI doesn't care what kind of expense it is, it's just Debian. But they're not going to classify hardware, travel, sprint. So Debian still needs to do that fine grained accounting and right now basically the reason I didn't make progress was simply because we couldn't get the data from the trusted organizations we needed so every transaction and so that's why I got involved in SPI and to solve it at the root and actually Sobel is doing a fantastic job right now putting stuff in ledger so I think at some point we can actually Debian as a project can go back and get the data from all the trusted organizations and then integrate that into one financial report because we really want to have a financial report about Debian because people are donating to Debian they're not donating to SPI or to Debian, Switzerland they don't care about which trusted organization they just want to see it as one is Debian. We have most of that data but we just miss in SPI which is working progress, they're having some problems with accounting which Martin said Sobel is working on that and this is just coming but there's people working on this, I don't think we need to fund an external accountant to get this more complex than it currently is. Well if people are working on it already then that's great. I'd like to propose that we differentiate a little bit between the concepts of well there's fundraising and then there's accountancy and treasury I think every single trusted organization given their legal status and the country where they are are going to have to do some accountancy no question whatsoever but in the light of the Debian project a trusted organization is really a treasury it's really a trusted body that handles our bank account essentially the DPL is the person that can say money from the air ghost somewhere else and if we accept this which is how the trust organization is defined then what we have is a virtual organization on top of all these trusted organizations which are treasuries and the accounting has to happen at the top level and as TVM just said for Debian in Germany accounts and expense categories that we had to use were tax compliant in Germany they're not necessarily useful in the Debian context so being able to really write expense reports into our financial reports on how we use the money is going to make it a lot easier for us to approach our sponsors and say look this is productive use of your funds and then that gets me into fundraising and I'll just one more sentence on that I think that a trusted organization can do in a like a universe of their own we need to coordinate among all of the trusted organizations to have one united one consolidated fundraising effort because it's not going to work if multiple entities speak to the same sponsors or if we don't know what the others are doing this has to be coordinated from above and we need to use on the ultimate goal and maybe we should get back to the main subject about out of the Debian projects so completely agree with the things that were just said around the trusted organization treasury blah blah blah I agree wholeheartedly with those getting back to the topic since you've asked it's not so much in the category of we're not doing because we're not motivated by them but maybe it's in one of the other questions I really like the Google Summer of Code concept because it's really about trying to find new people who want to work in open source and recruiting them into Debian somehow through a project like Google Summer of Code or what have you so any project that is short lived and is focused around bringing fresh people into the project from either a city perspective or from a skill set perspective would be great this could include funding DDS to go to events where we represent Debian to run workshops you know we could even become a Debian.edu and offer many courses and stuff like that becoming a Debian.edu has a side effect officially has a side effect of becoming a Jerome partner for putting on the goby do you all agree here that we could for example pay people to do some fundraising accounting etc I mean with Debian money the question is about just very quickly I'd be really careful about paying people to do fundraising that is going to contain this issue the question on the goby is what's not covering by volunteer work that could be covered by the I think there's a lot of stuff around fundraising that can be paid for but incentivizing people for instance telling me if I managed to get HPE to do platinum sponsorship next year then I get 5% of the cut I don't think you should do that and not only because of moral issues in this project but also because of tax issues it's difficult the list is fundraising it should not be a percentage like you say it's clearly just covering time spent just like Tails has been doing it's a only sane option but it doesn't need to happen just a possibility I just wanted to make the point that things that are important to Debian get done in the end and most of the things fundraising and accounting it's not being done at the moment but probably if we were on a shortage of money we would do it and we're not doing it it's not that big of a problem because we know that we have about 300,000 dollars at SPI there's no pressure to do it and I think the same holds for most of the things that are not being done Debian website is kind of old looks old but actually we don't really care that much about that because well our distribution also has that image of being a rather old and solid project it kind of matches the general image and I'm not sure we are a project who would like a really fancy website I really doubt that we need to seek additional ways to make progress in such issues I cannot see any of them that are so important and we cannot address if we decide it's so important I think it's time to go to the next question because I don't think we will add much more to the first one Do you know of widely desired features that we have been waiting on for a long time bike shits that was the obvious one thanks so my comment also applies to that category I don't think by shits it's not being done because nobody cares sufficiently about it let's agree on the goal of the book I would also really find ways to get rid of the 3,000 100,000 euros that you have in the bank accounts is to agree on how the Debian money should be spent which things should be covered by the project just have the consensus that if we don't manage to get the money spent that's okayish it's not a goal per se we shouldn't raise that much money I'm not asking the question I think we will have a more interesting question when we come to the next point so just see if we can find a few other examples for this question and then go to the next question do you know of any other things that you have been waiting for that have not been happening in the banking data for now so the auditors team is already in place and have the processes once they have the data from SPI they will be able to resume their work and get the financial report done one it's not a feature that I'm desiring but if I look at the bug count of central techniques like APT, DPKG they are just increasing over all the years and I would be happy if maybe we should do some cleanup there the very important tools that are very important for Debian should not be in this situation where the bug count is just increasing over years and never goes down so on the context of APT and DPKG there's a Google Sum of Code project that is going on right now to enhance the communication between the two tools and get rid of some of that cruft that we have accumulated over the years and I believe that going to question 3 of Debian having its own grant program I guess that would be an interesting way of enabling people to get things done on those Code projects that we desperately need I have another idea it's slightly not a feature but it might be Ubuntu does lots of stuff I'm not so glad about but one thing they do well is they have very pretty designs each version has a very nice background and stuff and in Debian the last one was better and I really liked Spacefun before it but we kind of don't care much about this and we don't have many designers contributing and that might be something that could be funny to found but maybe it might also help with reaching out to more people, new people, younger people different people and that might be interesting I'd just like to say that two days ago we were in here and it was told that there are people interested in working on web design and potentially also digital design it has been getting better already the problem that I see here with funding is probably going to put all of the creative folks that are currently interested off the job because two creative directors can't really work together right so we need to somehow find a team effort to do this and I think the most important thing here is not necessarily the lack of money but the go ahead from the project It's true that in the past we neglected that part of the project about the design we didn't see pictures and we've been doing better for we've done it for jesse for squeeze we've done it less not a good way for wizzy that's why it didn't work well but we've done better for jesse and it's going to be good for the next one stretch so I think it's only a matter of making sure that things are organized and making some announcements around the subject we only have a quarter left just on this topic instead of paying someone we could have prizes so that we could motivate more designers to participate and have a better result I think we've had a lot of designers in the past interested normally the block was working within the technical constraints and making things that people really wanted to push on the technical work even if you pay someone that doesn't solve that side who added the sentence about the per foundation grant can you say a few words about it the per foundation has a grants program when they have an elected board of grant committee members so you can submit a proposal at any time you describe your project, you describe the milestones you want to achieve the time frame and see how much you need to make that happen and then the committee approves that or not and they have to do pretty much like you guys doing LTS which public reports every month and everything so if for instance we had a predefined budget for a year then the grant committee could be team delegated by the DPL to handle that budget ok that's precisely the purpose of the question 3 if we have such a program what is the best way to select what project to fund currently it would fall down to the responsibility of the DPL which approves the project is this what we want, do we want a separate team do we want a vote by the DD body so I just wanted to follow up on that because I was president of the Perl foundation program I'm still on the board I think some of the important features of that are there is a grants committee that votes but there is a public process every grant proposal that comes in is posted publicly and everyone in the community has an opportunity to comment on it so there really is a consensus in the community reached before there's even a vote on it and also the grant committee's authority is delegated so much it would be like from the DPL it's a delegated authority and has certain limits set on it so when it goes over a certain limit then they have to push that up to a higher level but I think really making it a community process is the most important thing so nobody is ever surprised by some project being funded that they think is completely worthless or in fact actively harmful to the community I'd like to add a little bit to that I think it's important to realize that we're not going to be able to come up with a grants program by sitting down and discussing this on a mailing list by sitting down a piece of document I think we should get started and see where it takes us because we're going to learn in the process and being transparent about things I think is going to be the key to success So what's the first step to start it? Do you mean where we have lots of money apparently to spend I think you have to make the limits clear from the beginning to try something because in short projects it doesn't have to commit on big amounts of money but if you do a quick poll for example who would be for the grant idea in Debian and who is for the grant idea in Debian so like 50% and those who are against is this can you expand on why? I'm not so much against it but yes I can see the lens so you can see me I'm not so much against it but I don't think it's necessarily a good idea that would depend on the specifics before I would say yes or no so that's why I didn't say yes I'm for it I'm just not decided yet Some people are really against spending Debian money on paid work from the beginning so it's not about the conditions it's about the philosophy so is there anyone firmly opposed no matter what it's not the way to decide but just to see if someone really feels it's a few I don't know I think we know there are people who will be very unhappy so I'm not entirely comfortable with this point we just start off giving grants because of an explosion and people quitting and people being very upset and giving up on volunteer work and so on I'm fairly against but it probably depends on specifics I think if something is tried in that direction it would be nice to think of it as an experiment think about what are the questions that you want to answer out of the process and make sure that you can answer those questions as part of the experimental process for example I don't think that it would be a good idea to do it using Debian money it would be much more interesting to do it using external money that you know that in the future will be able to raise funds specifically for that it's acceptable to our donors to do stuff like that with our money if we do it with Debian money we won't have any idea on whether companies or individuals that are donating money to Debian would be fine with money for that in the future so at least as a first iteration it would be nice to do it purely using external money so I have an opinion on that is that the people and the companies who have donated money to Debian in the past trust Debian to use the money responsively and I think we have wide enough buffer that we don't really need to worry about that right now so I think the experiment could work with just Debian money yeah but then you don't know if it could be sustainable then we need to do some kind of sponsor survey or something but yeah it does not matter you have a grant program with a budget either you can fund 0 project or 2 or 3 but you adjust the amount of funded project to the amount of sponsorships that you have so I would respond to that keep it small whatever the source of funding I would say first grant should be no more than 5K absolutely keep it tiny and you're actually going to have a hard time raising external funding before you've had a successful grant one at least tiny successful grant could be tiny could be 1K it doesn't matter and if it's like the fact that you spent 1K of Debian money to see the idea to get a successful grant to then raise external funding before you do any more that's more where you find out if it's sustainable than trying to sort of hold off everything until you get external interest we've raised that amount of money in the past for outreach for example so I don't think it would be a problem to do it it was a really huge pain to raise that amount of money for outreach I know I did it that's a useful thing to know that would be a pain to reach I think that what you're raising the point about sustainability of the money is an important one and the more we deal with this the more this always percolates to the top on the one side we need to know how to spend money on the other side I feel that a lot of people are not comfortable spending the money we have in our bank accounts with the vault it would be more comfortable if we knew that 100,000 came in every year and we could use it for that I think it does also depend a lot on what my guess of how people respond is it depends a lot on what direction the grants are going and that I think not many people are thoroughly opposed when it's in the kind of outreach although I suspect there exist people who are opposed but Debbie and money has gone that way because previously it was Google Summer of Code that didn't work that way but I think if we have a more general grant program sooner or later unless you ban it from the start you're likely to end up funding what is seen as core people to do important core tasks and I think that then gets back to what we know from a lot of people at least used to be against maybe that's changed but people were very much against that I think if you compare this to dunk tank which basically just said we're going to try this with the release managers and borrow a lot from the LTS idea which is essentially everybody is free to work on this and if you're the core app maintainer and you can now decide for yourself am I going to do the work that I do anyway or am I possibly going to be able to implement this feature two months early if I get a 1K grant you can apply for it the community transparently decides for it and then I think as soon as we get to the point where we show results it's going to be powerful towards the sponsors and it's also going to probably foreclose most of the discussions in the project well hopefully not right back to just the issue of funding the experiment for the grant idea I think it would be really bad to put fundraising for the grant idea as a hurdle for the people that want to try it because it will kill the project so we have only a few minutes left so what would be the first step to try it would it be okay if the DPL was the one who selected the initial set of projects if you wanted to experiment or do you believe we would need something other since it's the first time we're going to have such projects and since there is no consensus on the idea project-wide yet we could have a vote for our first grant so that we could have either the project approves it or not the likelihood is that it draws it back towards the kind of core activities because people are not going to all vote for the same kind of outreach type project as each other the projects that get a lot of votes will be the ones that actually are also more contentious to support you could also say it's for newcomers outside the community sure if it's something like this I think it's much less likely to get opposition so far so I think the sustainability piece is quite important we've been very generous several years ago with a bite mark donation that really took care of a lot of our hardware concerns and this year HP has been exceedingly generous both with the conference and with hardware for Debian core services so when we look at the amount of money that we have in trust at SPI or the other TOs SPI is a one and a half years of operating budget between DEBCONF and DSA expenses on hardware refresh if there were no money so achieving sustainability on income is really important it's not like we have 10 years worth of funding sitting in the TOs so if I had greater clarity on the income if I had more consistent income then I'd be less concerned about spending a lot of that money but we are only talking about 5K just two things one thing on the fully second what Luca says because the thing is if you do something and you raise expectations and then you can't follow up on it because you have not cleared the income side that's not really worth it I'm again completely with OLSD2 to try something for 5K or 10K so there's no need to hold it all up by making sure you have the income side settled but you need to do both and so before you declare the experiment successful you have to have a way to also have the income side settled so that's just important the other way is I would deliberately not vote because the thing is you need to also establish a process that will half way efficiently allocate a certain amount of funding that you annually take and do not give to DEB CONF and do not give to the DSA hardware and do not give to mini DEB CONF and whatnot into those grants and if that always comes down to having half a GR you need to use this experiment also to establish a process that transparently but efficiently actually selects those projects and if people have complaints they shall complain against the process and help improve the process but have everybody vote so nobody can be against it is actually a very inefficient way of doing it and cannot be sustainable on the long run so we have to find something that will be sustainable on the long run otherwise the experiment doesn't work on the process side we are mostly out of time I think that's one last comment what I was going to say is that if you just say as a DPL I'm assigning now X amount of money to this experiment with no guarantee that we're going to continue doing it but this is just to see what would happen if we do this and then delegate a number of people to run it that would be not very contentious as an experiment to do that if it is made clear that everybody can have their say after the experiment has run even people who are opposed to it will have that feeling then at that point that they may oppose it after the fact if that is honestly done then of course and I think by doing it that way you also give us data to discuss about because right now it's all hypotheticals if we do the experiment and we do it that way then I think that would take away a lot of questions and would make many things a lot clearer and would give us a reasonable discussion which is very difficult to achieve right now last question before I cut the discussion if there was such a program where you had the opportunity to apply for a grant so that you can implement yourself that non-students or someone knowledgeable to implement it would you submit a proposal or would you consider submitting a proposal to the grant so nobody in this room would small tiny cave yet with that is that mentoring someone takes way more time than just doing the thing that you would have done yourself so yeah non-students I mean if you're the mentor you can accept to do it only if you have someone who looks knowledgeable to do it if this became a real thing and people saw other people getting money then suddenly people may be saying well I should get some of that as well so I mean that's a plus and a minus okay thank you everybody and let's continue the discussion