 Hi guys, my name is Andreas Tiller and I, oh, maybe I can talk without this nice penguin mask today. So you might be able to understand me better. Hi again, my name is Andreas Tiller and I want to talk about the DBNMED project and what the effects of the COVID-19 disease have on our project. At first I want to give a short overview about the project. Then I want to go into detail about our COVID-19 hackathons and I also do not want to forget to mention single contribution of persons to the first hackathon. Finally, a short summary, what I think was the main lesson we learned and what the future we might have. The DBNMED project was inspired by a lot of wine on the first step conf in Bordeaux and the concept about the DBNMED project with a different name at that time was created afterwards and on the next day I had to talk about the ideas behind DBNMED. For this DBNMED, DBNMED is a set of packages created by maintainers like Firefox, OpenOffice, GIMP and this crazy guy Andreas Tiller started with WordNet. I had never imagined that I would ever have so many, so totally different packages after about 22 years when I started with WordNet. But what is DBNMED for the newcomer? A newcomer sees a lot of packages and how to find the relevant packages amongst the pool of more than 40,000 binary packages. Well, there are DBNMED pure blends, for instance for schools like ScholarLinux, DBNMEDU for geographical information systems. There are others, but I want to talk about DBNMED which is for life sciences and medical software. In DBNMED we have so called tasks. This is a task for biologists if some expert in gene sequencing want to install its computer. Well, it's easy. Just install, apt-get, install, mid-biome and your computer is ready for gene sequencing, phylogeny and anything the biological expert might expect. We have some tasks in DBNMED. The biology task is covered very well. We have medical imaging, we have a very few packages in medical practice and some very, very small stuff for hospital information systems. We have some statistics about our team and what I think is a very good sign that in this graph is shown the status in 2017 where the majority of packages were maintained by one person and less packages maintained by two persons and even less by three persons. But now the situation is the majority of packages is maintained by two or three maintainers and even sometimes four or five maintainers for some packages and only a few packages are maintained by only one person. And I'd love to change this as well since the best idea would be to have zero packages maintained by only one person. DBNMED is very active to attract new developers and I've made some question here that amongst the 42 DBN developers and DBNMED as we have in our team, 20 of them confirmed that they are DBN developers only because DBNMED exists. This is quite interesting because DBNMED has about 1,000 developers and if 2% are joining DBNMED because of, well, in the scope of DBNMED very special, very small projects, this is quite good because we now have attracted some people because the DBNMED project exists and if other small projects would attract more developers we can build a very stronger DBN team than we are now. For instance, 10 of the 20 people above extended their activity to other fields in DBNMED and 14 out of the 20 above are active in DBNMED. Not all tasks in the DBNMED project are appropriately covered. Bioinformatics is very good covered. Medical imaging is also good covered. We have quite a lot of relevant applications but for medical practice the amount of free software is quite low because we have DBNMED but just lost 3 MED forms which had some issues and it's not properly maintained. Maybe we get it back, whatever. For hospitals we just can provide a stable but unspecific OS. Well, okay, if the hospital wants to run Vista which is kind of a free hospital management system we have at least the underlying database which is FISGTM. So far for the general overview about the DBNMED project now I want to do some graphs and statistics about the COVID-19 hackathon. We made our first hackathon in April this year and I have created some graphs which are based on the team statistics I'm usually having. Please note the R packages are missing here inside these graphs since these are maintained by the R packaging team and are not showing up here. We are doing a lot of R packages but it's hard to put them together into the graphs because other R packages are useful for other teams and it would be not sensible to merge them together into statistics. This graphics shows the committers to the DBNMED packages per week. Since 2006 we have a version control system in DBNMED. Here the gray area marks the time when we moved from Elliott to Sasa but any peaks that might show up here are not really relevant for what we are investigating now. If you want to have an answer about the sprints, ignore this peak. In DBNMED we are doing regular sprints one per year in the beginning of the year and they are marked by blue dotted lines. One sprint here, one sprint here, one sprint here. In the sprints usually you have local peaks as well here and this is quite interesting because this graph shows the number of people who are usually committing. These sprints attract more people than usual. Maybe the full team is active even if not all are on site but this is quite interesting. You also see the red dotted lines in our releases. Before there is a release we have only a few uploaders. It's a freeze time and the freeze time only release critical bugs can be uploaded. We obviously have done a lot to fix those release critical bugs then we have kind of spare time or we are fixing bugs in other packages but now to the hackathon. You see here an all time high of contributors who committed to our DBNMED packages. I'm really proud that we see this graph here which is way beyond the usual sprint activity and for sure even more than the normal activity in our team. Interestingly enough in April the first sprint has shown this high peak but the second sprint in June it has kind of a local peak but it is comparable to other weeks we had before but I think also in general we have more activity since we had the first sprint than usual times. If you go to the commits, the actual number of commits not the committers we also see the peaks in the sprints. We also see that before a release not so much is happening. We have two peaks. I cannot really explain. If we go back to the committers well there is not a significant number of committers that are beyond normal activity and I checked these peaks and sorry it was me who committed so much. I admit I forgot what I did. I wrote a route route and upstream update script and maybe I was testing it heavily and created automatic commits, whatever. Yeah, I was busy at this time but this is not what I want to show. I actually want to show that also here we have in the actual hackathon for the COVID-19 sprint the all-time high of commits in one week and also the second hackathon has if you forget those unexplainable peaks the second most number of commits as we see in the first graph not so many people but others we are busy as well. I admit the third sprint I personally also didn't commit so much at the package uploads I did other stuff preparing depth to depth conference so yeah, okay. If we look at the uploaders of DebianMids packages per week we don't see this heavy peak here in this graph. The rationale is that newcomers don't have upload permissions so the number of uploaders is not drastically increasing because these are the usual people who are uploading all the time. They are more active now but we had other times with more uploads and this is not astonishing but we have also local peaks here and the general activity of uploads increased since the first sprint. If you look at the actual uploads that we are done they are a little bit delayed after this sprint and my explanation for this is we had the commits in the week of this sprint but these commits did not lead to an actual upload they needed further commits until the uploads happened a little bit later this explains the delayed peaks according to this sprint. The uploads here and the alliot migration are fairly easy to explain here we also migrated a lot of packages from SVN to Git so these are artifacts these are some more similar peaks I also can't explain otherwise we see the peaks to our sprints related to our sprints we see also that after release the upload activity started again in some time here as well the usual image. What about the communication? The persons posting to the Debian mid-lists increase also drastically in the hackathon week you see and also these are all time high and considering that we here have our sprints and maybe some other peaks might be influenced by some spam this is not clear because we have spam filters but maybe some very active spammer in 2007 wanted to spam our list I have not checked this but you see clearly that the sprint has quite some effort but as all the other sprints have shown as well the first sprint was the most relevant one the other sprints are not so active anymore and here are the numbers of actual posts it seems in our sprint in 2012 we had a lot to talk I have no idea why we talk so much and here are also the sprint discussion the hackathon discussions related to the COVID-19 sprint in the green dotted lines or other peaks are basically also for sprints and this is maybe there was some discussion about names names are really tended to trigger discussions which nobody really needs but anyway I don't know so far for the graphs what about the communication we had daily JT meetings to organize our sprints we did a lot of discussion on the mailing list as you have seen we had IRC and we had some telegram channel where I have mixed feelings because it is a hidden discussion and I want to be open so please use our mailing list since we try to discuss there we had some really amazing FTP master support in our first hackathon I don't know how to express myself but well in both ways tons of kudos to FTP master for the great support we had lots of new packages that were processed in less than 24 hours I mean 24 hours not 24 days or 24 weeks or so as usual for some new packages it was a really great support I admit it was the most fun in my whole 22 years they were undevelopable life it's extremely motivating specifically for newcomers to get such support yet but to quote some kind of understatement of Scott Kitterman who was maybe the most active member of the FTP master team I didn't do anything beyond taking some time of work so I would have more time to pick up the FTP team and of getting the new packages in well this is this short statement for a really great work and I want to thank again and again for this great support in our first hackathon now I want to thank single individuals for their support ladies first and I start in alphabetical order this is Catherine Tiller my wife and I want to thank her for being tolerant to let me work even on weekends and on holidays we had Easter holidays in Germany thank you Catherine for your all time understanding that I can work freely on this great project you both Schuprikobber is an outreach student from last year she scanned the Covid-19 Birology relevant software which is not in Debian and she is qualified since she is a bioinformatician and she supported us by creating some kind of to-do list to make sure that we get all the software that is relevant packaged she did also some fixes on packages which was her task also when she was doing the Google Summer of Code work sorry, the outreach work Malia Asimani is a former Mentoring of the Month student I am running a program where I am teaching newcomers for one month where they can ask any questions they want to do and Malia learned it some years ago and drifted away somehow but for the Covid-19 Hackathon she came back and she learned a little bit again refreshed her knowledge on the package and worked on some other package thanks Malia for your work on this Rebecca Palmer is a long-standing Debian developer but new to the Debian Made team she fixed and updated Snakemake which is quite an important package for our workflows this is quite important work she also fixed some Python 3 package with some issues for Python 3.8 and she joined our daily video meetings thanks Rebecca for your support Zao Ikouan she wrote other packages for several Python packages she uploaded a new package in the Sprint which are named here and continued later with other packages a quote from Zao thank you for the opportunities I really enjoyed them Alexander Metashvili wrote some tests for the Authentic Python package and submitted a fix for this package well that's actually my work I had a lot of new packages which even preconditions for our packages and packages in progress with some licensing issues I also backboarded some packages because if users want to use packages on their stable systems they need to rely on backboards which I did here by I also did some infrastructure development I fixed some blends an automatically created package which is called the Thermometer I wrote an import for auto package test data into the ultimate divin database and I checked excuses meta data since we want to create a common QA page which should list everything what is in one task on one page this would quite helpful to get people working another thing is that I thought if there is Covid-19 an international pandemic then we could ask upstream would you like to free your software to make it available internationally or distribution so people can access them and this idea was not that bad because we discussed for 2 years with upstream could you please free it now it is free in Davion I was not successful for BLAD but I was successful for lockfit which has quite some reverse dependencies and now we are able to package even more packages very cool and GMAP is free as well so we have 3 packages not yet well this is a cool start I would have loved to free even more packages but it is only one week and this is writing to upstream sometimes boring about this topic so if anybody wants to help here please free your software please do it this would be really great help you don't need to code you just need to write an email or reach out on other channels I also did a lot of teaching in this hackathon for newcomers and so well actually my work switched from doing the technical packaging work to teaching managing inviting people and this was a very interesting experience next one is Andres Merkis he is a member of the DepecheM team and has 2 important packages which are valuable for the Debian team as well as the Debian Mate team he also packaged some javascript dependencies of streamlit which would be very interesting for some epidemiology applications as well as some javascript dependencies of the shiny server so having the shiny server inside Debian would be really cool since a lot of applications are based on it quoting Andres congrats and thank you for your tremendous effort to coordinate the hackathon and thanks to you Andres your contribution was very welcome Antoni Bilalonga he is a new Debian mate he has just fixed some Python 2 to Python 3 bugs which is very welcome and he also remained active in the second sprint in June and hopefully later Antoni, thanks for your work Ben Tris has an FSF email address and he is a non-programmer said he is not able to do much technical work but it was a very great contribution to have a list of possibly COVID-19 relevant packages he did some general QA work as he spotted issues in relevant Debian mate packages for instance he spotted some non-functional watch files and he told us by comparing Debian and the upstream version it was very nice non-technical work everybody is invited to do similar things like Ben thanks a lot Ben Dynarecy is doing a lot of stuff in the R team he has fixed a lot of R packages and while I was for some time the most active uploader in the team Titan is now gaining this position and I'm really happy about this because I'm not proud about being the most active person I'm just doing the upload because nobody else is doing it so if anybody else is doing it like Titan, that's perfect thanks a lot for doing this Emmanuel Arias is new in Debian mate he also ITP'd a package and it's dependencies which are quite a lot he also works now on some other packages and not to forget he helped me a lot with Titan and repeats finder he also talked to upstream he was the problem is if only some guy Andreas is asking upstream they say ignore this guy but if two or three people have the same question and are working in the same direction this helps a lot and so maybe it was just Etienne's question to upstream that it was finally freed thanks a lot for this he fixed also a lot of Titan too and three migration bugs GCC bugs and he joined all meetings Etienne, thanks for your contributions you are a very valuable member of the Debian mate team now Hamid is a friend from Debian Conf 19 he fixed minimap2 and nano polish almost quite important packages and he invested some time into it thanks Hamid Jun Aruba is from Fedora medical sick maybe a redhead employee the Fedora medical sick is basically the same like Debian mate but for Fedora they have not so many packages as we have but anyway Jun participated in Ausprins he provided a lot of nice work I think there is kind of a friendly cooperation between Fedora medical sick and Debian mate I am reading the list and try to give hints to our patches and the main contribution of Jun was to create a spreadsheet with COVID-19 relevant software specifically dedicated at workflows and this spreadsheet is showing what is inside redhead Debian and Kondar from COVID-19 relevant packages Redhead is not so much Kondar is a little bit more than in Debian Kondar is an alternative packaging attempt which has the main advantages that the user can run it there are good reasons to either use Debian packages or Kondar in different applications I don't want to stretch this too much but we are observing each other closely and try to adopt good things from each other this spreadsheet is extremely helpful as a to-do list and to coordinate work and newcomers can easily pick their tasks they can do and find new packages to include in Debian it also Jun also triggered a general discussion about Nanopore workflows and it's really a great contribution even if he is not doing any packaging Jun thanks for your work on this Michael Acruzo is a long-term Debian made member and he took him a while until he became a DD and finally soon after the hackathon he became DD and can now run totally self independently and upload packages but his main contribution in this COVID-19 hackathon was the communication to the kind of official worldwide bio hackathon which was not only done in Debian but in many other projects and Michael did all the coordination wrote reports joined video conferences and all these Michael also created the Git repository on Zaza as well as the wiki pages and he also assembled the final report about the hackathon and he also packaged my little script where I am doing routine updates and fixed some package which needed fixing Michael thanks a lot for this work on the hackathon and your work in general Ninesh Patra is a Google Summer of Code student his topic Google Summer of Code was actually to package software relevant for COVID-19 in the actual hackathon he started with Parasail interesting reverse depends he wrote out the package tests and fixes for some packages not only this one he joined all our video meetings for sure and have done lots of other contributions after this print Ninesh became a Debian maintainer meanwhile and to quote him thanks a lot for these efforts and congratulations on getting your talk selected Ninesh it was great to work together with you thanks a lot Oleg Wojnar is a Debian developer but actually not in the Debian mid team I think he is connected to the Debian games team but maybe we have to thank him the most because he contacted the ftp master team he documented our process and he created an interesting wiki page which is coordinating the work between ftp master and us he was also very busy in introducing guiding and sponsoring newcomers so he lifted a lot of weight from my shoulders which was in many cases only me or the guiding teaching sponsoring so thanks a lot for this Oleg but maybe his greatest initiative was to create the basal packaging team basal is a build framework created by gurl and also tensor floor which is a machine learning system is built by basal basal is also creating Android and the Debian Android team recently mentioned in their talk at the jma team is packaging basal which is not fully correct because it's a separate team because a lot of other packages can be built with basal so if we have basal then a kind of a new universe of software can more or less easily go into Debian and basal is in new now it will be processed soon once we have tensor floor machine learning framework in Debian it would enable lots of other COVID-19 relevant packages from epidemiology to bioinformatics it is widely used and can be very helpful thanks a lot Oleg for your really great work it is very influencing not only for the Debian machine is new in Debian Meet he started with writing some tests and close some bugs in some packages which are mentioned here but his main contribution I'm really happy about is that he started working in Java dependencies for instance for the SNPEF package and the problem is in Debian Meet not much Java competence always laggy behind and there are a lot of interesting Java packages which are not really good maintenance so we really need this competence who is brought in by Pierre currently Pierre is working on fixing Java dependencies and I really hope that he stays that busy as he is and contributes a lot to the Debian Meet team thanks a lot Pierre your work is really welcome the second Google Summer of Code students, we have two students this year his work was to write other package tests for Debian Meet packages in the package he actually packaged creating our packages is really easy he fixed some release critical bug and he wrote several other package tests and joined the daily video meetings I'm similarly happy with Panof like with Nilesh and it's really good to have two so dedicated Google Summer of Code students thanks a lot Panof Scott Ketterman is working on the FTP team he was working down a huge list of new packages in FTP master and it was really great to have you Scott thanks a lot, thanks again Sébastien Jean-Don is the Upsium author of Odyssey he was working on refactoring his package he is also Debian Mante and I can upload this himself he has quite done some quite interesting uploads of this package thanks a lot Sébastien Cheyenne Dost is a former mentoring of the Mante student for actually personal reason he had to do a forced break but he came back short after our first hackathon and it would be unfair not to mention him here since he had an interview since May 2020 and even became a DM now Cheyenne it's great to work together with you you are a very valuable team member Tony Menzel is a Debian developer who fixed two packages in our hackathon he is not really a Debian mid team member but he thought well it's a COVID-19 hackathon let's help these guys Stefan Müller had contacts to some upstreams and he was organizing the bug with them because he is a long-term bioinformatician and long-term Debian mid team member and he is quite busy to fiddle to find connections between bioinformaticians and Debian mid team members he is very keen on creating workflows in the Debian mid team and workflows are kind of connections between the different packages we have in one call to enable bioinformaticians smoothly to a sequence of single package calls and this is the default standard we should implement in Debian mid so Stefan was requesting the transition from packaging towards executing the workflows and also provide tutorials maybe would be an interesting thing well what is my summary about the effect of COVID-19 Debian mid project it's probably hard to say that COVID-19 had a positive effect because it's a heart disease and I would not I have to be careful with wording but every crisis has good chances and Debian mid has used the good chance we have the chances that we get new developers new motivation for the developers who are inside the project and due to COVID-19 the Debian mid project is it became more developed to even stronger project than it was before we became a really new drive and it's even more fun to work in this project everybody is kindly invited to share this fun with us and join our project I love to teach everybody who wants to join and I promise you will be happy even if you don't stay it will be a nice time for you and you can learn a lot about Debian packaging about the community and about life and universe the effect of COVID-19 on my own work inside the project has also shifted formally I was packaging something sometimes like crazy you have seen the graph with these peaks I don't know why I did this but my work shifted from this technical work more to mentoring organizing distributing the work between the people who can do the work and I admit it's even more fun than before to see all those skill developers doing stuff I could never approach myself the thing is I have also limited skills in several fields I'm not even a bioinformatician and don't really understand all those programs that are used in the Debian team it's really cool to see who will join who will dedicate his time and do some effort to bring this project forward it's really cool to see and I'm really happy about this and so again thanks a lot to all who contributed to this great hackathon to our sprints and to our project in general so my talk is online you can find it at my usual talks page my talk is online as my t-shirts are online you can see them in my bag here and that's all for my talk talk outside and I'll see you next year for RealDebConf bye thanks a lot Andreas for your nice talk it's time to see the questions so the first one is trying to understand the purpose of the task why would you use the one to install multiple applications that perform the same task yeah welcome yeah yeah I answered the question in the etherpad the thing is it is not one task that is solved by well I wanted to tell you about it we have assembled in one so called task or meta-package I have a strange echo very sorry I switched off the stream to repeat myself in tasks we are assembling all those packages that can be used by experts to do their usual tasks they do day to day work they need more packages to assemble all those packages made in gene sequencing phyrology or epidemiology where people are using a set of Debian packages and to find this set is actually making very easy by assembling them all in one task I hope this answers the question thank you so the last we have what is the difference between these three hackathons and past the unmet sprints for instance hackathons and sprints since 2011 we are doing team meetings which we call sprints it is mostly weekends maybe Thursday to Sunday also we try to assemble as many as possible developers we are between 10 and 20 people for those sprints usually it is some nice location and it is a meeting in real life so we meet each other which makes quite a difference to this online conference these sprints are very helpful to come together with upstream developers and you see on these graphs that they had quite some effect on our world the actual hackathon what I am reporting out here is the Covid-19 hackathon which was triggers by the need that we want to package as many as possible Covid-19 in different packages so this is dedicated to Covid-19 it was not planned it was quite out of the boot that somebody said there is a general bioinformatics sprints organized by an international group of bioinformaticians and the deviant mid team was joining it and we continued it for because it was so successful and so we started the hackathon was the first one the most successful one and we tried to add to the success in June right before DevCon and it was we had increased activity but it was not so successful as before Andreas thanks a lot for sharing your nice talk with us and for all your work and the team at deviant thanks a lot you're welcome