 Okay, welcome for the Debian Science Boff. My name is Andreas Tiller, I'm a physicist working basically in the Debian MED team and because medicine biology is closely connected to science and we are covering common things, I'm also always engaged in Debian science. I just have prepared some slides which show who is working in the Debian Science team. You see the Debian Science team is uploading a lot of packages. This is the number of packages done by different people. I realized that I did a lot in the last time, but these are mostly our packages. Anton Gladke is quite active, Sylvester Lidry has more or less initialized the team. Ola is here in the room and these are the active people and you see they are continuously contributing. On the Debian Science user list, we are discussing common topics about what should be packaged and kind of interaction between users but mostly developers which could be quite good if there would be some more users asking questions like what could be packaged, what's interesting for the users but maybe this becomes some more use in this sense. There is also a developer mailing list. We have some technical things will be discussed and specifically these two people are actually not Debian Science members but they are doing QA work, they have a box here so the box of the report is not really contributing to the team but they are asking the team to do something. These are the people who are hunting bugs in the Debian Science team so you have also Ola and me are mentioned there because we are fixing bugs in our packages and these are the people who are contributing to the version control system. This is in 2018 is a little bit short because I did not switch this graphics to Salsa and I need to do some work for this but you see that we are constantly committing things for the packaging and this is my team definition which I use in the Debian Meet team it's not so good working in the Debian Science team because in the Debian Meet team I experienced that I had a problem Saturday night, 12 o'clock and I was writing to the mailing list and on the Sunday morning when I opened my laptop somebody said, oh I have solved your problem from yesterday and so I wished that the Debian Science team would be more like this that we cooperate more and people would not just stick to their own packages and reach out to others and help others a little bit more. This is, unfortunately you can't see it here on this screen this is a bit sad, I need to explain this projector has a problem, right? I hope it stays. This bar reflects how many people or how many packages get contributed by only one person this number of packages gets contributed by two persons three, four, it goes quite steeply down this is not really a team because if you would have a team the number of packages that are only maintained by one person would be basically zero and then for instance in a Perl team the graph goes like this so most packages are maintained by two or three people and not only by one person which shows that this team has a problem well not really a problem but it's not really teamwork but it's a collection of single package maintainers which I would like to improve so I have prepared a wiki page which exists since several years for the Debian Science booth and I switched to this wiki page and we can also take notes in Gobi if somebody wants to take notes here this would be a document where somebody could do some notes for this booth these are all my slides I imagine that we hand around the mic and you say your name and why you are here and what you are expecting from Debian Science and how you want to contribute to Debian Science is okay for you and maybe you can start because you are sitting next to the mic say your name, what you are doing and then turn around the mic and then we know each other and then we talk a little bit about Debian Science I hope you will find this Okay, I'm Rafael Li from China, Shenzhen I'm a robotics engineer I work on robotics systems mainly you work on logistics our company designed the automated logic operating solutions for logistics for the existing industry they use some computer vision to solve the problem but for ordinary computer vision implementations it cannot tell different packages we have packages from different sides and different colors so the pattern matching in traditional computer vision cannot work so we develop algorithms and social structures to solve that problem to make the classification automated you said robotics or logistics both because we have this robotics task you should know it are you aware of this? Debian Science has created some tasks which contain packages like mathematics, meteorology and we have also wait a moment the screen is not good here we have also robotics it's a difference between my screen and this screen so we have robotics so the Debian Science team has created different so-called tasks this is a blend term and this one is collecting everything that is connected to robotics this is a robotics task which has applications and we have also robotics develop these are packages which are used in developing applications are you aware of these packages? sorry, no it's good then I can transport some information to you this is the point of this part and if you have some applications which are missing here then we should package this are you aware of how to build Debian packages or not? we are using Debian you are using Debian have you built your own Debian package? yes then you can build official Debian packages by finding a sponsor and we put it into official Debian ok, well consider that your task is subscribe to the mailing list and say, hey, I'm the robotics guy I want to contribute here and you also see your friends it is José Luis Rivero and Dimitri, no, this is probably a sponsor but the maintainer is José Mattia is also a sponsor, José maybe it's just José but he would be I think Leopold Paloma Alavedas these are both Spanish guys wait, it's coming up here he is also working in robotics these two guys and if you work together you can make you can cover the robotics task even better and because we have no logistics task if you have some dedicated packages for logistics you can add the task with this one ok, if you have questions now or later, I'm here for full week ok ok, then pass the mic hi, my name is Norimitsu Sugimoto I come from Japan I worked as a programmer I worked post-grace GIS programming and time mapping programming on GIS data and GIS data is based on science I study science and science I know in the project many brands I know I understand many brands ok, if you are working with GIS do you know that we have a Debian GIS team? GIS team I know I know both Debian GIS are fast time yeah, so maybe Debian GIS has also a dedicated project like Debian Meet and Debian GIS they are both scientific brands which have dedicated packages so maybe you find some some software here and maybe you can work together with the Debian GIS team this would be interesting for you you can pass the mic hello everyone my name is Wentin Yang I come from Taiwan and I study law focusing on information law so interesting in user privacy those kind of issues actually I am not familiar with Debian but I use Ubuntu so I just run into some topics and go into the room and listen ok, that's all if you study law there is kind of a blend which is more or less that called Debian Lex there is a mailing list but it's mostly not used so maybe you can write an email here maybe we can discuss something but it's a difference if you talk about law and law issues or if you talk about privacy privacy is more general which is not necessarily law also but thanks for showing up here hello my name is Cher Xu I am from Shanghai I am working in a company which is doing Internet of Things IoT and I just jump into the company so I don't know what the company do sorry and I am almost nothing to do with IoT I am an SRE guy so I don't know too much about the Debian science pretty like that I think there is also some IoT team but it's not listed here I was planning to go here to pick something about that but I am not prepared for discussing about that ok, pretty like that fine that you are here, thank you ok, I am from Shenzhen China and I was interested in Debian science because my personal is kind of my personal interest related to mathematics and physics I was planning maybe I can get some package to work on my own PC and visualize some mathematics phenomenon or physical principles on my PC and I think I can learn a little bit of things from this both I hope you can learn is that we have these task mathematics and mathematics development and mathematics relevant applications and I have noticed that even all this, ok my name is Cong Yu Liao and I am from Taiwan and I am working in E3 my major job is to study and developing algorithm for panoramic video and need some computer vision such as disparity find the disparity of image that sounds interesting I don't know if you have some support here but maybe you I always recommend that you subscribe this Debian science mailing list and discuss what you would like to see in Debian and so this is always helpful maybe there are some friends just hanging around on this list and are too shy to write something and then you find some friends there and this has some chance I can't promise maybe you want to pass the mic or you come close over Hello, my name is Artu I am from Brazil and it's obvious because of your t-shirt and I just finished my masters so I was working benchmark applications to perform scalability analysis and that's one thing that I am interested in to performance and scalability but I am also interested in discovering new tools to get more data related to Debian I believe that Debian has a lot of information and we could make science with it and I am here to discover some tools that are already available in the Debian infrastructure I recommend to read the task here we have data acquisition which is probably relevant for data analysis and if you have some ideas how to restructure all these tasks maybe we could add some more which fits more your field the idea of the task is to up get install science this data acquisition to have everything what is inside Debian 2 on your computer not only installing one package this kind of meta packages I explained tomorrow in my talk about blend techniques this should give you some idea this is not a static thing we invented this is somehow in flux and if it can be made better please provide some information Hello, I am Steven I am living in a small city in Germany and a member of a hackerspace there so to us I think the electronics packages, the electronics task would be most relevant to what we do there and we do a lot with Arduino and small embedded systems Thank you I am from Taiwan Taiwan is a beautiful city and I am studying biology so I am not cross computer and I just know Debian before I don't know Debian Khan and I am so I am here Thank you Hello everyone I am more Joe Chinese master student and I am a member of Debian science team Recently I am working on programming packages and Julia language and MKL Hey everyone My name is Yunxion Chai and I come from I come from mainland China and actually I am majoring in chemistry and Debian is actually my working environment on some scientific computing The packages I am currently using is basically for some quantum chemistry computing and actually sometimes I grow max for molecular assimilation and that's it I am basically a user I am not a DD yet Wait a moment please This is Debian team which is probably where you can find some friends like Debian Geese or Debian Meet or Debian So if you are a chemist they are really in need of people who are working together with them So if you could learn how to package and join this team this would be really really good because there are well only 5 people about which are working this and they could use some help So if you join the team you are more friendly and kind people and looking for some more contributors Yeah, thank you Hi everyone My name is Xu Xiao I am a student of linguistics and I am going to start my master in computational linguistics this year in Europe I used Debian Science linguistics during my bachelor project and actually it's basic Python and RTK Well, my suggestion is that maybe Python or PIPI should be included in Debian the linguistics package because it's essentially used in every single computational linguistics project Yeah, as you see we have a different linguistics task and if something is included or not included it's basically depending on who really does the work if you want to get something included then it's the best idea if you include it yourself it's not that hard we everybody of us started some day I also did my first package it was a little teddy putting around on the screen so I did to make my son happy so you can learn that it would be a good idea people in the linguistics are working let's have a check it's not so many people as well it was just sponsored by me here I think it's Karthik mystery I think I have seen his batch here on the Debconf so I wonder why he isn't here in this bof though there are some people I think you should watch out for Karthik mystery I've seen this batch on the Debconf I have no face behind it but talk to him work together, build a package you need and put it in Debian this is my recommendation we share this one actually you said that you use NLTK package and NLTK package maintainer do you use SPACY that's also a famous computational linguistic package SPACY yeah actually some of my research project covered natural language processing but it's not my major but I maintained some package recently I no one month ago I started to package SPACY but then I aborted it because I don't have enough time if you are interested in some packages like SPACY you are welcome to contribute please share the mic we get everything on the video stream the contact should basically be Debian-science at this.debian.org because if you just do too discuss each other you waste the chance that somebody else joins the discussion so I would really strongly recommend to discuss on the mailing list for instance this cardiac mystery might have some additional input to your discussion so please use the mailing list and discuss there and I trust him that he is also a good mentor than teaching you something because he is doing a really good job it's a good I like this conversation inside the box because it somehow has a potential to get new contributions okay my name is Ulle Streicher I am working for Debian-Astro which was in the beginning also part of Debian-science and then separated into its own blend and for Debian-science I contribute whatever is we need and is more or less common for others as well actually as a comment for instance if you are working for robotics and so it might be an interesting thing to say oh we are now 4 or 5 people working for robotics let's do a separate blend and do more fine-grained tasks than just this robotics task this is what happened in Astro and maybe in others so this is a good idea to say now there is a critical mass of one-on-one hand packages and on the other hand maintainers and then you can build your own group which might attract more people than such a general science blend okay hello so my name is Ralf I am working at a university in Paris and I am interested in formal methods program modification start solving SMT solving things of that nature so these are tools which are part of the logic logic task and I have contributed a few things to this task okay yeah so do we have further questions about something we should do in Debian so I think the essence of this was we have basically we cover the fields of the people here in this room and we could add more packages to this and my advice is don't wait until somebody else has built a package just do it yourself because it's some time to learn but you know that it will happen if you really want to and in Debian it's a way you have no upload permission if you are a newcomer but you can ask for a sponsor and this is Debian developer with like for instance Ralf we can good sponsor or I do it also myself then we check your packaging and upload so if you find a team like Debian science your chances to find a sponsor are higher than if you are just a random package because I try to assemble people with common interests there and people with common interests have better chances to get something done so this is the idea of Debian science or if you have I'm also running in the Debian media team kind of a sponsoring program mentoring program is called once where I teach people to package something maybe someone defines the time also for you to teach you how to package because the entrance to Debian is more easy I have a general comment so I think that Debian science team is somewhat atypical in Debian because that explains part of the problems that you mentioned because it's not centered around one computer technology because there are many teams which are centered around a certain programming language like your camel team or pearl team or peasant team or stuff like that or around KDE or GNOME or things like that and that's of course not the case with science because they may use any technology so maybe it's not really fair to compare the statistics with what you see and the pearl team so maybe the team which is closest in nature to this team please do not laugh as the games team for sure they should have very similar problems definitely not more things because we have language teams which are technology oriented which have application oriented teams like games or like science or all the other plans so when I'm doing packaging I have touched from the technology that the software was written at least 10 maybe 15 programming languages which I don't speak myself but I try to do the installation method reading how to install software then you can print a little bit of package but this makes I'm very bright from the technology but I'm not very deep if there are problems I always need to address the Java team, pearl team and ask for help and this is maybe also for you the case if you find a scientific application you need for your topic but you have no idea about the programming language you don't should be afraid about this because we have all the language teams and you can ask there and there's also I just repeat what I said before there is a Debian mentors list Debian mentors help you doing the packaging and despite the fact that I'm a Debian developer since 20 years I'm frequently showing up there and say I need help for this or that problem and this is a very nice experience when you get the help after a couple of hours which you can't do yourself and there is a lot of knowledge hidden somewhere in the Debian universe and you should just try to find the way to this knowledge usually by asking or mailing this and I just try to create the Debian science list first and then we can redirect you to the places where you can find help I have another general comment to do a possible problem with science related packages is that there are sometimes badly perceived in the community for a certain reason so software is coming from universities from research institutions have a certain bad image which is that they are badly maintained that they are one short developments that is some PhD student or some researcher for his research project he writes a software and puts it on his webpage and then the PhD student finishes his thesis and goes away or the researcher goes on a different project which is no longer maintained and for this reason scientific software often has quite bad image in the community which is not completely wrong so there is some foundation for that in my opinion I observed it as well some more like issues in publication and then forget the software something like this but you find examples for this and you find examples for things that are definitely better maintained and so we have I agree on the deviant development list, the scientific software has some bad images from some people specifically the people who are doing the ports to other software the scientist writes a software in MT64 and don't care about ARM or even more strange architectures or strange is the wrong word it tastes unusual architectures which is a good reason because who wants to do gene sequencing on his mobile phone nobody will do this and so there is no real need that this software really runs on mobile phones or other internet of things typical devices the argument the porters have in this case is that if the software is well written it can run on any hardware but the scientists just don't care and we also have limited manpower and then we say okay we provided for AMD64 and this must suffice for all practical cases it's sufficient and other cases yeah who wants to port it he can perfectly file bug reports and patches any more comments others we have seven minutes left so I would like to know if Debian has some tools for example to access data from other tools for example UDT and get data from archive as well in an easy way and you have this kind of tools to facilitate this work to get data from other I'm not sure what you mean about data but we have the data about the Debian packages these pages are generated from UDT this is everything in UDT we have also data I go to the Debian Meet package because we have what's interesting for scientists in Debian Meet close to every package has this information I'm very keen on it close to every package has a citation which is very important for me because it's an additional motivation for the scientists that they are my publication has another Google hit on a well searched website and so we provide some service for the software also and they are on the other hand more happy if they see this package for them we have also some data about registries of the software so this is metadata we put into the packaging and then it goes to UDT and you can drain it from UDT directly and for instance it is rendered here answer to the question? yes I would even expand on that citation approach which is I think very important for they are not just the authors but also for the users for example in astronomy there is a big effort now ongoing that the people should cite actually the software they use and that's about if you have a system which is ready then that's quite difficult to find out what's there and what could be cited and if you have all the information right there then it's much easier for the people to find out what they actually use so what I'm doing if I'm precise start to package a new software I do the Debian packaging and I'm seeking for the publication belonging to this and add the information this is for me if it exists the package is only complete if this information is included in the packaging I hope that every member of Debian Science will do the same maybe we could somehow also ask the upstream to put their preferred citation methods on their web page or we do that in the moment already in astronomy but I think that's worth everywhere maybe but I have made experience that 50% of upstream is very responsive and the other 50% just don't realize so we can only cover 50% with these requirements I think it's a question to teach people in general when writing scientific software to be aware of that they maybe want to be cited somehow it should be easy for others to find out I fully agree some more comments if there are no further comments I'm here the full week if you have questions just approach me maybe I can redirect you to the people who know better than me and thanks for attending oh there's another question by the way I have to talk about the technique behind how these web pages are created and how meta packages are created and so on so when you create a kind of experiments to publication, article, paper stuff like this you want people to reproduce those experiments using your scripts and use your data I have no scripts and no data this is what upstream provides I take software if they are providing data then yes but my question is does it make sense to create packages for to put together the set of scripts that I have to try to make my experiment yeah sure more reproducible well I have my personal policy is I package every software I'm using every software even if I have a very private stuff I'm using on my computer package because if I want to install it another computer I'm way faster starting from 3 computers you need to install it's worth the effort to do the package and if some colleague of mine approached me oh I need software this or that I will not manually install it I create a package and the package is installed so it makes sense because maybe you want to run your software in a cluster also and then you will be happy to roll it out in all the machines whatever application and finally the packaging is also some kind of documentation how you install the package sometimes you forget oh what did I the documentation is the packaging codes with the rules file documents what you did to install the package so it has several advantages so I think we have to finish now thank you for attending and maybe see you tomorrow maybe you see you in the R I don't know how many people are using R this statistical program I'm doing this as well now