Debian Med - a Debian Pure Blends for medical care and microbiological research
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
Title |
| |
Title of Series | ||
Number of Parts | 90 | |
Author | ||
License | CC Attribution 2.0 Belgium: You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor. | |
Identifiers | 10.5446/40254 (DOI) | |
Publisher | ||
Release Date | ||
Language |
Content Metadata
Subject Area | ||
Genre | ||
Abstract |
|
FOSDEM 201385 / 90
2
5
8
10
12
13
14
15
17
19
21
24
25
28
29
31
32
34
36
39
40
43
44
46
50
51
52
54
55
57
58
62
65
66
67
78
79
87
88
00:00
Projective planeSoftwareComplex (psychology)Medical imagingRight angleGraph (mathematics)Point (geometry)Lattice (order)Open sourceBitPlanningLecture/Conference
01:06
Forcing (mathematics)Slide ruleTranslation (relic)TwitterSoftwareData structureComputer animationLecture/Conference
01:46
Data managementBitSound effectData managementPhysical systemSoftwareMultiplication signExpert systemComputer animationLecture/Conference
02:32
Lecture/Conference
03:13
Computer iconSoftwareLatent heatMedical imagingFocus (optics)Information systemsProjective planeLecture/Conference
04:23
SpacetimeCovering spaceInstance (computer science)Latent heatDimensional analysisField (computer science)Lecture/Conference
05:01
Demo (music)VotingDistribution (mathematics)ArmDifferential (mechanical device)Single-precision floating-point formatCodeMedical imagingLine (geometry)Projective planeCodeArithmetic meanDoubling the cubeError messageDifferent (Kate Ryan album)CuboidSystem callSource codeJSONXMLUML
06:24
Data structureExecutive information systemData modelModul <Datentyp>FreewareField (computer science)Noise (electronics)BitElectronic mailing listEmailOpen sourceSoftwareSoftware developerLattice (order)Web pageDistribution (mathematics)Physical systemSimilarity (geometry)Data managementRevision controlData structureTask (computing)WebsiteProjective planePoint (geometry)Extension (kinesiology)Instance (computer science)Group actionGame controllerLocal ringNumberTheory of relativityCartesian coordinate systemOcean currentProcess (computing)Perturbation theoryWikiClient (computing)Insertion lossLatent heatHypermediaSoftware suiteInternetworkingFlow separationDistributed computingPlotterPlanningCubeControl flowGraph coloringPhase transitionLecture/ConferenceSource codeComputer animation
12:22
Group actionGame theoryOperator (mathematics)Device driverGoodness of fitLecture/Conference
13:17
Modul <Datentyp>Modulo (jargon)ExistenceGraph (mathematics)PRINCE2SoftwareData structureMultimediaPoint (geometry)Decision theoryField (computer science)Proper mapFlow separationFreewareMoment (mathematics)File formatDerivation (linguistics)Endliche ModelltheoriePatch (Unix)NewsletterPhysicalismArithmetic progressionSoftware developerHand fanSource codeXMLUMLLecture/Conference
15:59
outputModul <Datentyp>DeterminantString (computer science)Distribution (mathematics)Axiom of choiceVariable (mathematics)Musical ensembleWell-formed formulaStaff (military)Server (computing)Software developerInternetworkingSource codeJSONXMLUML
16:59
Boss CorporationPatch (Unix)Distribution (mathematics)WebsiteWeb pageDatabaseLevel (video gaming)MetreSingle-precision floating-point formatIndependence (probability theory)Computer fileMedical imagingNeuroinformatikX-ray computed tomographyUser interfaceSoftwareStaff (military)Computer programmingWordTask (computing)HypermediaMeta elementRepository (publishing)MultilaterationTheory of relativityMIDICuboidWeb 2.0Lecture/Conference
19:35
Local ringDescriptive statisticsElectronic mailing listBitTranslation (relic)Element (mathematics)
20:19
Translation (relic)Instance (computer science)Web pageCanonical ensembleSpacetimeDescriptive statisticsPhysical systemForm (programming)Expert systemNP-hardComputer programmingWordIntelligent NetworkProjective planeCodeMilitary basePhysicistLecture/ConferenceComputer animation
21:32
Alpha (investment)Computer programmingAndroid (robot)Revision controlVotingInformationLecture/Conference
22:13
Sheaf (mathematics)InformationMenu (computing)Meta elementMIDIDecision theoryBiostatisticsGroup actionDifferent (Kate Ryan album)Medical imagingLecture/Conference
23:00
SoftwareEvent horizonWeb pageSoftware suiteCASE <Informatik>Latent heatTask (computing)Point (geometry)Lecture/Conference
23:56
CodeWeb pageField (computer science)Task (computing)Web browserSoftware bugAdditionPhysical systemTrailXMLLecture/Conference
24:35
Web pageHaar measureSingle-precision floating-point formatComputer fileWeb pageSoftwareSheaf (mathematics)Web 2.0Probability density functionLink (knot theory)Software developerMIDIGraph (mathematics)Instance (computer science)Flow separationDatabaseElectronic mailing listSoftware bugPoint (geometry)Group actionPerspective (visual)Service (economics)Source codeJSONXMLLecture/Conference
26:14
Hausdorff spaceTask (computing)Hard disk driveSoftware developerRevision controlMetropolitan area networkRing (mathematics)Data miningElectronic mailing listMIDIOnline helpOrientation (vector space)Direction (geometry)BitSource code
27:06
Arithmetic progressionComplex (psychology)Multiplication signRow (database)QuadrilateralData managementFreewareSoftwareEmailStudent's t-testRight angleComplex systemLecture/Conference
28:40
Software developerRoute of administrationPoint (geometry)Software maintenanceSuite (music)Bus (computing)Atomic numberMusical ensembleElectronic visual displayVideo gamePhysical systemProof theoryProduct (business)ExistenceProjective planeTerm (mathematics)Noise (electronics)Source codeLecture/Conference
31:03
Graph (mathematics)Degree (graph theory)Sound effectMultiplication signPoint (geometry)Graph (mathematics)Monster groupLecture/Conference
31:55
Point (geometry)Electronic mailing listEmailGraph (mathematics)BitLecture/Conference
32:37
AuthorizationSoftware developerData managementSoftware bugSound effectMassComputer programmingGraph (mathematics)Graph (mathematics)Lecture/Conference
33:29
AdditionCirclePolar coordinate systemLattice (order)Extension (kinesiology)Video gameReal numberLecture/Conference
34:11
Real numberVideo gameLattice (order)Single-precision floating-point formatView (database)Universe (mathematics)Lecture/Conference
35:01
Sound effectBitGraph (mathematics)Cartesian coordinate systemSoftware developerEnterprise architectureMultiplication sign40 (number)WindowConnected spaceStudent's t-testField (computer science)Task (computing)Letterpress printingQueue (abstract data type)DiagramLecture/Conference
37:00
Stagnation pointExistenceStrategy gameGraph (mathematics)Expert systemProjective planePlanningStudent's t-testDirection (geometry)Focus (optics)Ocean currentSet (mathematics)Computer programmingMehrplatzsystemOperating systemSoftware developerExtension (kinesiology)Distribution (mathematics)AlgorithmGoodness of fitGame controllerBitCodePoint (geometry)CodeBit rateLogic gateMoore's lawPhysical systemWorkstation <Musikinstrument>Musical ensembleMoment (mathematics)MathematicsSingle-precision floating-point formatPatch (Unix)Disk read-and-write headControl flowArithmetic progressionRoboticsSource codeJSONXMLUMLLecture/Conference
41:13
Multiplication signTask (computing)Web pageDescriptive statisticsSoftwareData managementPhysical systemMeta elementVideo gameAuthorizationComputer programmingSource codeGraph (mathematics)Phase transitionObject (grammar)CodeTable (information)Revision controlVirtual machineServer (computing)DatabaseExtension (kinesiology)FreewareContent (media)System administratorMetreGame controllerProjective planeInstance (computer science)Maxima and minimaState of matterMathematical singularityUniqueness quantificationSubsetMusical ensembleForm (programming)MaizeLecture/Conference
45:35
Point (geometry)SoftwareClient (computing)Programmer (hardware)AuthorizationInstallation artServer (computing)Computer programmingMathematical singularityCellular automatonRobotService (economics)XML
47:03
Well-formed formulaGene clusterSequenceOffice suiteService (economics)Physical systemComputer programmingIdentity managementPoint cloudGroup actionUltraviolet photoelectron spectroscopyFreewareSoftwareTask (computing)Noise (electronics)Ocean currentMilitary baseProgrammer (hardware)Flow separationLecture/Conference
49:15
AreaComputer programming10 (number)FreezingRevision control
50:05
StatisticsBitProjective planeSet (mathematics)WebsiteView (database)Point (geometry)Square numberHypermediaExtension (kinesiology)Lecture/Conference
Transcript: English(auto-generated)
00:00
Yeah, this is the last talk of the day and so I do not want to do some complex scripts also. I have some images, some graphs and so on for you and to make sure that your dinner will serve right. And I hope it will be not too boring for you. Davian, it sounds like medicine and so but today I do not want to talk about the medical topic
00:27
but rather how to structure a project around some software which is not so well covered in free software. I have a short intro then the main point is to demonstrate the Blends concept on the example of Davian Meade.
00:48
And then I have some graphs which I try to use to find out how the team works in Davian Meade. A little bit about the history. This idea was born 14 years ago in Bordeaux.
01:05
It is a liberal software meeting 2 or DEPCONF 1 or so. I am not really sure, not on this slide. In 2001. The question was at this day somebody was starting to translate the installation instruction of the medical software.
01:29
And I thought how stupid it is to translate installation instructions instead of making it easy to install. Nobody has to read the installation instructions. So we started after having some wine with my first talk about Davian Meade.
01:45
The name was not there but we started working. It was a little bit influenced by Alco, yes, but it worked somehow. If I say Davian Meade, what are people thinking? It is a practice management system and I have also always to fight this impression
02:01
because it contains a practice management system. But it is not something or not only intended to run your medical practice. The better explanation is it is a Davian pure blend for medical care and microbiology research.
02:20
The reason why we also cover the microbiology research is that the first time there was not so much medical software, pure medical software like what the doctors are running to do their documentation and to run a hospital or so. And so we needed to start with something which is interesting for people in medical care.
02:45
And my own motivation was I am working in a medical institute and they don't use this. Nobody is using this. But my idea was if they decide to use free software, then I want to be ready prepared with something they could use.
03:03
And after 10, 11, 12 years of work, I would say, well, we are not ready. They could start thinking about using it now. So what is my impression about Davian? Davian in principle is a collection of software.
03:20
Here you see some famous icons and some people behind this. And this is a big advantage of Davian because you have some people you can talk to. Perhaps this is a team. I started with WorldNets, a little bit specific. But if you are a user which is certain work field, you see a lot of packages.
03:45
I think we have 30,000. And you will not find the software you want to use. And you are frustrated. And how will I find, for instance, a medical software? And for this we started this DavianMID because it has a focus on this medical software.
04:01
This image is in principle guiding the people, be it for medical care, be it for chemistry. Or we have a DavianGIS project for information systems. DavianEdu and all these so-called blends are focusing the user on a certain topic.
04:24
And my idea is if we have all these blends, we can cover by certain interests. The whole 2G dimensions with certain fields and these are spreading out. For instance, DavianMID and DavianGIS are working together.
04:43
So we can find users who are working in both fields. And if these people start working together, we can fill up the user space with specific users. But finally we can fill up the whole user space to run Davian.
05:03
What about the name blend? Formerly it was called Davian internal project. But we wanted to have some differentiation between things like these arm ports. And so this is also internal project. And so in 2003 there was a name custom Davian distribution chosen.
05:21
It was a total failure because from the sound people imagined that it's something else than Davian. It's Davian customized to do something. And this name leaded to so many confusions that we finally changed it. And I had the idea to say Davian internal solutions.
05:41
Well this is a funny thing. Once you start discussing about the name, the people are popping up and start bikesetting. And think they have some very strong opinion. There were people who never even wrote a single line of code or they were not known.
06:00
But they had a strong opinion on the name. And so we ended up with Davian pure blends in short blends. And finally the good thing about this name is if you don't have no image of what blend means, you need to read the definition of what it means.
06:20
And this is finally much better than just assuming something what it is not. So we have similar structures in Fedora. It is called special interest groups or seeks. And we even have the two things. Fedora medical and Fedora science which to some extent are also covering the software we also have in Davian.
06:47
Fedora has not so much application as we have packaged. But well I'm reading their mailing list and try to give some hints. And if they start something I said well there is some package for Davian.
07:03
Perhaps you can steal some code or so. And this is I hope that we can develop a stronger relation to these people. Because finally we are doing the same work. We try to support medical care and we try to doing it inside a larger distribution.
07:22
And in my opinion this is the main point. There is also open source medical. At least I have read it sometimes in German news sites. But I'm not really sure. They somehow just copied out task what was already done.
07:42
Put it on a web page or wiki page. And that was all. And I try to read their mailing list and ask them do we need some help? Can we do so? I have no problem if they copy our stuff it's free. But if there is no substance behind it it's just a little bit confusing for the users.
08:03
And if a user comes and sees well there is a lot of noise. A lot of what they have done but there is nothing. Then they come to us they perhaps expect there is also nothing. And this is a little bit unfortunate. Then there is Ubuntu meet.
08:22
I checked the web page. I don't know these people. In my opinion they did the first big failure to contact Alice. I don't know why. They are lagging behind the official Ubuntu. I don't know I'm not so sure about these Ubuntu versions.
08:44
And they have a quite different focus. They are just very specific on the system you are using in a doctor's practice. And there is free BSD ports which has only biology.
09:03
But they have a lot of packages. They have from the number not as much as we do. But they have a lot some others and so. And this is also interesting if you are a biologist. You might be interested in this. So in short Debian Meet has by far the largest team and the amount of packages.
09:26
And I also want to stress this. The team is quite large. We are at least you can cannot really obtain from the activity. But we are at least 20 people. I have some other numbers.
09:40
Perhaps more who are working with us. So the purpose of a blend is to make a certain topic hot inside Debian. And you want to attract users of a specific world's work field. Just to drag them into Debian. Before we started with this approach.
10:01
People took Debian and did something else like this custom distribution approach. This is okay. It's free software. But the problem is you have two or three people behind this. And they start from some status of Debian and add some medical software.
10:22
And this is just work. And this work needs to be done by these people. But maybe they get a different job. They get children. Whatever good reasons it could have to stop working on this what you are doing currently. And these projects mostly died because of bad management of this project.
10:50
And so the idea was to stay inside a distribution. And teach users and developers how to work together with Debian with a blends team.
11:02
There is one positive example. There is BioLinux. I started with Red Hat. I've seen that it's not so good for them. Use Debian then use Ubuntu. And always made a very separate thing. For instance the installation of their packages went to user local.
11:21
Which is... I don't want to qualify this. And now they work together with us. And they say well could you please package this software. And we are doing it. And we are taking over other software they have packaged. And do some quality control and so.
11:41
So the idea is a blend is a way to advertise Debian in a specific work field. So we try to advertise this for instance to the BioLinux people. They could use Debian. They could work together with us. And we can do something together. And they now profit from the Debian made effort via Ubuntu.
12:03
And then they derive from Ubuntu. But they do it directly and do not make their own stuff. Why do we want to consider Debian made as a model? Well we were one of the first blend. The idea was originally born with Debian junior in 2000.
12:25
Debian junior is unfortunately not so good maintained anymore. Because it hits basically one or two drivers. And they move to other tasks. And I really really would wish that some people would do it.
12:40
And I always try to advertise if you want to care for your children at home. If they should run a nice. It should be nice for the children. A nice operating system. Just try to reactivate this Debian junior stuff. It's basically checking what games are nice for children.
13:02
What is making easy desktop and such stuff. This is kind of a good idea. Well this is the blend's idea to prepare Debian for a specific user group. Why not this user group should be children. So as I said Debian made was one of the first.
13:23
Besides Debian junior we had also Debian edu which is working quite good. And we now have Debian science and dbchem. I have shown you some graph of what other blends exist. And one idea is also Debian made is a model.
13:43
Because in principle medical software has a really bad chance in free software. Because well you know doctors have money. They don't care so much for the money. And do not understand the concept of free software. They just do not understand. I can tell you I'm working the institute.
14:01
And after I tried hard and had a one hour talk about free software. They said well this is interesting. Can we have this? But then these people were not the decision makers. They understood it but they wasn't able to make the decision to do it. So I can tell you medical software has a really bad chance to survive.
14:28
But in Debian it did. And this is the point. You have a lot of multimedia software in Debian and GIS software and so on. And this is without strong structure behind this.
14:43
There are people who are caring for this. But they don't follow this very strict model. I try to apply to form a solid team working together and so on. And they don't advertise their stuff. So if they would use these methods I'm also showing you later.
15:03
They could be much more effective inside Debian. And you could also do some say Debian physics, Debian mathematics, Debian whatever. What you imagine, what field you could cover. Because it's just proven that even software with these chances can do it.
15:24
Well I said that I want to form a team. And on the other hand the work of the Blends team pushed Debian into a field of quite dominant propriety software.
15:41
The same reasons because they have not so much free software. They have a lot of propriety software. And Debian made something several people just know for the moment. And perhaps they can use it later.
16:00
In short the Blends want to turn Debian into the distributor of choice for medical care. So if anybody working medical care asks, oh I want to use Linux, what should I do? Should I use Fedora, SUSE, whatever or anything other. The natural answer should be Debian. This is my goal.
16:22
I do not want to fight other distributions. They can do the same. I will support it. But my goal is if it's about biology or medicine just use Debian and you are served well. So what tools are we using in Debian Made?
16:41
And what I developed or which I formalized with just substituting the Debian Made string to variable so that everybody else can use it and is using it finally. There is a so called package Blends Dev. Blends Dev is doing nothing else than creating meta packages.
17:04
Meta packages are defining dependencies inside Debian. You can say there is a package made practice. And if you install made practice you get everything what is inside Debian and it works in medical practice.
17:21
This is a canoe made. This is a program where doctors can do their documentation. It's a database, PostgreSQL and a user interface for entering medical data. Or you can say install made imaging and you get a lot of packages.
17:44
I think about 30 or so which are dealing with computer tomography images, extra images. And you can handle all this with your computer just by installing one single package. Or mid biology is the largest package about 120 packages which all can be used to handle gene sequences
18:08
doing phylogeny or something like this. And this Blends Dev package is working on the so called task files.
18:20
The task files are defining relations. It says depends package X, Y, Z or so. And it checks if this package really is contained in Debian. It might be a package, might go away. Or we even include intentionally packages which are not yet in Debian.
18:42
And if Blends Dev realize oh this dependency cannot fulfill it reduces the dependency to suggests. So it can suggest anything which has the reason if the package enters Debian later. You can install it anyway or you get the signal or you can get it with install suggests with enforcing up to do so.
19:09
Or you can use some private repository. You can set up in whatever derived distribution. And this meta package also works.
19:21
So this is Blends Dev. And we have a so called web center which has task and box pages. So I will show this for short. I will show it later a little bit more. So this is basically a list of made bio.
19:44
Which has all the packages which are available in Debian and are relevant for biology. And you see a lot of things here. You see well the package name. You see a short description, a long description.
20:01
And it is not by chance that I have chosen the German locale but it is even translated. This is important for people working medicine because we cannot assume that everybody understands English. So you can translate the description and we try to force people to do the translations.
20:24
So for instance this means fix the translation. Or you can if we find a not yet translated. We can find a not yet translated. Yeah this is not yet translated. This is English. Then the user could go to this page and translate description.
20:45
The idea is if I want to attract biologists to this page. These people are the most competent people to translate this description. I'm a physicist. I basically will not understand what this package is doing.
21:02
I can do packaging but I do not really understand what this means. What is inscribed here. So I want to have the experts translating this. So this is also quite user based. So you can help the project without being able to do some coding.
21:21
You can just translate something. Or you can do a screenshot. As we've seen when browsing we had some screenshots here. Here's a screenshot. So use the program. Make a screenshot and upload it here.
21:40
And you always find a button where you can do it. I think it's also easy. You find information. If there is a new upstream version available. Which is here. And all you see there is everything is fine with the versions.
22:05
And you can also see how many users are actually using this package. If you see here you have 8 users. You have 122 users. There is some information about the package. And you can base the decision what you want to use later.
22:24
If you install the mid bio meta package you can install these all. But then you can pick from the menu which one is more interesting for you. And we have also different colored sections.
22:45
We have some packages which are new. So they are expected to arrive inside Debian. And we have some packages which we are working on. And this is also some information where you can find it in the VCS.
23:05
This package is in SVN. You can work on it. And so we want to avoid people to make duplicate effort. And they could start working on this. And in the specific case on biology and also in Debian science.
23:25
The software is quite frequently written by scientists. And scientists are doing publications. And the publication is connected to the software. So we also have a way to add these publications.
23:41
And this is something scientists are really proud of. And this is in turn something the people come to Debian. Because we are rewarding their work. And this is also one point I want to stress. So this is the task. These are the task pages. And we have also pages.
24:01
I do not want to show so much on the web browser. That's why I can show it later. Where you can find all the bugs which are currently open to the packages in this field. Also with the idea people are interested in a specific field. And will be more interested to fix these bugs.
24:22
Than going to the general bug tracking system of Debian. But these are the bugs relevant for their software. As I just said we have some handling of scientific situations. We are also creating a bibtex file for all software inside Debian.
24:45
So I think it's about 150 or so citations connected to any Debian package. And you can download a bibtex file and do some citations if you like this.
25:01
We have also these QA based pages where you see. This is called GDPO. Where you see the packages in sections. Usually if you are a Debian developer you know these pages.
25:20
Everything what is underlined is a link. You can click on the link if you download the PDF. Well it was a long list for our team. And we found out some subsectioning to find the relevant stuff inside the team.
25:42
We are using the ultimate Debian database for several purposes. For instance for the web sentinel citations. We have these prospective packages I have shown you. Which we are working on. This was the yellow section. This is also in UDD.
26:01
And we have also some use for the uploads and the bugs which were fixed by team members. I will show you some graphs later. And we have written the Debian mid policy. This is if newcomers want to know. How can I help Debian?
26:21
Or Debian mid actually. How can I become a member? And everything is written there. And this policy is also copied by some other teams. I think we basically copied the first version from the Perl team. Thanks Gregor. And what we developed a little bit in a different direction. Because it is also more user oriented.
26:43
And finally I started last year some so called mentoring of months. Mentoring of months is... Well I noticed that several people are really shy contacting us. I said well I have a question which might be too trivial.
27:01
And people will not answer me or will laugh at me. And I said well no please ask your questions on our mailing list. And others will learn from it. And I will mentor you for one month with a goal on a specific package. And on the other hand I have learned these people wrote an email.
27:21
Well I am quite enthusiastic. I will do this and that. And then you hear nothing from these people. And my turn is to... I also want a grip on these people. If I want to... If I spend my time to mentor you. I will also have the right to bother you with questions. Give me some report. What did you done?
27:40
So I have... In Free Software you have no grip on the people. They are volunteers. They will come and go. But if you try to get them involved you... Or at least I think it is a good idea if you try to catch them a little bit. And this mentoring is spending time on two people.
28:01
And it worked last year for four months. I had four students. Two ended up with uploads. One was nothing. And one was really really complex system. And we are really wide approach.
28:21
So it was not to expect to get it finished in one month. But we did a good progress and we hope to get it soon. So this mentoring of the month is also something I really would like to recommend to other plans or other teams. Because you get new people involved.
28:42
And regarding getting new people involved. I also thought, well, who are the people in our team? How did they... Were they related in the first place? And it has turned out that in our team with uploads and so on we have 23 Debian developers and Debian maintainers.
29:08
And the interesting point is that after 10 years we have 10 DDs who answered the questions. Are you inside Debian because Debian made exists? 10 people.
29:21
So one Debian developer per year came to Debian because Debian made exists. In other terms Debian would have 10 Debian maintainers less if we would not make this noise about Debian made. If you imagine that there are way more important projects than Debian made.
29:43
It's a leave project. We could get a lot of more Debian developers if we try to catch these people with these topics they are interested in. We have also 3 DDs including me. They were in Debian before Debian made existed and the other came by chance.
30:03
But these 10 people I'm very proud of. And 7 of these 10 people extended the activity to something else. They came because of Debian made and now they are doing other valuable stuff for Debian. They are doing QA, they are doing working on the policy, whatever.
30:24
So Debian, this is proof that Debian profits from Debian made and it could profit from other plans as well. And out of these 10 are active in Debian made.
30:49
I'm sorry, I forgot. I forgot what I wanted to say with this ad. Yeah, they are definitely all active, whatever.
31:04
Now I want to do some graphs because I also said, well, who is... The initial question was who is in Debian made because in the beginning I was not aware of this. I started with these graphs I think in 2008. It was in Argentina in the Debian conference.
31:23
And you see in this graph the uploaders of packages which are now under the hood of Debian made. Debian made started in 2002 but there were some packages before and I count them in. And what you see is the first 5 years.
31:41
This was not so much. It could be demotivating if you see that nothing happens. But at some point in time the people came and this is the effect I want to say. You really need to last long if you want to catch more people. An important point I want to show is please look here at 2011 and 2012.
32:05
Something happened. It is not the peak for me. I do not want to point this one. I am also not proud that I am on top here but I am proud on the fact that here are many people.
32:21
So we are a lot of people and not only one person who is active. You also see 2011, 2012. This is a graph of the activity of the mailing list. The names have changed a bit because here also are users active.
32:40
For instance, quite from the beginning the author of GnuMate is active here. He is no uploader. He is an upstream developer of the GnuMate practice management program. And he is always active from the beginning. But you see also after 5 years people are stepping by.
33:03
And we have also not only increased in the height but also in the mass of people here from 2011. The same effect. In this graph you have the bug hunter. So I just did the graphing. Who fixed the bug on the day-by-night relevant package?
33:24
The same graphics, 5 years nothing. Then more people coming by and from 2011, 2012 more people. And here you see the commits to the DebianMate VCS.
33:40
Well, here we start in 2005 with the VCS. So there is no longer history because we are using SVN since 2005 and in addition Git in 2010. And here you also see this increase. Does anybody have an idea how it might be that these two years show this increase?
34:06
Well, the solution is we did sprints. We did meet each other. And meeting these people means inviting other people in real life who have to some extent never heard of DebianMate or regarded it as something else.
34:22
So after these meetings we started with new people. Oliver Salut started in 2011. Laslo Kayan started in 2012. And these two people became under the 10 most active people.
34:46
We are more people than these. We are just graphing the 10 most active. So you get more people if you meet, come together, real life meeting one weekend. This is very helpful.
35:03
So here you see the same effect on a different graph. Here I am graphing the main packages. You see the topics are microbiology. This is a dark blue and a little bit softer blue is if you want to develop biomedical applications.
35:27
These are by far the tasks which are covered by the most packages. You have for biology 120 or more packages. And you also see it started slowly.
35:40
It started at all by the microbiology. And in 2011, 2002 the graph increases here and here. In the other field it is not that practical because the most people on these prints were microbiologists. But whatever you see in all fields that over five years it is not much.
36:06
Then you find the team and get the people and get something done. We hope that we also get something more for the medical practice. There is something in the queue. We also get something to cover the work in hospitals.
36:25
Hospitals is a really, really complex thing. This is really more deviant enterprise or something like this. Actually what I mentioned in the mentoring of the month in the beginning is something called Wistar.
36:41
Interesting enough it has nothing to do with Windows Wistar. It is way older. This is a complex thing and we need to do some work with the upstream developers. Actually this upstream developer was one student. Now we established the connection to upstream to get more things done.
37:02
You see what I wanted to demonstrate is you get something like a healthy grow. Because you are using the strategy of the blends to include upstream, to include experts, to find people who are interested in deviant because this project exists.
37:26
I have some warning, you have also seen it in the graphs. If the team is growing it doesn't mean it's less work for you. This is wrong. I learned this becomes even more work for me.
37:40
On the other hand my work has drastically changed. In the beginning I was doing all the packaging myself. Currently I'm more or less working with teaching people how to package, doing quality control, trying to be very strict with quality control and holding talks.
38:06
It was also interesting for me that being strict with quality control is if you have upstream who started some coding and never imagined that his code would be included in a distribution. Usually this is the first point.
38:23
Tell the people that it might be interesting to distribute their code in a distribution. Because I don't assume so. They don't maintain nice upstream tarballs as we would like to see this. Talk to these people and say would you be interested?
38:43
Well yes and would you like to do the packaging? I can't package. Would you agree that I explained how to do it? Yes, we can see and some of them agree. In this situation I'm always thinking should I be really picky as I'm for other people.
39:04
Should I start a little bit more relaxed about Lin-Cham warnings also? I learned no, these people want to be picky. They want to make really good quality packages. This is wrong to assume that people who are not so really deep in the technical details.
39:25
We have frequently to do with scientists which are picky on the algorithm to do their science but not on the operating system side so much. But they are interested in getting really strict education in this and I have very good experiences with this.
39:48
So my strategy is making the blend tools even more attractive to on the one hand the users and developers
40:01
and making the blends idea more attractive for teams inside Debian. I'm talking about blends since seven years or so, maybe longer. And the idea has not yet the grip I wanted it to have
40:21
because Debian developers are to some extent very focused on single packages. Not all but you have a lot of Debian developers who have a focus on package, package, package. But my idea is to put the focus on a set of packages
40:40
because there is no single user which uses a single program. Users are using sets of programs and these sets should be more in the focus to work together and to form a team around. So the package should be maintained in a VCS and not somewhere else.
41:00
And this VCS should be accessible for more than one person. So this is, we have a movement in Debian, a strong movement in this direction. But it's not as strong as I would like it to be. Well this would be all for my talk. We could start with questions or I could do some more demonstrations on these task pages,
41:23
whatever you are interested in. So I'm just asking for questions now. And I can easily fill up the time until seven or you want me, you want to stop me and to go to dinner or whatever.
41:42
So you said professionals are not really using free software. Do you have any idea how, who the users are? Well you mean for the practice management systems? For instance, yeah. The authors of this practice management system are, as far as I know,
42:02
all medical doctors themselves who gathered some skills in programming to a different extent. This is, sometimes this is a problem. Yeah, but I have seen one practice management program
42:20
which was written in TCLTK. And it was distributed with PostgreSQL database server with some fixed version in the upstream table with object and source. And 10 other free software products, everything in one thing. And put this on a hard disk and start this also.
42:44
This is a no-go. And there are other doctors who are quite good in programming. And you must look at the code because there exists perhaps 15 programs who claim to be able to manage a medical practice.
43:07
And 10 of them aren't. Okay, so thanks for this information, but in fact I was asking about the users. Yeah, okay, sorry. And the users are basically the authors and their very close friends.
43:25
Okay? So you can say the user base is about zero. This is how life is currently. And I do not expect this to be different in the next couple of years.
43:41
It is also the same, even longer, zero usage phase I have in the graphs. And they are doing also some conferences and there is, I think there's GlueMate. We can check how many people have installed GlueMate.
44:05
This is in practice management. So there is Clinica, which has a popcorn of three. Clinica is a very, very light system which was just programmed by a friend of a doctor in a one week thing
44:26
and it's actually not really, well the name practice management system is not rectified for this Clinica. And Tango is more or less to control cameras. This meta package contains everything which could be useful in medical practice.
44:43
Not necessarily only practice management system as it's defined. Then we have FreeDiams or FreeMedForms. This is also, FreeDiams is also to manage medics.
45:01
Or it has a database about the medics which can be used. It's also used by GlueMate users. So we have a popcorn nine. You know popcorn, not everybody needs to agree to popcorn. But this is basically what we can say. FreeMedForms should be installed by some users.
45:23
GinkgoCatX is a very interesting medical machine system and it's quite professional systems and you see a lot of users. This is very useful in a medical practice. And the GlueMate client you have seven users. And the server has 18 users.
45:44
So it's basically programmed by a German doctor and it's basically used in Germany because nationalization. So let's assume 10 installations and three of them is running the other themselves and some are in other places of the world.
46:02
Well, what should I say? This has cool features because the author implements everything. He's missing in the proprietary software. And the problem of this specialized software is
46:21
the proprietary programmers are not practitioners. They don't really know the needs and these people know what they are doing. This is the main selling point I would say.
46:41
It obviously doesn't sell enough. But yeah, this is what I can say. Yeah, that's everything. This is what we are in preparation, this OpenEMR. This is a bit better covered by users.
47:01
It's not Debian package ready but I know that it's used by more people. And the reason is that the main programmer or the program team has a very vocal person who is running on conferences and say,
47:22
well, use this, it is very good. And the more noise you make, the more users you have. The GNU-made people are a little bit shy. So it's also in free software development, there is some advertising, marketing stuff involved. And we try to get this as well.
47:42
It is a web-based system. Yeah, we'll see. Is this question answered so far? Any more questions? Can I show as well?
48:02
So we have these tasks in Debian Made. It starts alphabetically with the biology and we have also here some task cloud which is a subsection of all these packages which is just command line based.
48:22
You can run a cloud and do some gene sequencing on clusters. And we had formally something about dental practice. There was even, it was similar. In Debian it was a dentist who wrote a program
48:42
and built a package. It was in 2003 or 2004. But this dentist moved away and this, if you don't develop, keep on developing it, so it was removed from Debian. So there are some remaining, but we have no support for dental practice currently. You can run this GNU Made in principle
49:00
in the dental practice, but it's not dedicated for this. We have epidemiology, which has basically several R packages. And we have something about psychology.
49:23
Oops, why does it show up here? Something is broken. This is, I think for this specific field, you have a lot of users and it's quite active programming team
49:42
and we are always lagging behind with upstream versions, currently because of the freeze. And yeah, something from the neuro Debian team. Neuro Debian team is some very, very active people.
50:01
In principle, they are also forming a blend. We want to start using this and they have also a cool website. They just, to some extent, they do even more than the Debian Made team because they are running on conferences and have a booth for their own project.
50:22
But here the problem is that they have invented their own tool set and not using this one. And we try to merge this a little bit from the statistical point of view. So I think we should stop here if there are no further questions. Thank you very much for your interest.