Debian Contributors
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 |
| |
Subtitle |
| |
Title of Series | ||
Number of Parts | 199 | |
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 | 10.5446/32506 (DOI) | |
Publisher | ||
Release Date | ||
Language |
Content Metadata
Subject Area | ||
Genre | ||
Abstract |
|
00:00
Group actionTerm (mathematics)Source codeConfiguration spacePlanningEndliche ModelltheorieMereologyElectronic mailing listSoftwareMultiplication signData miningWebsiteView (database)Division (mathematics)CASE <Informatik>Key (cryptography)Video gameBitDecision theorySystem callRight angleSoftware developerOffice suitePhysical systemForcing (mathematics)Game theoryPhysical lawState of matterWeb pageEmailStability theoryMoment (mathematics)Social classDivisorPoint (geometry)Price indexScheduling (computing)Computer fileAnalogyUniform resource locatorTranslation (relic)Thresholding (image processing)Sheaf (mathematics)Parameter (computer programming)Flow separationException handlingConfidence intervalSoftware maintenanceSoftware bugCommutatorLoginMusical ensembleData managementBuffer solutionProcess (computing)LogicRepository (publishing)Factory (trading post)Product (business)Token ringLine (geometry)Form (programming)DemosceneVirtual machineCartesian coordinate systemFormal grammarLink (knot theory)QuicksortGoodness of fitStatisticsGreatest elementFormal verificationSolid geometryProjective planeScripting languageGastropod shellState observerFile Transfer ProtocolBeat (acoustics)BuildingStaff (military)Statement (computer science)Order (biology)Identity managementNormal (geometry)SequelInformationSpecial unitary groupNegative numberCore dumpQuery languageExecution unitClosed setRange (statistics)DatabaseLimit (category theory)Complex (psychology)SummierbarkeitSound effectData storage devicePower (physics)TouchscreenTask (computing)Entire functionAreaNatural numberOptical disc driveThermal expansionImplementationAxiom of choiceWordNumbering schemeArithmetic meanConformal mapInformation securityNumberPresentation of a groupRule of inferenceWeb 2.0Game controllerSlide ruleWikiGradientDescriptive statisticsFingerprintSet (mathematics)Incidence algebraData structureMathematical optimizationLogic gateHeegaard splittingCountingSingle sign-onCoordinate systemINTEGRALType theoryProgramming languagePort scannerTraffic reportingAddress spaceDirectory serviceTimestampInformation privacyAuthenticationCollaborationismRandom matrixSinc functionExistenceTrailField (computer science)Graphical user interfaceRow (database)RandomizationVotingMathematicsWritingCodeEvent horizonDependent and independent variablesComputer virusOcean currentFocus (optics)System administratorProof theorySorting algorithm1 (number)File formatWeb applicationDefault (computer science)Lecture/Conference
Transcript: English(auto-generated)
00:07
Hello. Welcome. I'm Enrico Zini, proud Debian contributors, which is a new hat, a new role,
00:22
a new title that we are creating in Debian. That's my advanced slide tool. The people
00:44
in the back, can you read it? Okay. So, the concept we are moving in, just to contextualize
01:01
where we are, is the Debian diversity statement. Before I put this slide on the screen, who among you had never heard of the Debian diversity statement? Okay, a few. So, maybe a year,
01:23
year and a half ago, possibly two years ago, Debian accepted this as an official statement, as an official document in the project. And this basically, it's a text I like a lot. And it says that the Debian project welcomes and encourages participation by everyone.
01:45
No matter how you identify yourself or how others perceive you, we welcome you. We welcome contributions from everyone as long as they interact constructively with our community. While much of the work for our project is technical in nature, we value and encourage
02:03
contributions from those with expertise in other areas and welcome them into our community. And it's, we documented what basically most of the project was already feeling, which was interesting because we had a huge month-long discussion on the Debian project
02:26
mailing list without flames. And it was only about the nicest way to paint this beautiful shed, which I think came out very beautiful indeed. Now, that's what we feel.
02:43
We have a problem in that most of what is visible at the moment of Debian contributions is package uploads. However, there are several ways to contribute to Debian. Like we have Debian developers, maintainers, Alioth committers, bug reporters, wiki
03:04
contributors, translators, mailing list contributors, and so on. And the list goes on. We don't even know how many kinds of contributors we have because the edges of Debian are blurred. It's so vast an ecosystem. People who report a bug,
03:27
did I write bug reporters? Yes, are also contributing to Debian, although they currently have, well, they previously had no title previous to Debian contributor. People who edit the wiki,
03:41
who do translations, who run an event that Debian never hears about. Somebody running a fantastic Debian booth in South Africa, we may never hear of them,
04:01
but they are Debian contributors. So we are failing to acknowledge this. However, if we are to spend time making a list of lists of possible contributors, then it's a mess. Therefore, Debian contributors. Easier, right? All those people, now we can call them Debian contributors.
04:25
New title in Debian. And we say Debian is a do-worker, see? That is, in order to change something in Debian, you just get on and do it. There's no voting on how we want to change things, with the exception of init systems, that I do not want
04:46
to discuss here. On the other hand, that is another case, probably where do, where just getting on and doing it would have helped more than trying to discuss first. But
05:00
yeah, the issue was, well, anyway, we do not want to talk about this here. So generally in Debian, you want something to happen, you do it. We call that do-worker-see. And that's a title that follows the same idea, because there is no process of approval.
05:23
When you contribute to Debian, you are a Debian contributor. That's the definition, by definition of Debian contributor. The first rule of Debian contributor is that if you are a Debian contributor, it also applies to Tautology Club, and I'm very happy about
05:44
a nice self-sustaining definition. And when you are a Debian contributor, you get a fantastic privilege to have your name on a list. Right? Great. Because before this project, you didn't even have that.
06:05
You can also choose not to have your name on the list. So you're not going to be a Debian contributor whether you like it or not, because I believe that may put some people off contributing to Debian. You can see a summary of all your contributions get acknowledged,
06:22
you get thanked, you get reputation for the work you do in Debian. And it is unacceptable, I believe it is unacceptable that we still fail to do that for so many parts of Debian. Who's translating Debian in Bengalis? No idea. Does any of you know?
06:45
Is Debian actually translated in Bengali? Possibly, yes. But that shouldn't happen. There should be a page listing people who do things, because they are the people that make Debian happen. Therefore, we fix it.
07:07
So we fix it by having a list. A list. There you go. Fixed it.
07:27
Federico, are you here? Federico, Federico? No, shame. First contributor of February! Okay, I may come back to that. Anyway, yeah, we have a list.
07:55
So, that list is built by data that is sent from many teams in Debian who set up their own
08:05
data mining. Each team takes care of its own data mining, because Debian is too diverse to expect otherwise. So, the first design decision I took in building the site was
08:23
that I'm not going to do data mining, because then it would not scale to the size of the project. In my whole life, I may never know that some bit of Debian exists. And I'm not joking, it's not an exaggeration. It's that vast. Now, since teams will be responsible for
08:50
saying who are, for listing, contributing, who are their contributors, teams that currently have no motivation to acknowledge contributions get a reason to.
09:04
There are teams that need to get a job done. People do the job. The job gets done. Fine. Nobody has the time to think about tracking who does what in order to thank, because maybe it's all nice, selfless people that only care about getting that job done.
09:25
Well, excellent. But it's nice to have a motivation for people to do it, and of course make it as easy as possible to do it. And it would be good if our full range
09:41
of diverse contributions finally has a chance to become visible. I'd like the list to also show all bits of Debian that we know of. I want somebody who would like to contribute to Debian to go to the site and use it as a shopping list to see what people are actually doing,
10:05
and saying, well, I don't like to translate, I don't like to package things, I would like to build web applications. Oh yes, there's people who do that. Let's go and get in touch. Oh, there's a link to the team. Let's join the mailing list. I would like that to happen.
10:27
For bits of Debian, we don't suspect are there. Anyway, Debian developers, the way we see them, are not really software developers, are more like
10:43
DevOps's admins, writing little makefiles and shell scripts for deploying stuff into machines. But then people who would like to develop software, currently it's hard to see how to
11:00
get into Debian, because we package software that has already been developed by others. Although, well, there's people who develop software, like me, building that list or data mining scripts. So again, I want to change the image of Debian from that of people packaging
11:22
software to what Debian really is. A huge ecosystem of vast integration work, coordination work, non-technical work that is as important as everything else, like well-being at the Debian booth, selling t-shirts, designing this wonderful t-shirt. That's as good because
11:47
otherwise nobody would know about Debian, right? So, make that visible. In the list, there's a problem of people missing in action, so somebody stops
12:11
contributing to Debian, maybe they're still on the list. Well, we just split the list by year, and so the list is self-managing, because the people who haven't contributed
12:27
in 2014 are in this list, but not in the next one. So, as little management as possible, and if somebody stops being a contributor, because they have something else to do,
12:42
then automatically they are still listed as they should be, but we don't get false data in the current contributor list, because we're not selling marketing by number of accounts, we're just thanking people who do something. So, it's a list that manages itself,
13:09
and Debian developers are not automatically Debian contributors. It's about what you do, not about what you are.
13:21
Which maybe the Debian missing in action team may appreciate, but the point is not to trap people, it's to thank people who do things. In terms of design choices, now this is a technical solution to a social problem, which is something we are very good in doing in Debian,
13:42
although people say it's impossible, but we don't care about what's impossible, or we wouldn't be here in the first place. But anyway, since the focus is social, the design choices are deliberately simple. I had to make an effort, not to make an effort,
14:06
due to my obviously OCD nature of, you know. So, the point is we make a good effort to thank people, but we don't need to go out of our way to do so. It's fine to say thank you for
14:22
fixing that thing in the wiki. Thank you for fixing stuff in the wiki last month is enough. I'd like to go and say thank you for editing the third line in the fifth page in the wiki with correcting a grammar error, because it's unnecessary complexity in the data mining.
14:44
When in doubt, relax the technical requirements. Time granularity, do you track the exact hour of the contribution to the minute, to the second? A month. It's plenty. Store begin and end time of contributions, you contributed since February 2005
15:06
to March 2010. Fine. If there are gaps in between, whatever. If somebody didn't contribute in 2008 and 2009 and then came back, I still say from 2005 to 2010.
15:24
It's too much hassle to track gaps, so we don't really need it. Latency is okay. If you make a contribution and it shows up a week later, two weeks later, fine. We will eventually thank you. Which means a complex data mining task that takes three
15:48
hours to run can be scheduled once a week. There's no point, there's no real gain to run that once a day. It's okay if some contributions are lost as long as most are seen. I mean,
16:07
if we do data mining of the whole Debian backtracking system to see who reported the bug, maybe there are spammers, without maybe, that there's plenty of spam, but we can say, well, you need to have interacted with more than five, ten bugs.
16:27
Because that's easier to filter out spam if we raise the threshold a little bit. And it's okay if after a while you'll end up in the list after a while of using Debian and reporting the odd bugs and replying to the odd bug. And that's fine as well.
16:45
We're not trying to track each little my new thingy, but, you know, see who's there doing something and thank them, generally. That was explicitly asked at the
17:10
Cambridge Mini Conference. They said it's unacceptable to have a conference without cats, so I know that there's an issue with privacy.
17:21
One scenario that could be quite scary is that you are at the office playing Wesnot because it's quite morning and it segfolds and there's a helpful dialogue coming up saying, oh, would you like to help the developers fixing this segfold? And you say,
17:43
yes, of course, I'm very glad to help the developers. So you click okay. And so a bug report goes into the Debian BTS and then a couple of weeks later you appear in the list of Debian contributors saying thank you for helping the game team
18:02
during your work time. A time granularity is a month, so maybe it's still fine, but, you know, that's unexpected behavior. I don't want to turn every Debian activity into a scary thing that will go and tell the world that you are something you didn't expect to be.
18:26
So ideally it should be opt-in. Maybe you get an email telling you how to be on the list after a while that you show up as a Debian contributor to the system. Do not publish email addresses by default. You report that bug, you get spammed, maybe not
18:44
a good idea, although that is so at the moment because the BTS will publish emails, but, well, I don't want to add privacy issues. And allow people to manage what's credited to them.
19:02
If somebody wants to have two online personas, they should be able to. Maybe I'm one kind of Enrico doing work and one kind of Enrico doing my private time. I may have the game people doing private time and the networking people doing work and I may prefer them to be shown
19:27
as different personas. So as long as I use two GPG keys, otherwise be quite hard, or even if I use only one GPG key, I should show contribution to that GPG key in one side,
19:43
contribution with that email address on another side. I don't want to be snooping on people, spying on people, reporting on people, none of the kind. I would like people to be in charge of how their contribution in Debian will be presented. Hide some kind of contributions. I
20:04
don't want my contributions into PHP packaging to be shown to the world, which is something I cannot do on Ollo. They still believe I'm a top-notch PHP developer because
20:21
I don't know what I did sometime that they mined and no way to tell them to get my name off their site. I don't want to make another Ollo that decides they know better than you who you are. And also to merge and manage Debian identities. Some bits of Debian understand you,
20:45
well, think of you as an email. Some bits of Debian think of you as a GPG key. Some bits of Debian think of you as a login name. And I should be able to say, well, I'm Enrico, these three emails is me, those two key fingerprints is me, those two logins are me.
21:07
Then since I'm a Debian account manager, I would still like to see everything, though I promise I won't tell anyone. Because when somebody applies, and that's negotiable,
21:20
but I wouldn't mind if that would be accepted because it would make my life easier. If somebody joins Debian and I could see all their contributions just to say, yes, of course, you've been so active in Debian. Perhaps people can say, I will only look if people saying,
21:42
go and look. But it will be interesting to have privacy conscious access for limited amounts of people. But yeah, okay. Pretend I didn't say that. Now, in terms of teams,
22:04
teams in Debian that submit data to the site, submit it must be as simple as it can be, or nobody will do it. Now, if somebody has two hours a week to contribute to Debian, I'm not going to ask them to spend one hour a week sending contributions data to the site.
22:26
I'm not going to ask them to log in the site and click a button every time you contribute to Debian. It's something that should happen automatically and should be easy to set up and easy to maintain. The main activity of a Debian team should be to do what that team does best,
22:43
not to track their own contributions. Otherwise we become a bureaucracy, which we are not good. Some teams can track their whole history. FTP master has a SQL database with records of all
23:05
uploads in Debian history, at least after a year, some year. Some other teams only know who are the members. Yeah, that person is a member, that person is not. Fine. This range of team diversity
23:26
should be able to send data. Data is collected, is received by the site using an HTTPS post of a JSON file plus an authentication token. Basically, one cool invocation.
23:52
It's a JSON file that is simple to build and post, and I made a script that does all of that for you, although if you don't like the script it doesn't take much to do it by yourself, as
24:04
Steve can witness. For the Debian wiki, it went like, oh, I'd like Debian wiki information to be submitted to the site, and Steve went, okay, is it okay by email and date, star date, and yes, sure,
24:23
great. And it's like, well, I'm scanning the thing with the daily maintenance. Anyway, how do I send it to you? I said, here's how you build a JSON file. After 10 minutes of typing Perl, data was in site. It's really that simple. And you can build JSON with almost any programming language nowadays.
24:52
As we speak, the tool to do data mining is getting installed in several machines.
25:01
So, in possibly a day or two, there's been a hiccup this afternoon, or I could be able to announce it even here, but soon indeed, matter of few days, anyone who can access a Debian or machine can set up a cron job with some data mining, and I'll show you how it works.
25:28
And I'll show you how it works. So, can you read it in the end? Bigger? Better? Okay.
25:56
So, this is the site. That's the main list. It's logins at the moment because
26:07
there's still no identity management implemented to choose if you want your name to be shown. So, I figured that your Debian or Ali of login would be safe enough to show.
26:30
There's a detail of contributions. So, this person maintain packages developed
26:46
as email. Well, again, if somebody wants the email to be public, here will be the email addresses. Well, the presentation is still being worked on, but these are all the
27:01
various identities of K-Stoff, and these are the contributions done under each identity. Yes, just because there was a date field in Django and there was no month field in Django,
27:28
but we can tweak the rendering to only show the month. Indeed, that could be a good idea. We have a list of data sources.
27:51
These are... Forget about the proof of concept ones. Those were my experiments when developing.
28:07
Wow, I didn't know of this. That's great. Thanks. Okay, sorry. So, these are all the
28:33
teams that set up some data mining. Now, if you are a team, say you are the Debian
28:43
new member process, you automatically get your own team-based statistics with sortable columns.
29:23
So, the data broken down by contribution type, by team. As team, you can say, let's see who are the application managers active at the moment in the new member process, and there you are. Nope. So, that's an extra value as a team. You set up some data mining,
29:49
you get your own team page, so you no need to develop your own. Yay. And then anyone can set up, anyone who can log in the site can create a new data source.
30:09
So, let's create a data source for Debian presentations.
30:27
A URL with a list of talks, an authentication, that doesn't exist by making it up,
30:40
an authentication token, secret that's used to send data to the site, and some implementation notes. Submit. A Debian developer can do that, and there you have it.
31:24
Automatically, you get listed as a dump administrator. You can add more, and it's just for documentation purposes. Who do you send an email if it breaks? And you edit, and it tells you how to post data to it from the command line,
31:45
and you can say what type of contributors are. So, you can say speaker, description.
32:06
This, it needs to be entered from both sides, because when it gets pasted into web pages,
32:25
you can say this person is a Debian talk speaker, and this person, and this is a list of people delivering talks. So, it needs both strings. It's a bit structured,
32:41
so it's easier to translate, hopefully. Submit, and now you can start sending begin and end date for speaker activity in the Debian presentations theme. That's it,
33:05
as far as setup goes. Data source management is over here, and one easy thing we can do is see if there's a data source that's broken. So, time of latest submission.
33:28
This data source last submitted data three weeks ago, maybe it's broken. Yes, of course, it's a proof of concept. So, that's as much as maintenance of this site goes. Okay, it's broken. Let's send an email to Enrico. Done, which will tell you it's a proof of concept.
33:48
Hopefully, this will be picked up soon by people actually maintaining it. So, even management of the whole site should be just about programming the features in it,
34:01
but it should be sort of self-managing. People can go in, configure the source, and set up data mining, and it's a simple matter of self-documenting web forms. So, should be manageable. Debian presentations. Ah, what's that? No, I don't want that. Delete.
34:41
It's interesting, if you'd like to show up as a Debian contributor, and your team does not currently send data, well, just create a data source for your team, as Federico has done, who is, therefore, having written a very low latency submission thing, is the first
35:07
contributor of February. So, yeah, if you want to show up as a Debian contributor, write the data mining for your team and make sure you are in the list. Okay, so about writing data mining scripts,
35:30
there's this command line tool called dc2. You write a configuration file
35:40
describing how to do data mining for your team, and you run this command, and it will go, do data mining, send data to the site, and it's done. You just put that in cron, and done. Now, what's in the configuration file? Authentication token,
36:09
configured in the site, or it wouldn't be able to send the data, and several sections, one for each contribution type. So people who are committer to some git repository, you say
36:30
committer, scanning method git dirs, and the git repository is in here. For each contribution, you build a URL to the committer, to show in the site, in this way. That is all you need
36:53
to have a Debian collaborative maintenance data source. That and one line in cron running it.
37:05
I'm not sure it could be made easier than this. It can. That tool needs to be installed on Alio. I'm working on it. After which anyone, even people who are not Debian developers, can set up
37:24
the data mining. You only need a Debian developer to set up the data source and put the authentication token in it. That's it, right? Scanning methods, it looks directory tree,
37:45
look at file owners and file timestamp, and if you own a file with a timestamp of three days ago, it means that three days ago you contributed something. You can do that on a CVS
38:03
repository, say the Debian web team, CVS, and you have the Debian web team data source. For example, that way. It's nice when you can give four line examples that could be put into production, and can work as hint hint, wink wink, not too much,
38:29
everything documented, what you can put, and so on. This will scan git directories, it's a bit like the file scanner, but will only look at the right bits of a git directory tree,
38:42
looking for file ownership, because if you look at git log, you can have any name in it, maybe some processes that give commit access to the people actually committing prefer to look at files, depends. Anyway, if you prefer to scan the git log, you do it that way.
39:09
You can, it can scan mailing list archives, even compressed. The tool is already installed
39:21
on master Debian org, which has all the mailing list archives since the beginning of Debian. I wouldn't be sure about since the beginning of Debian, but perhaps a few months later. So it will just look at who's in the from header of emails in that,
39:45
in the folders found there, and you can add the black list or white lists. So if we are scanning Debian dev announce, the list is moderated, so every from address
40:00
is a good contribution. If we are scanning say a team mailing list, where, sorry, here, a team mailing list, the team has five members, but there's other people posting in the list, then you put the five member in the white list, and run the scanning,
40:23
and you see when those five members have been active. It's limited, it does a portion of Debian lists, it's not so smart, but it's a start for several teams. Now we get better with time. You can run a SQL query on a Postgres database, so anything you can mine off UDD
40:46
or the FTP master database, write the SQL query and run it, and you have the data source. For example, that's how you get all package uploads into the site, including a nice URL
41:04
to link at the developer PHP page for that person. Or scan subversion repositories, that's it at the moment. Any new code that I add to it automatically adds to the documentation,
41:29
because the tool is able to write its own documentation. So yeah, there was a bit of engineering there, but still interesting code to read.
41:44
So yeah, let's say we take Colab main subversion, Colab main data source,
42:29
which looks like this, look at Git and subversion directories, straightforward, right? Add this, and it explodes. Data source, the authentication token in the file,
43:17
the stash post, and you're done. You can schedule it in Chrome. The reason I haven't
43:22
done it is that I don't want to take responsibility to maintain a one-liner Chrome, well, 300 one-liner Chrome jobs. Then I can talk to the maintainers of Colab main, I hope that's not me, and ask them to set up a Chrome job. And still you get a preview
43:46
of the data that would be submitted. The submission itself would look like this.
44:00
So you have several records like this, and you say this person identified by login with this login has done contributions of type,
44:21
one contribution, well, has done contributions of type committer from this date to that date. The magic of JSON makes it in random order, but still, that is the file format. So it's still easy to write it by hand. That's about the status of it at the moment.
44:48
As this tool gets installed in Debian production machines, I'm going to start sending announcement to Debian Dev Announce with all the example snippets saying, can you please
45:01
Chrome schedule everything. Timeline for now is, well, this bit, have at least FTP master starting to send data, because that means that the site will be as good as anything we have now
45:20
in terms of showing contribution, except slightly better, because currently we know about package uploads, we know about that. So the site should at least show that, and then anything more we can have, and we already have the wiki, it's enlarging our view of Debian. It's showing us
45:44
a bit more of what Debian actually is, and we will never, I believe, cover a hundred percent, because Debian is fast, still. In about two weeks, there's going to be a Debian
46:04
single sign-on sprint in Coln, where we are looking into how allowing anyone with an Alioth account to log in Debian single sign-on site like this, and that is the first step into having
46:21
everyone doing identity management. At the moment, I only show things that can be related to Debian developers or Debian maintainers or people with Alioth accounts, but that's quite restricted, because even people with Alioth accounts, not all of them. I'm being
46:44
extremely careful not to surprise people by putting them on the list and not being allowed to take themselves off, but as soon as anyone can log in, then identity management should be, I'll implement identity management, and then you can log in the site and say
47:04
I'm also that email, I'm also that email, that fingerprint, that gpg key is mine, and so on, and build your online identity, show how you want to be shown, maybe add a short bio for yourself, have your own little contributor page, why not, and that should be enabled about,
47:26
should be thinkable about two weeks from now if everything goes well in the sprint, so hopefully in a matter of few months, we should start looking into the faces of people who have done something in Debian but we never thanked before, which is something I really much
47:44
look forward to do, and that's about my presentation, I don't know if there's time for quick Q&A, one question, two, one, one or two questions, one, two?
48:04
I said convenient for me to get to, what was hard about that? I can repeat the question if you want. Christoph, and then, okay, oh well, okay, you too anyway, you won the competition. Just an idea, like a wishlist pack, eventually we can separate, you can separate the list of
48:30
teams which could be, have volunteers, contributors, and the actual implementation, so there's a list, there could be a list of teams, you don't know, maybe, but something else
48:47
does know, and if there's a list, it could be like a wishlist pack for the team to implement, to contribute, and in, for example, the DEPCOM volunteers.
49:02
Like a request for packaging, but for teams, yes, very good idea. Good. What do you do with teams that do not send historical data, just current ones, do you keep the old one or just show the current one? Good question,
49:28
teams that don't have historical data can't go, can't really go back in time. On the other hand, so that would be really difficult to track what happened five years ago for that team,
49:42
however, rule number one of Debian contributor design is relax the technical requirements,
50:10
therefore, we just don't back, back history, and my plan would be to have a configuration in each data source saying, I don't know anything from before that date, and then,
50:28
and then when you show a person's contributions, let's say, random person,
50:42
you, in here, it would say, since before we have data, so a team with no history will start sending data, then from that day, the site will keep history, because each day, the team will say, that person had a contribution today, or is a member today,
51:05
and the site remembers that first seen as a member on that date, and then going on, and you can get a threshold saying, anything from before this date shows as, from before we have data, and that's fine. Okay, those were the two questions,
51:30
thank you for listening.