 So, maybe today we can show you how to use Kutog, what is Kutog? I'm not sure, it's very interesting to show you how to install Kutog, but anyway I can... But I think we should skip that part because I think the main point would be more talk about the Debian Kutog server rather than... More about the usability concept of Kutog, what can be done? What we have currently? Okay, so this is the Debian Internationalization Server Kutog server. There are already some PO files which were installed by Christian. Mostly for those projects of PO Depomp, Debian Installer. Just for the native package, just at least changes. There is also a Kutog project, it's used for the translation of the web interface. So you can make some changes, but please don't, it's better to do that in the stream Kutog server. And the DDPTP project will be... It's currently not here. Yes, and it will probably be used at the end of Kutog. Well, at least filled with the real Q5. Yeah, so the entry is both by language or by project. So the concept of Kutog is everything is arranged by project by a language. So you have a combination of both. So do you have to live with that in the Debian server? Currently no user is logged in. Maybe we can have a look at what regular user which do not have an account can do with the Debian Internationalization Server. So the fastest would be the Debian Installer. Okay. I hope so. Did you update the... I don't know, I was waiting for a presentation. If you don't want to break this. Yes. So we are currently running a 09 whatever version of Kutog. Yes, 10 is ready to go. And it was a pretty much hacked version by Nikola for many performance issues. So the time is long. Is this one task? It may be quite long to open a project. The first time a project is open after the Kutog server is restarted, the server is reading most of the files to gather the statistics. Okay. It's not really reading all the files, it's reading some leakage for the statistics, but it makes a lot of files to read. So here we can, maybe we can have a look at something more interesting than the French language. If you can find... You should look Amaric at the beginning. Yes, because there are some... There are some un-translated... Some, yes. A lot of them. Just be careful, it's not a number of strings, it's a number of words. So don't be afraid if you have to translate 1000 words. It's much faster than 1000 Gtex messages. Okay, so... When Christian installed the PO file, he tried to have the same architecture than on the current... With all levels and etc. But for those who translate the DI, there are not many people translating DI in this room, so we won't go this way. The most important... What's called level 1 is the core of DI. So there is only one file. If we have... In this language, there is no level 2, level 3, 5, so it's not listed. But for other languages, we would have seen level 2, because they imported level 1 and 2 only. And these files, the important part is they are linked with the DI SVM. So the AM that PO will see here is a copy, is an SVM check-out of DI. But not only check-out, but you can check in the work. So that's the main new feature I will start with this organization. Okay, so the first step will be to click here to see more advanced functions afterwards. Then we see there are still several performance issues in this place to talk about. As you can see. The main problem is the total development value. They didn't test widely with a lot of things. So we may use a lot of performance. The largest source they have is OpenOffice. Yeah, OpenOffice. OpenOffice is very large, but not many files. Bitfiles, but not a lot of files. And we have small files and a lot of files. Well, I think DI is a little bit different from Debcom files. Debcom files, we have a lot of PO files. And for other ones, we have one big PO file. I don't really understand why it's taking so much time. It's very slow. Not faster reactions than I left. Yes, maybe... There are some more years over there. So it may be a solution. I think there is one very close to... Yeah. The yellow one is longer. Just the content of the PO file. It's like an offline translation tool. You can see the train in the translation. I think this train is cut because it's too long. But if you put the mouse over... The size is increased. If you double click on one string... Okay, so you're not logged in, so you can't do it. No, it's impossible. You can only view because you're not logged in. So we might need to log in. Can you log in? No, I can't. I can write. I will log in. So one user wants to log in first off to register. Then you can send the user a mail with some registration information. And afterwards they will be able to activate their account. Does it work with HTTPS? No, I don't think so. Definitely mouseable. Not of that survival, but I think the puto server do that. It's possible to use a patchy to proxy the... I don't know if you have seen the difference while we were explaining. I logged in as W2, ordinary user. This user is registered for French language only. So he just only sees French and is registered for pure.com, PII, etc. So he's first registered himself for French. He said, okay, I register for French language. And then you can register for various projects. For each of these projects, it can be given different rights. Okay? I go here. I immediately jump to the I and then another one. So we see French again. So I go here. I go click. I go click. I get additional functions. So Google One is just a standard user, so you cannot change the translation, but you can make some suggestions. Oh, yes. This one can translate. So basically, you can define an administrator for the French team, and this administrator will be able to give some rights to the various users, and it can say that Google Two can only suggest that Google One can submit a translation. There are other rights which can be defined like synchronization with the subversion repository or other features. That's very complex. In the administration of Google, there is a big generality. It's a little bit too big because it's not very easy to manage. Many projects, many languages, it requires to give admin access one by one to each project. But the good point is that you can delegate some administration processes. So there should be now two to observe or there should be one of you admin team who just manage the top level rights, language by language, and then project by project. So in the next version of Google, you will have also a button to say that this translation is fuzzy. Sometimes it's useful just to say no, it has to be really remote. You can also add some comments. It was something missing and now it exists. Otherwise, I don't know if it's interesting to see all the functionality. It's just a translation tool. I think any user can use it out of the box. There are no instructions to give to the users. It may be a little bit slow. This country is slow. I think a man, not too much, not too many people, we want to really translate online. At least for the team we currently have in the end. It will be more convenient for some new teams that are popping up. We are translators, don't know how to check out, don't know how to launch cableball or POS or whatever. And these people translate most of them online. Some of them go to an internet cafe somewhere and just translate online. So that's the man for that. But already for us, the first interest would be to give access to the material. Our main challenge is to import the material first to the server before we can really use it. And at the beginning I think we could probably use it as a gateway to purify. Compared to what we have currently for the AI and things at the end. Well the challenge is to import the DDTP. Don't have no main interface, what? What is your main interface at this point? There is only a web interface. All the remaining tasks you do them on the server. And the first check out I made for the AI. I have tried to do it anyway because you can see at least the remaining functions if I can remember my password. So on the PUTL server, the PO file I just installed, you can just put the PO file and it will work. The PUTL server don't have to be restarted. You just put the new PO file and it will be dejected when you navigate on the interface. Do you remember PUTL, can we use a section to show without key still? We can still do it? Yeah. The German people. Yes it is. Where is the why? Okay. On the web interface there are some other functionalities. We only saw the use case where the user opens the PO file and then translates the first string, the second string etc. But of course when the PO file is translated like 99% you don't want to see all the strings. And there is some filtering functionality like you can export, you can skip the translated strings and just translate the frizzy and the translated one. There are some other functionalities. We saw that the user can make some suggestions. Afterwards the main translator can reuse the suggestions and validate them or just reject them. You can suggest, you can validate and then you can commit. This is what I probably have in the last. My user with Google Chrome has everything, all rights on all languages. And if I go to DI and hopefully it will be faster, I go to level 1, then effort.po and file arch. Now I show editing functions. This sucks. You have to always activate this. Show editing function. This effort.po now I can also... Maybe just for it here, we can allocate some strings to a user. So we have a big PO file with 1000 strings. You can say that the first 100 strings is allocated to one user. The second 100 of strings is allocated to another user. And when the user log in, he can click on translate my strings. This is something new for Bitfine. Bitfine could be convenient for the future. The other part is not working yet, right? The user cannot allocate the strings yet. No, it's working, but here the file is fully translated to... So the user now can allocate some strings for him? He can, because he can't do anything. I don't think the user can. The user can allocate strings for himself. Is there an interface with commands on the command line? Can you help me? No command line, no. It sticks to the message. But everything should be on their roadmap. Everything is registered in the text file. So it's possible to... Can you change the text file without preload or restart the portal server? I don't know. Last time we mess with that, there are some problems that... Freedom was not aware when we changed the POD and we already have the stats generated. It could confuse you. You have to do that in a lot of files. Last time the problem was that it was running in parallel, like it was updating the characteristics of the portal and at the same time we could change some files to it. It was already running at the same time. Yeah. It should work if you update the PO file in the middle of no other processes. If the user is logged in, it's a trial. Maybe it's not safe to do that stuff. I want to see a relic. What was the language? R-M. But this file is interesting. There are some fuzzy and translated strings. You can go there. A-M also. Yeah, you can go there. And probably as we are still logged in as myself. Just try to show the functions. You always have to reactivate this. Each time you go into it. That's not good. So you can see the interesting things. Yeah. And you can also state it that I can't allocate because I do not have all rights and all languages. That's not true. And here we can see that it's kept with two or three strings because they are translated. It's a special character. Yes. If I remember correctly, this is probably not active correctly for Amaric. But these characters are supposed to be related to the language. So depending on the language, you will not see the same characters. It's a bit what you have in Wikipedia. But still need that. So you have all... And this translator doesn't have rights to translate. It can only suggest. So this will not be kept in the profile in the file system but only aside. The problem if I tick on it? No. You can tick. You won't break anything. Because it cannot be validated. Okay. And then if you list Am... you have Am.po that's pending. So the Am.po has not been modified because it is currently pending. Okay. We see your changes. Suggested by... As you see in the comments, you come back. At the top you have a comment. I can get one, select one downstream to the user. But only he can translate it. I'm not sure it's possible to say only him can translate this stream. But you can give a goal to a user. His goal is to translate this file or 50% of this file. And if somebody else translated the strings then his goal is finished. It's nice to know. My strings. I have no strings. Because no strings have been allocated to me. Yes. But they cannot allocate some strings to me. Okay. So I think you need to define a goal. But maybe it's better to switch to the French. Frankly speaking, we will... we will get lost between... Okay. I think that everybody can see by himself what are the possibilities. Probably people around should register and they can set up a playground for everybody. Okay. So I think you can go for it. But entering the details is not really the point of this box. And just ask Christian to give you a login and the right permission. Okay. You can come back here. People create login for themselves. So everybody can create a login. Actually if you go on the server to try to create a login then you get a mail. You validate the login. These are the checkbox. You see here. These people validate it. And the admin of course can remove users. Okay. I just found about people that couldn't remove users. So we can see a list. And if I go to a project... It's not too long. Okay. If I go... No. If I go to admin. Sorry. I've been defined already. So I cannot name the project. Which project I will admin? For instance, DI. And then for DI in German. And say... Okay. I can update the file from templates. So that means this project will be... The profile will be synchronized with the POT file. Don't do it. But it will take a long time. And here I have a sign. So nobody unlocks the user account. View. Constations. Default. These are the rules. Don't change. Yeah. And if we are... You can say that even a logged user can suggest that this is not really what we want to do. This is what is in Rosetta. Then this is not really a good idea to know. A logged user... Any logged user can view, suggest, review and compile PO files to grab them. This is more or less what we want. And I can add user in the left. I pick up all users that would... I pick up Rhonda. Who registered for this language. So only these ones said, okay, German is my language. So... Yeah. That's also usually the issue. You have to control click to add writes. And for instance, if I want Rhonda to be the overall admin of DI in German. Superboss. I just give all writes and update writes. And then... The sign writes is necessary to the user. And administrate this project. So if I give administrate on that, we'll be able to do this work. This is a way to delegate writes. So they implemented already a kind of good infrastructure to administrate and delegate etc. This is much better achieved than other parts. You can administrate the languages by yourself and say, you see for instance, for Amaric, these special characters are here. Are the ones which we could see... You can set them up. The plural form can be defined. And all these settings are in text files actually. In Etsy Tutorial. These are the files that are modified if you have seen logs from Truro today. Yes. I modified some things in EtsyTutorial.print. This is mostly because of these things. And if we go back to the interesting thing and what we should do now, here is the place where everything is stored. So space is a little bit hard. So here are only director in Valid Putol. We have the Debtcon project. Vi project where we were with all this... Vi.org is no longer a project but I kept it there. Yeah. Ah, yeah, left on VAR. So when we will work on the DDTP, we will just add the DDTP directory here. And then maybe we can... I don't know what will be the layout, maybe inside the DDTP directory we will have all the languages or all the various package names. Or... That's the main problem. It's interesting to see, yeah, how is the data coming? Because that will determine how we can... We assume it has many magical scripts to generate the materials. But basically, last time, we decided to go with full structure. Right? A, B, C... Yeah. C, S, C, A. Okay. So if we... make a table of this and just copy it to Truro... That's what we did last time. And it crashed Puto. Yeah, let's do it. It crashed because it was too slow or it was... To token handle everything. It hits all the memory and then starts... It should be much better When Puto is starting, he will just create all the PO file and make a statistic file for every PO file. It's really, really slow because it's not only counting the number of strings and the number of untranslated strings. There are a lot of checks which are implemented in Puto like the spell checker or he will validate that there are the same number of parentheses and the original untranslated strings. So it takes a long, long time. One way to solve that is just to make a small C program which will read the PO file and write the statistics without the extra checks. It's really, really fast. Puto will not notice that the statistic file are faking. The statistic file will be updated when one user makes a change in a PO file but it will be slow it will be smoothed much more than doing that for the first time Puto is in. Then, on startup, Puto will try to read every statistic file. It's also a problem when you have 50,000 files to read it takes some time but it's much faster than reading the PO file and making some computations. So maybe we can make something much faster that way. The second time you run Puto it's much, much faster. In the web interface when some data are in the Puto memory afterwards it's much, much faster too. So it maybe reminded that when we restart Puto we have a script which kind of navigates in the various projects just to load the statistic file and make sure a user won't have to wait for 5 minutes to open one project. Point is out at 11 so we have time. But just to give you an idea Puto on my box it took more than an hour to load the DI directory and now it's only 5 minutes. So there were some improvements in the extra material meeting. Did you know all the file stuff from Puto to make own translation user files? I'd like to write some main interface or something like this for Puto but then I must know the files or to change it how to select some description to some user that it's not nice if one downloads one description per email and then the same description is still online and some other make a translation. So you would like to log one time and say a lot of people translate the DDDP by the mail interface and the DDDS also uses it so it is an important part of the puzzle. Yeah, doing mail. Do we imagine a translation or a translation support? Yes, maybe one option could be to only activate the navigation on the DDDP in the Puto interface. And maybe use it for the usage or maybe activate Puto fully for some languages because you can remove writes. But you remove? Yes, maybe it's possible. You can try it. Can I go to DDDP? Yes, yes. French, just here? No, just on the I. There's one? French. Oh no, there's one. What is on the I? Can you list the... And remove French to 10. It's okay? It could be an option to remove the French one. That works quite well. It's a super user. It's DI. At that moment, we saw another problem. Because now the statistics don't match. Maybe this is also a mix of them. Usually it checks if the statistics file is older than the PO file. In that case, you know what you will do. Probably you will try to rebuild the statistics file. Okay, you have seen what it does. It's not. It's not. So French is a rude method, it's not a good method. You only see the level 2. Loveless. But I have seen that. I think remember that you have level 1 because it doesn't check every time. Okay. But that is easy. Because I can add a French PO point. Can you check the content of the DI directory? Like this? No, the proof is you're the parent. The parent. So what? Can you... DI-FR? No. Is there something like DI-FR? Put all DI. Put all DI. FR, but press. Somebody seems to be editing this file. There is the main difficulty with... Can you show the content of the DI-FR? Put all sets default writes to every project when a new project is added. This is also... You can change the writes. You can say this file is currently under review by the main interface. So you can probably comment... It's a bit unscripted. You can just comment the non nobody. It's a nice part. Okay. This is for for the root DI. Yes? This is the place to define writes. And actually, as I was finishing, put all default files and you often have to edit manually to go faster than the web interface. So you can also imagine having offline tools or other tools and if you go back, you will set the other one again. Probably you will see the other one again, but you need to refresh. Or refresh. So it will be easy for people to see the files again. It's missing the stats. These will move only the PO files. It's probably other files. It may be called by one of my patch. It's one of the patch, the one that doesn't always recurse. You're reaching the middle. Yeah, we're reaching the middle. The main point is what do we want to do with PUTO? Not what does PUTO that. Everybody can go and see it. But what do we want to do? What else we have to say today? It has to be restarted, but we have to define what we want to do. What we have to say is up to us. My main idea is we should push importing the DDTP. That's the main point to do. So we have to talk about this maybe in the DDTP session or whatever. What should be put in the meantime until the DDTP session is to try importing some stuff. For instance, just just like last time we needed per language. It was the best way to test it instead of doing everything. So we can do that again. But I think first we should install the PUTO. Sure. There will be improvements, but don't expect much improvements, I think. Because most of the improvements are already in what we use here. How is PUTO upstream moving? They are fast moving the code or they are like eerily moving the code? There are still some functionalities. I think the performance issue was not the main issue for them. But they appreciated the patches that I sent and some of them are included in the trial now. But to make a patch, the performance was big, right? Yes. This patch is already in the new package? Yes. So we should first install the new package and then we can pair language on language base add all the stuff only to test the performance for the restart. Second question you can patch PUTO, I think. I can't, I cannot do language in all the stuff. Is it a big problem to check then I can write maybe an interface can send the PO file to a translator and make a doc file in the directory. Can you check it? The PO file. The French PO doc file and it is in the directory then you are only in view or suggest you know nothing more. I think it's possible. If the translator send it back then I can exchange the PO file and remove doc file. It's one way. The fast way. I just don't know how it handles the chains. PUTO has some memory translation, right? Or not yet implemented? I think every time there is something it will call a function which check if the user is allowed to do to do it and I can add something this function should work. So the main point would be to get an account for this on the server. So that can change and also test the automation of the push to PUTO. If we import and you created this chess table you could import it now. It's not useful. Now this is all PUTO. Why is it still creating this? It's a lot of files. A lot of files, okay. 20,000 or so. Okay. Yes. Quite a performance, actually. We killed PUTO. There are probably some improvements that we have been discussing yesterday with more memory, more artist. On the hardware. We also try to change a couple of treats on the file system like Maus would do that to help people faster file system to save themselves some time-stamping. And we can disable some stuff in PUTO like some checks we did not at least receive. If it flies then we can start hacking. But I'd like to start the main interface. If we have a fast machine a faster PUTO and all that stuff then we can translate all the files to PUTO and activate the main interface and then the translator can continue work without stop of the translation. And if I don't change the interface from the main server then also the web interface that we have will work and we can change all the translators to the main interface. So the other goal we had was also to explain what we also do. Not true. I come back to our schedule after the scan. So the point is PUTO is not only PUTO. No. Far from it. Yes. At this moment the most useful part is not PUTO. It's all the material. It's the need to stop the generating the files. IE-TNN the servers start generating it. We have like probably 8 or 9 robots running. Each one of them does a different thing. Basically we generate the entire material or tar ball with all the PO that is used by the the WWW or international infrastructure. We have also some local accounts that we plan to organize now. Nicola has graph status kits and some informational in his home account. Eddie is generating companion files. Toma has the robots for the main list the old 10 robots that keeps tracking off. I think project SMIC is also using the robot right to track reviewing about tracking reports and all that. Yes. This is Toma's robot. Yes. It's tracking robot which tracks the main list. Which is when you register for most of the translations not all of them is using this robot some things have we have in fact four robots in the wild three of them have different patch one of them is entirely rewritten in Python one of them feeds a database one of them receives directly the message and our bot cross the main list so most of the things are using our bot but we still have other things to have their features on our bot and we didn't work with that Puto is running on that server we have a full mirror with all the three suites and all of that is on that so it makes available to generate statics from any timeline any testing that we didn't have passed so now we generate for both we still have accounts we basically it's a very easy server to have an account you just need to contact me with GPG message and I will create the account for you we are just moving for some security new security resources on that server so we are allowing only SSH key login and we will change a couple of things like having Apache proxy for plural HPVPS probably Cesar will arrange a couple of another IP address you activate the MX structure on that server so we can grab mails we can have email infrastructure in lists and other things better with that, I figure out how to do that and there is only one wiki we already spoken about it on the first meeting that we are closing it moving all the information we have to the main that in wiki and this wiki we remain only like this is the url for this is url for this so this is basically what we have most of the robots are up to date with the CES in fact we don't change them in a long time which basically gets the entire archive of the list process it without the pseudo URLs and register what is what that's all the information we have over there and the the material that generates this page then our web server grabs it from the material this is one of the important part we are here on the main web server official dvm.org architecture but the material that's pointed here and this is used daily by open status for to the conference is pointed to yes you should point to 5 if I point this vo5 it's there it's there it's there it's there it's there it's there it's there it's there it's in there in fact true is already part of the dvm infrastructure and if we break true if the robots break we are breaking these pages We have several things in production at this moment, and probably only in the hands of a few people, that's the main weakness we have. The knowledge is not very widespread, but there are several things converging. We probably need to organize these things better, for instance it could be good. To point here we have official LTM pages, but if you want to reach the robots, it's pretty hard to find. If you don't know the URL, you have no chance to find it. So this is something that has to be improved, probably. An ID would also be replaced with DI status page, which is semi-official, and to point directly in Putol, because Putol is able to do all this, maybe not as nicely, but quite well. There are a lot of small things running already, I think. Do you want to show the robots from Thomas? The output from Thomas? Yes. I don't know if everybody is familiar with it. I have to find the URL. If you go to the main page of True, it's the same thing. www.hypingand.dev.net slash drdm. Don't change this one. This one. This one? Yes. What is delayed? Expensive. This is easy to understand. No, no, it's DL10L10, DL1010. DL7? Yes. Here? Yes. No, DBM-L10N. DBM-L10N. DBM-L10N. DBM-L10N. DBM-L10N. DBM-L10N. Yeah, this is mostly the crawling robot results. So if we're trying to go for... For the restance for Brazilian? Yes, still. We need to... We have plans to split it four lanes. You can see PT by package. And this is generated by two robots? Yes. And the crawling robot? This is already the stats of the base. And it also automatically closes the bugs per day. The bugs that were done. Here we can see the information for package, file, translator, status, and less sending message. So it's also interesting for the translator. When the page is sorted by a translator, one translator can see if he did not just forget one translation somewhere and did not finish the review or anything. The process is very hard for me. It's very, very difficult. Like this? Yes. Yes. Still a coding problem with Android. Yeah. Yeah, but I think it's because he used different mail clients which do not send the same names. This from... I translated in Portuguese. Yeah. Really? It's a BTS. It's a BTS. This is probably because I reported a bug. This package in ILO. Yeah. And the special treatment. So that's exciting. AD components are... AD is an ADP holder. Okay. So I18M.EM.NET to the ADP. Without the devianto. Yes. To the ADP. No. To EDD, Y, P. No. Y and P. P. That's it. That's it. Compendium. Peel. Docs are the information. Peel are the results. And if anybody knows how to use companions over the I, I would really appreciate it. The log file may be interesting to see if there are some errors in some PO file. I think if you go to the French one, yes, for example... There are two broken files. I see them also in the puttering port logs. It's also... The component is also interesting because you can see if one string is translated in different way, in different PO file, you will see it here. You can try to harmonize the translation between different PO files. So what's yours? So in the component, the error is not really interesting because it's a concatenation of the various errors. But if you pass the error and then if you search for sharp-sharp-sharp, you may find some interesting... Sharp? That's it. Okay. All the errors. But if you can go through the errors... Yeah, in the middle of file. Just click. Go down there, click somewhere. Here somewhere and now fine-est. Now fine-est. So what's for this? Next one. This one is not interesting but it's a difference between... Okay, it's a difference in the headers. It's a difference in the type of... Of the comments. So let's try the one. Next one? Oh, that's working. Sorry. It's really... It's really huge. But it may be interesting for... QAM translation. It's interesting for QA, I think. It's also interesting to give the component to translate to the new translator. The main... They use it in their translation too. Yeah, KTranslator has support for you. KVL, sorry. Yeah, but I never... I suck at configuring that thing and I never succeeded in configuring KVL to use the component. So AD is able to... I tried to find some information to use it with GI but I couldn't. Yeah, but you will find all this kind of stuff. So that's not really useful information. Okay. Because you are polluted by all GI stuff. This is something to improve in Thomas' scripts. They shouldn't pick the GI UDAB. Because the UDAB are listed twice in GI and in PODAB. So if Thomas listens to this someday... Maybe we can't forget text to avoid the... This stuff, yes. This kind of stuff. So this is the first... AD considers this as the experimental stuff. Later it should be run by the robots and not in GI and AD... We should start thinking of building everything that works to a common place not among the records. But multiple things are in common places. No, mine is not. No, yours is not. So yours should move. I was just testing if it's working. Yes, but it is. I've been checking it for a while. We did not explain what is the main use of the compendium. It's to start the translation of a PO file. You can... Message merge through it. And the PO file will be filled with the translation whether it exists in the compendium or if the screen is closed. That's the use of a compendium. I expect translators to know about what the induction is. The induction name is pretty... Our main goal is to organize what we are doing on foot or not exactly do a course about translation. Eventually we don't know when probably to some time it will become a software machine. I think one of the things to do will be to move all robots to common place. There is one advantage with the .NET machine is that it's, I think, much easier to get some login on it. Now developers, of course. As long as I know you guys, it's easier. Maybe it's easier also for to make some development, to test them. But please, perhaps, I don't know, perhaps we can have machine machines. We can still have iating.dev.NET for forever, if as long as X-Trim doesn't exist. And then we can have a new iating.dev.NET machine. What I will... has that profile.NET machine still exists. And they gave us 3D, which is a new machine. And do you think that such things that are quite moving qualify for that old machine? I think so. Because then they go to a super infrastructure of security backups, much more than we have. They are quite conservative. I do not expect X-Trim to disappear, but still, it's about our machine. But I don't see it happening, like, before landing. Well, we still have... What could be a target is moving old things. I want to finish. The things that are pointed by the Debian website, should the Debian.org website currently are in the Debian.NET, iating.dev.NET website, that's probably not a very good idea. Because we make some official Debian resource depend on unofficial ones. So this is something we should try to move quite soon to iating.dev.org. It would be nice also to see the material country. Country, the important is the materials. I think that we can do that, even on double, double, double on master. Well, at least we can... I don't like to move it for a people account. I don't think we should do that. We should move it for... It was in a people account, so we don't have to move it back to a people account. But, well, I can try because one of the things that we could do is use LDN. It's about our machine. But in a project, do you use the home directory? Yes, on the LDN. Yes. LDN. Or I can tell... I also say LDN. So let's go for LDN. I was thinking in Portuguese. Yeah, I was in French. But I can also speak with Joey about having I89 kept up in the tour. It's usually the guy behind new machines. We also have Wagner, which is able to give us VM machines. So, it's also... Well, I can open I can open, I think, the only request to record and see what's necessary. The main problem is we have a quite low quality guarantee I think through being broken or not broken, because this is something with... The anonymity is very small for truth. They are almost the same size of the essay. Yeah, but I consider that the only real anonymity in the team is a little bit. I have... I have super rights, but basically I can use them for breaking anyone. So, it's certainly not a good idea and I wouldn't require it on a daughter machine. It would not be a good idea to give them to me. I think we are somewhat weak because if something is broken in some place we depend on Felipe on Cesar to fix the stuff. We are not critical but quite important to have information that depends on the system. Do you need only that space? Or... I think we can move. I would try probably we can have a cron job on the master on the website doing this. Probably. Yes. Yes, I think the material can be gathered together with the build of the site. I think we can do this. No problem. This is the past in the book. Now, I have already looked deep in org. I don't use it because I cannot conduct clients of the main system and some stuff. But you have an own group for this. You can put more than one user for the school. You have an own directory for your maids you can put scripts on it. It was not dependent on one user on one own user. No, no. Org. Org. Your request to the SCA. Six years ago I requested Joey. Joey has Joey. Joey shows it. I would try that. I would figure out the best way. That would be interesting to figure out the best way to organize stuff and make them more official. And of course, generate some documentation. The same thing. That would enforce documentation. Explain how this is done because we have seen when Denis resigned that we were heavily dependent and it would be a very bad idea to reproduce this and unsure. Maybe it's only a problem if you need a touch of memory or to use power. I don't know the scripts. The scripting is a little bit heavy on this because it needs to come back files. Copy and keep doing that a lot. But not memory intensive of this. There are no other operations on the PL5. It basically opens the package and grabs the PL5. What do you say? How long does such a run usually take? Last time I checked it and it looks like it depends on the new package which appeared. If it comes in like 5 or 10 minutes. It just feels very good. It's incremental. But the first time you run it it's only one hour. I think it's a little bit more. It's two or four hours. I would say that for the first run it would take like two or three hours. But I think we can create the first run on a different host and arsenal. We can import the selection. It's already run. So we don't need the first run on the machine. You should say how much memory, how much CPUs how much hardware space you need. If you think there will be one hour, one minute or something like this. Then they can choose the machine. If you need a local mirror, some machines don't have a mirror. Yes. It should point in your mind. Good. So this is mostly what we are doing on Shuro. A lot of stuff already. More than initially we wanted to set a put-off server finally. And currently the useful part is not the put-off server. It's all the other stuff. What's your page? It's on Necral. Necal. No. Ne. This is a woman's name. Necal NP. No. NP? Yes. This one. Is this all? Yes. Choose a reverse. Choose a Wow. So it's a statistic for every language and different kind of statistic for any P.O. or P.O. for any P.O. Maybe the more interesting are the P.O. Bitcoin fires. You can see how it works. Well, the French one is not really interesting. Maybe you can select your reward. Instead of months, you can put your... You can see the Chesh or Galician. These are the ones that are faster. Yes. Yes also. Moving fast. Just have a look at the URL. The URL page. So we can see it's moving a little bit. If you replace P.O. Bitcoin by main.html main.html main.html main.html main.html Yes. Here you can see the bubble enemy campaign. It's probably the release. No. I think that's the release. No. If you select the testing distribution, you can see that the number of things increase after the release. It's probably here. Oh yes. Yes, but that makes easy to stop the activities and make a kind of funny contest. Sometimes I'm using also unstable and unstable DTS and I can see if a lot of translations are lagging in the DTS. So it's time to start the enemy campaign. If you select something like the monthly report you can see when Clyty came back. No, it's still main. Yeah, okay. Month or monthly? Months. You go down for weekly news? Yeah. On week 23. She's back. She's back. And it's also interesting to if you have the number of translated stream, the number of un-translated stream in the existing PO file and the number of un-translated stream because there are no PO file. So language can focus on removing all the the rules and the reds and then reduce yellows can be a good target for team or not. Depends on the you can do some contests. Yeah. It's not very nice. It's because you want to see another language after I'm done. The interesting thing is this nice language that works in the last year and this is a 5000. So make like 4000-5000 translation. You're pretty lazy for French. I think it's no problem. Also, John, if you see this and this Yeah. The ratio is lower but I think that obviously everybody is waiting for something to happen. Well in fact, somebody who lost motivation when you have all the come and go, come and go, come and go. If you see the come and go there's one coming. Yes. People still is not like 100% sure that this time we go. Yes. Why do you decrease that much? Because it's a time that everything stops. This is also back in and then that and there was a main problem and I really liked this stuff and this was the last time this was one year and I re-voted and all the input includes import used translation. Which is the keep growing green line. The packages and the count of not translation the count of descriptions that are in set are used in the database and this one is the count of description in the database without the outdated with the outdated this is without the outdated in set this is translation. Yes, we see in the database I think we can mostly close the session.