 Thank you for coming to my talk. My name is Noriko Mizumoto. Today I'd like to introduce how the localization team is doing a fantastic job. Before we start talking about the localization, I'd like to introduce a little bit about the art of technology of Japan. Maybe some of the people know about the Toto's company in Japan. This is a little bit the embarrassing experience from five, six years ago by myself. It is not the first time, but the total technology is so adverse and every year is changing. That time, the first time to see this one and so this device. Then I got in the room and then do the things what I need to do and try to exit from the room. But before I exit the room, I have to press buttons or I have to use the lever and try to find the buttons, levers, everywhere, but I can't find it out. I will show you the entire device, the state of the technology from now, embarrassing you. This is my call. So I did everything down. I need to fresh it, but I can't find the lever or buttons to fresh it so that I can get it out from this room, right? And then at that time, this is the one. It says already English, but when I was in there, there is no English, only the Japanese. So if I am the English speakers, totally I have no idea which buttons to press. And then finally I got found this device. It says, hover my hand, it will flash. It will first explain for myself because flashing means flash the buttons will deliver, but not hover my hand. It just hover my hand, it's fresh. It's surprising, embarrassing technology I encountered. Now these days, total develop this one. Now no lever, no hover the hand, it's enter in the room, the lid is automatically open. You did everything you have to do and then walk away. Flushing down, flash, lid to back, no hover, no hover hand, just walk away. It looks after everything. So this is totally different from the localization. What I wanted to tell you is this is the one. This is the one and that one, if you guys only speak English and there is no English in there, there is useless for the people who doesn't understand the Japanese for this technology. I wanted to tell you how the localization is important for the fedora itself. English fedora is sometimes useless for non-English speakers. This looks like an English written fedora page. And this is how we translate it. It is translated in Polish, translated in Czech, translated in Albanian, translated in German, translated in Japanese, more and more. And in French and other languages so that every non-English speaker can hopefully understand what is going on, what is the technology that the fedora is providing to the users. To do the localization, every application needs internationalization first, otherwise localization cannot be done. It's kind of the preparation to down the proper localization. But I am from the internationalization team, so I just give you small tips about what is internationalization, what is important things to remember. It's both sides of the localizer, also the developers. First, when developers create their own application, I like to emphasize that please think about to make your application the internationalization tool to keep compatible. It's like a get-tekist or logging tool, so depending on what languages you are using, it is probably easier for you guys to surf the internet which one is for you. But make sure that the internationalization tool keep compatible so that makes much easier later on to the application internationalized and then localized. The second one is write in English. So if a Japanese developer tends to write in Japanese partially or entirely, then in open source, more other developers can join, contribute, improve, or send a patch to improve the year coding itself. It is good to have everything in English so that other developers can understand your application itself. And from the third is more problematic for the localizer translators at plural forms. Without the plural form, sometimes it's really hard for the translators to translate. For Asian language, we do not have a plural form. So singular or plural, we do exactly the same translation. But for the French or the German, the other email team has a plural form, so they have to have separate strings to translate singular format and the plural format, even with the sentence exactly the same. And the numbering substitutions also the problem for the translation team. If you use exactly the same substitutions, then we cannot change the orders, but many language has their own grammar different from the English. So sometimes we need to fit the substitution, but if substitution doesn't have any numbering, we can't do that because it's being fed in the order which comes from the first, in English format. So the numbering does help a lot, which gives you another example, like one, two, three, four, five, so it's in the English order, but if it translated into Japanese, we have to change the order. Number two comes first, three, then one, and two, and five, otherwise it doesn't make sense. What does it mean? Output is second of August, Tuesday, 10 o'clock. In English it should say Tuesday, second of August, 10 o'clock, but in Japan it is a different order. That's why we really like to have numbering for the substitutions. And the last one is do not cut it off the words in the sentence. It says that this rule is applicable to black, the substitutions. The one sentence, we are easy to translate, but if cut it up in each terms, then it's really hard to give a correct translation, and then we try to give a correct translation, but it's going to be the literal. So that once that is consistent in the one sentence in the program as an output, it looks like sometimes funny translation, so it impacts the quality of the application itself. And then that means the user face quality is sacrificed because of this. So make sure that how it looks like before cutting the sentence. There are more probably the questions about internationalization, but as I said, I'm not coming from the internationalization, and it's lucky that we have a couple of internationalization teams on board on this clock, and one person called Praveen is going to give a session about the globalization audit of Fedora Atomik. Please join his talk, and then any internationalization problems or the questions, I believe that he can answer more in details. Okay, let's talk about the localization. What we are localizing. We localize websites. We localize, of course, Fedora itself. This is an installer, and the package name is Santa Conda, and this is Santa Conda. We do a lot of the packages in Fedora as well. This is a login screen. This is actually the GDN, the combination of the GDN and then GNOME shell packages. Once login, we'll receive this screen, a setup screen. This is coming from GNOME initial setup packages. That means many of the Fedora translators also contribute to the GNOME project itself, and then some of us even contribute to the LiveLelfits as well, or the Mozilla project. We work multiple projects to form the nice, complete form of Fedora. We work on the Wiki page. It's more free format that each team may have their own language team pages that we do translate the Wiki page as well. And then last one is the documentation, the installation guide, system administration guide. Many of the guide are published from the documentation team who will localize this guide as well. But at this moment, there is some publishing issues on the documentation team side. So we do translate, but unfortunately we cannot publish or localize the documentation at this moment. And hopefully the problem will be solved soon. Is there an item that you can translate? Yes, we can translate, but cannot publish. Why is there the name Fedora Translate that's strange to me? Talking about the translation quality? It's weird. Yeah. Normally we don't translate at brands. So Fedora Workstation has to be Fedora Workstation every time it is. And normally in Polish, Fedora should be Fedora. Yeah. Many of the languages, Fedora is Fedora, we don't translate other languages. The question that I can answer, because it's my English, was your question? Many of the questions, why did you name Fedora Translate? I will not translate your name. Your name is your name. We use some pictures from Polish language. Because otherwise it looks like you're using, it's not using the plural in English. It's an S, a D, a plural. Someone may ask why you have computers. That's why you have them to read the whole sentence so that you don't have to use the plural. The plural has no plural at all. The plural is the plural. So you want the plural. But language, language rules are strong. Yeah. Because in the way that it sounds natural, it works without changing the way. Yeah. We use foreign language for everywhere. And so there is no exception. It's a complex language. And do you translate Fedora Workstation? Maybe it's so long, Fedora Workstation. No, we don't. Maybe we can discuss about the later in the workshop that we're going to have. So let me move the next slide, please. I'd like to introduce where the translation activity sits in the development cycle. This is the Fedora schedule. The only shows up the key milestones. And there is a software string phrase date. This date is exactly the developer has to stop, edit, or add or delete the strings. But they have to finalize the string part of the application. So that the translator can start translation and make it 100% translated. If the developer break this date, then we give the 100% translation. But still the application will not be localized 100%. So this is the date. It's very important, the possible from the developer to the translators. And the next comes to the deadline of the software that we stop the translation and give the pull back to the developer. So the developer will start pull the latest translation, build, make sure the build is the same, and put it back to the Koji. So that will come up as a release, next release. Let me scroll down and we'll see the details of the schedule. In here, for the translation schedule, we will explain more about, we do have a QA test period as well. So more detailed schedule we can find in terms of the translation team. Also, as I explained, we do translate the documentation. We do translate the website. It is depending on the website team or the documentation team. So usually we go in the documentation team schedule. When is the string freeze date for the release note? Or when is the string freeze date for the website itself? And the last one is upstream project schedule. As I explained, some of the QAnon part will come in Fedora. Some of the Libre office part come in Fedora. Some Mozilla part come in Fedora. But which part will come out? It says that exactly the which version of the QAnon will come in Fedora 25. Which version of the Libre office will come in Fedora 25? So according to those information that the translator will work or focus on those versions to get in Fedora 25. Language change. We do have 81 language change at this moment. And 875 translators registered. I think we are one of the biggest sub-projects under the Fedora project. I don't say that all of the 875 translators are active at the moment but at least they are registered. And then they are potentially come up and being active in and out, in and out. We have our own homepage. It's Fedora project slash Wiki slash Elton N. All the information in there. We do have a bi-weekly meeting on RSC as well. Yes. Registered translators is a people in Xanatar teams on the platform Fedora.xanatar.org. This is how you count it. 875 counted from the trans-at-list mailing list registration. Okay, so the number of persons registered are trans. Yes, that's right. So there might be more who are not registered in this mailing list but only in the local mailing list. But basically we asked the new translators to join this main mailing list as well as the local mailing list. So when we are writing, we are writing to many limited persons. Yes. Okay, how to join the language team. If new and interested in localization, this is how to join the language team. We do have a more kind user friendly guide written in the Wiki page which can be found in the homepage. Basically what the new person has to do is create FAS that's every country to do. And subscribe mailing list that we just talked about. Once trans at the list of Fedora project, all that this is mailing list, every translator required to join. The second one is language-specific mailing list for a Japanese trans-JA mailing list and French has their own mailing list. In this mailing list can be discussed about the language-specific program or the goals or whatever the problems and they can speak in their own languages as well. And sign to the Zenata because most of the packages are registered in the Zenata and we are using the Zenata as our translation platform. And then we can start translating. In the previous Wiki page guide consists of six sections. But to start with the translation, we only need one to three sections. Four, five, six is more like additional information how to proofreading, how to make sure no error for the packages. But basically we need only one to three sections, that's all. And in section 2.3 there is a section called Introduce Yourself. This is most important part of the join team. It's not the technical part, but this is a communication part communicating with the other team, with the other translators. Say hello to others to start the team working is very important. If you are here and then other team work over there and then there is no communication then it's funny, isn't it? It's like working on the same language team then it is nice to help each other. Say hello where I am the new, can you help me blah? Then the other person can help you but say nothing and then the other person can't help you at all. So the communication is important and this is the very first part to introduce yourself into the team. I just brought up some neighbors language teams. There is a coalition team led by Pioto. There is a Czech team, there is a Hungarian team, there is a Slovak team, there are the German team as well. If anything, any questions, any translation, probably talk to the coordinator, the coordinator are happy to fix it or discuss it. In the initial keynote we learned a little bit about the Fedora Hub and then translation team and the design team are looking to connect the localization team, Zanata workflow, translation work process connected to the Fedora Hub so that we are more easily to access Zanata, easily to learn more information from the hub in one place. I can't have anything to show you today but there is an infrastructure workshop on 4th August Thursday starting 1.30 to 3.30. We are free to talk about third equations, give us suggestions as a translator what we want to have in the Fedora Hub. So please encourage yourself to join this workshop to talk about how to include the Fedora Hub in terms of the localization. Let me introduce a little bit about Zanata, how we work with Zanata. Put in the web browser fedora.org then this login screen comes up. It's a very simple work, click on the right-hand top log in then the second login screen comes up. This actually comes from the H1 OpenID given by Patrick. I can see him in this room. This is his application. This application connects Zanata to the FAS system so that we make sure that only the translator who has FAS account can access to the Zanata.org application. Put the username and password from the FAS account and then we can go in the dashboard of the own dashboard. From there, say if I wanted to continue the translation previous work, then go to the history activities and click the link. Say I used to work the Anaconda yesterday, click this link and then I can move to the Anaconda project page to start the translation or start the new project in the search field on top, put in the project name to search and then enter then it goes to that project page. It looks like this. This is the Anaconda project page. Some project page has multiple versions. The other project has only one version, most likely master version. If there is only one version, then the translator simply works on the master version and then that translation will go into the latest release. But if the project has multiple versions, then what we have to do is currently we are in the Fedora 25 release development cycle, find out the F25 branch in here. If there is F25 branch, then work on it. That will go into F25 release. If we pick master while there is F25, master is for the future. So it's not wasting, it's going to the future, not the current release. So as long as there is F25 or the current release branch working on that branch so that it can go to the current release, at this stage there is no F25 branch. That means Anaconda team doesn't create F25 branch yet. So we still be able to work on the master branch. Anaconda team will copy the master branch, create a new F25 branch and upload it. From that point, the translator can start F25 branch for specifically the F25 release. Click the... Because there is no F25 branch, click the master branch this time. It shows up like this. Click on language, say this time click the polish and then right-hand side, number three, it shows a document. Most likely the project has only one document, but sometimes one project has two documents, and that means you have to work on the two documents to complete that application, especially that application. This time only one documentation, so click the button manager. This is called the editor screen, where we can add the translation, edit the translation, delete the wrong translation. The left-hand side is a source, English, so we can see it, but we can't touch it. So even click it, we can touch it, we can edit it, we can edit it, we can delete it. So it is safe. Sometimes we say we... Localize in the editors, then easily to accidentally touch the source English, and we change it. That can lead to the error when they... when developers package that translation so that they give up to include the translation. But using the Xanada, we kind of touch the English source, so it is safe, safe way, not accidentally touch the English code, but we can translate it in the right-hand side, there they add the translation. Can you choose a language source? The language source is the left-hand side in English. No, but I mean, so for example, if you want to use Russian on one side and say Japanese on the other, because we have people speak Russian in terms of English. Yes, you can. Yeah. So on the menu on the right-hand side, there's very little on the right. Yep. Technically, it can be done, but if the source is Russian, and then the developer requires the Spanish translation, imagine how many translators can transfer from Russian to Spanish. That means some men do understand Russian and the Spanish at the same time. It's a very small amount of resource. So in the beginning, I explained that right in English. So that many of the translators do understand English, so that can be translated in many other languages. So if even the developers speak Russian, then I encourage that developer to grab some English-speaking, well-English speakers to translate his Russian into the English first and put the translation as English source so that the many other languages can translate. But put as Russian as a source, it's really hard for us to translate because I don't understand Russian, so I can translate from English to Japanese, but I can't translate Russian to Japanese. I just explained a very brief, essential information about Zanata, and lucky again, we have a Zanata-dependent team on board on the flock, and he's going to present a more detailed Zanata session, which happens today, 2.30 to 3.20. So please join his session with a bunch of questions and shoot him until he died. Okay, that's all from me. My name is Noriko. Any questions or any suggestions just come to me after the sessions, or you can send me an email, norikoadridhead.com, or I'll be on IRC. Usually I'm on the FIDOR-DUSTY-11N channels in the Austrian time, 9 to 5. So just send me my nickname, it's noriko. So any questions, I'm not sure I can answer everything, but yes. Can you tell us the history of localization, of localization, of localization? History of localization. I have a lot of language testing work over the phone. Right, localization history and the testing group. Localization team actually started almost the same time in where the Fedora started in 2003. Before the Fedora started, the Red Hat has Red Hat Linux, not Red Hat in that place, Linux. Red Hat selling the Red Hat Linux, or the open source contributed. Red Hat decided to have template the Red Hat Enterprise Linux and Fedora, so that Red Hat will fork the Fedora as a Red Hat Enterprise Linux as the other doing the same thing. Since then, most of the Red Hat developers contribute to the Fedora, so as the Red Hat translator contribute to the Fedora localization itself as a beginning. And then we started to more encourage the community people to join and take the lead. Many of the language are now led by the Anong Red Hat, but their period community translators. Even the Japanese team has their own lead, not me, but their own lead. So we are getting bigger and bigger. In terms of the testing group, this is a little bit programmatic area that we need more resource. Preparation for the testing is a very important part to keep good quality for the localization because we work in the Zanata, but we don't know how the string comes up in which situations, but through the testing, we actually see the application as the operating system, so we can see how the string comes up and check it, and then the quality is good or bad, and then we can change in the Zanata. But the testing is, we need some resource to communicate with the release engineering team, also the QE team to set up the test days and prepare the testing itself. And at this stage there is only one or two people from Red Hat, but differently we showed up the resource and we like to have more people join it on both for the testing group. This interaction do you have with other projects? Sorry. I guess that's a lot of the terms that I've been translating would be very similar to many other complete applications. And you said many people also work on certain different projects. Yeah. Is there any way to share previously done translations? We do... I haven't tried to count how many packages we are translating. It's so many more than 100. Then in Zanata there is translation memory. So all the translations are kind of stored as a data and then we can refer to it. Then there is a copy button so we found the similar translation or the good quality translation in the translation memory. We can simply click the copy buttons. It comes up in a field in the editor. And probably a little bit tweak needed. So basically we can share the database within the language team. And then how is the quality done? You mentioned quality translation to people both or how is this managed? There is another function called the review. Once translated it says as the status is translated in Zanata and then another person called a reviewer has a reviewer permission in Zanata that person reviewed that strings and market reviewed approved or rejected. Approved means this is good quality and then the status changed to the reviewed approved or rejected so that somebody has to redo the translation but if marked as a reviewed approved that means the quality is good. So everything in Zanata is a pretty much smart user friendly translator friendly platform which is also used by the open stock project as well. So we like to spread to Gnome itself. Gnome is using at the moment downline pretty much in manuals. How do you do the translation team and tell it to the project to not to explain that are not reviewed? Yes, a small translation team say only a few active translator is at the moment focusing on 100% translation rather than the review. So those team haven't got the review at all but probably from now on we should encourage those teams to review more essential like Alacrondo or the essential packages to make sure that the quality is good. Does it exist in Zanata or the way I as a French coordinator I want to say this project is under the review process. If the content has not been reviewed or approved then we can't go live in user hands. Is there a way to tell the project manager this is our way to work for this translation so we can manage quality as it exists? There is such a feature but that feature is only handled by the developers. So developers can decide the translation can only be used as review the translation but if key use that feature that can affect all languages not only in French. It's a translated decision if you have the approval of the tool you cannot say it's mine I agree. It depends on the way the Polish community is organized. I agree on that on the workshop prepared so that I don't know I believe that Zanata team will join that workshop so that part I can't do anything at all but Zanata team can do something else for us I can share with you the information of this some power community actually I would like to tell you that I often contact with translators if I have some messages but it's not like that I cannot contact the translators I can and I do it. I saw them with translation I get my basic talents I introduce myself to the local language as well as the people who are working with me on my list I call it and then I go over to Zanata make my account and then I start in translation and two years later I will recognize I still translating here I'm not part of a particular project I have just the crew membership CLA and the other one but I'm not the translator officially I have just CLA I said nobody puts me in a group so I have no it's all languages and to look we have there right now and with Pumpskill he's already active in translation and what I do then is I see dot cross info and the username and I don't see the CVS 10LNs you know yes and so I have to look up in Zanata but in Zanata I don't see it as me active or not it's the next issue I just see it as an account no actually you can actually check individual statistics if you have accounts like this so if you go to Zanata is that it or I don't find it I have pretty all yes so it's open for public so anyone that has some account in Zanata you can actually log in and search for the user you can see his individual contribution throughout the year or whichever time period it is I have to see the address for them yes it's not a problem but how about there is some miscommunication between you and your community but the the law approved problem still stands so I think there is probably some miscommunication between you or I'm not sure I'm speaking yeah we better to look closer at your problem but in general so as you don't want to see this so and if people really just go to Zanata and they are officially they have some rights because they cannot go and vote for the answer or they cannot go vote for Presco because they have no other group so it's your problem and recently and there is a spam problem actually there is 4 words existing for localization it's the authorization rules of the past but they are not very new but recently because of spam problems we translator need to be in the rules we don't want to use the week for example and so this revealed that the past were not actively used anymore so there is a question actually to should we add every translator that kind of group or should we not and this is mostly a technical or process way of thinking the biggest problem is are the translator active members of the federal project or just a little answer some mistake and goodbye how do you make the translator part of the federal project more globally and it's something that looks give a problem to me and I need the past to defend it maybe we finish this sessions and get together closer so we can discuss here there is no markup or the automatic status that this person is active this person is active more likely in terms they regularly comes up in mailing list and then help the people or submitting a translation itself it really depending on the teams but for the Japanese if the people one translator one one packages then that person one post into the mailing list hey I completed this package translation somebody please review this translation and then other persons comes up and give a review so sort of the communication back and forth back and forth I called those persons are active why the other person just read in the mailing list do nothing is in active ok thank you very much everyone thank you