 Okay. I will come here. So basically we are having this three hour track for globalization. And if you are knowing globalization is nothing but it's a part of IITLN and LNN. So it becomes a Umbrella. GLNN is basically a Umbrella that consists of internationalization. In short we call it as a IITLN and localization in short we call it as a LNN. So globalization is nothing but it's IITLN and LNN. So let me explain the internationalization. So what internationalization is that internationalization is nothing but adding a specific port to your existing software that will help you to make sure that that software can be localized. So when that software can be localized that means that it helps to make sure that your software will work uniformly across multiple regions. So what this means is that suppose you have some application that is showing an English UL strings and you want to make sure that the same software will run in different languages. Then there comes an internationalization part that will make sure that whatever the strings are available for that software that will get extracted, that will be extracted into some form say like a port file or any other form and that will be available for the localization team for translation. And they will do the translation and submit it back to the upstream and upstream will include it into the next release and thus that software will get all such different translations for different languages. So you just need to run that same software in your language and thus that will make sure that it will work uniformly. So that is what I mean here. The software will work uniformly across multiple regions. So the next point like I said it includes adding an idl support in your port. So basically what idl and idl support is nothing but there are multiple idl libraries say mostly used is a get text library. So you can use this library to extract the UI messages from your software. So that is nothing but adding the idl and support in your port which allows your strings to be translated. Now internationalization also helps to set for that particular language or you can say a locale. That language setting like a date, number, currency, sorting and paper size. So these are the basic things that is needed while developing some locale for your language. Locale is basically a plain text file which defines everything that is needed for your languages. Suppose your language is having certain format of date, certain format of number, certain format of currency. So all this information is included into a JFC locale file and that locale file will get used whenever you switch your language from one language to another language. So that selected language will get these all settings effective. So you will get all the environments set for your language. So that is why this JFC locale is very important. Then comes a part like a discard input text. So internationalization also provides certain libraries like input method iBus is there. So it allows you to type in multiple languages. So input methods are there that will help you to type in your own language. So iBus is the main basic input engine currently in Fedora. So there are multiple input methods that you can add to the iBus engine and you just need to select and start typing. Now we have, I think, very good number of input methods available for major languages almost. So if anybody sees that their language is not having an input method, if needed they can find some already available input method and they can submit it to Fedora and that will get included into Fedora. Now they can write even and they can submit it to Fedora. So I think most of the languages are having an input method available but if anyone wants to contribute they can contribute it. So this is all about what internationalization basically is that. So I would like to add further that internationalization contains three main parts like input, display and fonts. So these three are very important things. So these three make sure that your system is running fine for your particular language. So I think we will see now what recently Fedora IDL has developed changes in Fedora. So let's start with Fedora 24. In Fedora 24, internationalization team has proposed two changes. Those are nothing but landpacks installation with RPF dependencies and GDPC token sub-packaging. Now landpacks is a very important thing in Fedora because it allows you to set your Fedora system with all the needed translation. But remember one thing that landpacks is available for very basic and certain number of packages only. It's not available for all the packages in Fedora. So some packages allows to separately create RPMs as a localized RPMs for that package like say if you have a LibreOffice. So there are multiple LibreOffice localized translation RPMs are available. The same thing is now for the Jellipsy. But in Jellipsy we are not having a translation but actually we are having different local files that is getting created in their own RPM files. So landpacks is this thing. So if you are aware that when we were using a YAM, there was a YAM landpacks plugin. So YAM landpacks plugin was taking care of all the things. But as we have moved to the DNF package manager, DNF package manager is providing certain restrictions to resolve the transaction. That's why the same YAM landpacks plugin cannot be ported as it is to the DNF landpacks plugin. But we tried our best to port it and we tried to manage to port certain commands in DNF also. But we cannot add this functionality that suppose a Fedora is getting installed in certain language. Then whenever that certain language the Fedora is installed, you should get all the landpacks automatically installed in that operating system. But with the DNF it is not possible because DNF is not allowing a real resolution in the transaction. So what YAM plugin was doing is that when installing Fedora operating system, it picks all the initial packages that are needed to install Fedora on your system. And then if you are planning to add a certain different language than English, then it will check for those settings. And accordingly those settings it will add the landpacks available for that language into the same and upon the initial transaction. So what happened that it re-resolved the transaction, it added the needed landpacks and a whole transaction as a one transaction was getting installed there when there was a YAM. But the same thing is not possible with the DNF now because it was not allowing to re-resolve the transaction. So only the initial set of packages required to install the Fedora was getting installed as of now. So this discussion went on for a few Fedora releases. DNF landpacks plugin was unable to do that. So there was one concept introduced by RPM that is called as a RPM tax. These tags are nothing but big dependencies and rich dependencies. So what these big dependencies are nothing but that you can add a certain line called as a supplements line into the spec file and you need to build it. So what this approach is that you need to have special meta packages for the number of languages you want to make available the landpacks. So we created almost 60 to 80 landpacks packages in Fedora currently. So what you need to do with this approach is that you need to install like DNF installed landpacks happen FR. Suppose you want to install a French language on the landpacks. So this will make sure that all the basic packages that are available on your system like suppose for LibreOffice it should be LibreOffice 4. Then if LibreOffice 4 package is already installed on your system it will pick that LibreOffice French landpack and it will install when you give a command like DNF install landpacks happen FR. So the same applies if there is any Huntspil packages available like Huntspil FR or SPLFR or Jiripsil landpacks Jiripsil local FR is also getting installed. So this actually is for the users that at least they have some way now that where they just say a command like DNF install landpacks happen their language code and all the needed landpacks transitions will get installed. Now with this thing we have solved this thing that re-resolve is not needed but with the supplements tab it is going to pull all the needed RPM packages when you say DNF landpacks hyper that language code installed. Now this is the way to install manually. Now suppose if you want to make sure that if you select a French language and you try to install it in French the Fedora then whenever you see whenever you install it and start it and you want to see the French translation for LibreOffice it is not available there. Because the thing is that landpacks packages are not available on bus station live ISO because they felt that if we try to add such a packages then it will increase the size of the live ISO. So that is why they decided that the landpacks packages should not be added. I actually given a suggestion sometime back that there should be a ISO, proper ISO, DVD ISO should be available from the bus station team but I am not sure they talked to someone. I think they are still with the thing that there should not be any DVD produced for one station team. There should be only a live ISO available for one station product. So here comes the problem that it is having a size limitation it cannot improve packages. So we are still trying this issue to be solved. I think if needed we can discuss it later on also. So this is what the change was there. Landpack installation with RPM with dependencies. But the thing is that if you try to install Fedora using a net install ISO it will work because net install ISO allows that thing because it picks the packages from everything required. That is why it is there because it is a net install so it can pull the language packages. But if you try to install from the live ISO it is already having a very limited set of packages. So there is no option but when you do up an installation you need to make sure that you apply the DNF install. Landpack say whichever code you have used for installation you can use that. That is it the landpack installation. So it is not yet complete so we are just exploring what solution we can get to make sure that whenever a user install Fedora you will get every landpack available for that language. A JLPC local sub-panaging. This was the change I think that is mainly for where needed to make sure that very minimum ISO can be produced. Say for I think RPM OS tree and ATOMET thing they needed that whole local archive file is not needed because they think that all the local files are having a fit of size over size. So they don't need that much. So they decided that local archive file can be split into multiple local files and they basically divided this single binary JLPC file into multiple archive packages. So now we have multiple local base RPM files from JLPC and whenever you select a language the particular JLPC landpack that whole FR suppose that RPM will get installed upon your system. So that was this change that we split this single local archive file into multiple different local files. That is what here is local sub-panaging. Fedora 25 I think changes. So in this range we propose basically three different changes. That is IBOS Symoji typing, Unicode 9 and IBOS Typen Buster multi-label support. Let's go through one by one. So IBOS Symoji typing this was first we introduced in Fedora 25. So this change basically propose that you can type emoji characters using IBOS in your favorite editor. So I think in this change basically IBOS package has been revealed with the Node.js emoji once data file actually. So Node.js is not an actual dependency for IBOS but only the IBOS uses data file which contains the definition of the emojis. And just pick and build IBOS and it allows you to type emoji characters using this IBOS. That's what I think emoji picker is. So yeah it was basically that it allows you to type using partnership E and emoji characters. So that was the change that this was proposed. Even if you needed to type a particular emoji code point if you are knowing that the very basic method we have is that partnership with you and the Unicode code point. And if you do that you can type your emoji character in your document file anywhere you want to type it. So this was the first that we introduced emoji typing in IBOS. Then at that time the next Unicode release or 9.0 version was released. So we just make sure that all the packages that are having the Unicode dependencies, Unicode standard following. We make sure that those packages are updated to follow the Unicode 9.0 standard. These are basically GNMC files and some form files are there. IBOS Typing Mooster 19.0 support. If you are aware that IBOS Typing Mooster is like I dance thing that allows you to type fast in your language. Now IBOS Typing Mooster was there. Now there was a need that IBOS Typing Mooster was not allowing you to type in support more than one language. Now how to type this in more than one language. So IBOS Typing Mooster, so there was some modification done to the IBOS Typing Mooster. Suppose if you are using certain languages like Hindi language, so you just go select a Hindi language. In the setup you just need to make sure that you enable the add direct input method. So add direct input method will allow you that whenever you select that Hindi input method, you will be able to type in Hindi as well as English in the same input method selected. So you will not need to make sure to change from Hindi to English. But with this setting, you are able to type in Hindi and English as well. If you want to know more about IBOS Typing Mooster, it's basically a suggestive text input. And this is actually the home page where you can get the information about IBOS Typing Mooster. This is meant to be in the world of my right hand value. I think it is working on it. The pipeline is to try and only have one single input method. At the moment we have this situation where we have to have different words with which language. Actually they are all the same thing basically for a language so we are talking about it. The main point about the feature that I was just describing was that it is impossible to have mixed languages in your dictionary basically. You can teach Typing Mooster different sentences as well. They don't want to be in different languages. Another thing that is new in Typing Mooster recently is that it is separate to the Dora feature but we also include the emoji in Typing Mooster. There is a lot of fun in the Dora for that. Even if Mooster turns off because it is a little small. So if you have a unique code input with lots of problems, the smarter Typing Mooster is. Dora 26 has made some big improvements to Chinese. I think the future now can support a multiple sentence of candidates. I think it is nearly my thing. It seems it is not so common. There are a lot of other short words. For the people around us I am trying to give them a little more time. But I took a few screenshots. This is the old version of Dora 24. I am telling you now. It means if I simply input. If I simply input. It means if I simply input. Yeah, the old sentence. This is my screenshot from Dora 26. It might be 2.1 or 2.1. There is the multiple. It increases the water level. It also makes improvements to the way to get involved. There are a few changes. There are also a few changes. One is the new Surrey font. The new Surrey font is 100% weak. Selection font is original and the Surrey font is fixed. Who has submitted this change? Pumbul. Pumbul. There are some changes to this opinion. So, I think Pumbul requires some of the back-end. There is a lot of new Julian. What is new Julian? I am going to go into the Julian language. But there is no need to. There is a lot of common code. There are some changes. These are not just formal and general changes. We are also going to do some work on the font pack. There are some patch bags. Another thing is emoji. The hand-manpacked is not really improved. It depends on the time I want the images to be small. So, I think we haven't really thought about it in terms of context. We are actually right about it. We have the Firefox. In the past, I tried to approach the Firefox in that time. It is a package language. It would say it was based on code. I think you are right. It does actually be great experience. Even with PFC, we mentioned before. Actually, we are starting to follow all the locales now. We are now in a big relation in terms of locales. On the other hand, I think you would think like LibreOffice would increase the size of the package significantly. I think for a ton of workstations, we have the recent kind of what to do about language packs and locales. The theory of how much size increase would be is package and locales. LibreOffice, I guess, is really the main thing. Maybe it is what it is going on. I guess modularity might help. As we see, lots of things happening right now in Fedora. For each release, we used to have, say, 20-plus features of change proposals. Normally, in any developer, developing any change proposal feature, his first objective is to get that feature done rather than looking into the language perspective already. When that feature landed in the Fedora, say, in particular release, maybe in the next release, we can take a look at that change proposal and see whether the admin aspects have been handled or not. So I think that can give us a good idea of what change proposal we should work. So, like, I don't know, like, VLN was the one, the flag pack is coming. There are many, even minor changes are happening, which actually affect itinerant side. So I think we should maybe do the itinerant cycle for each Fedora. So I think one of the things we should change is what other proposals have been there in the last release. That we should do, and it can give us a good idea. I'm interested in this thing that I could think to do before landing to look at releases and sort of changes in the last release and consider that. I think it's actually about dynamics. Trying to do something behind the flag pack. I guess it's very important, so I think it's very important to improve the experience for the users and the projects for the world. Then further, we could extend this also to other things like just for a while. We've been thinking a little bit about how we might want to do this, but we kind of wanted to discuss, you know, just to see what we wanted to do. So currently the frame of the release problem is gone. So if we do a live installation, then we don't get land packs. So let's say we want to get a land pack from a language mastery. Another big problem, one way is to install the... There's no way to send it to others, so it's graphical. There's one long way of trying to do something, and the first option is kind of non... This is the first time we must do the project, to do something in a kind of post-installation. So this is where we can rely on, if you don't mind, recognition and additional up-to-date transactions to... I don't remember if we need to do a live version. Let me install it. No way. It doesn't matter, it doesn't matter whether you're installing that post-install or whatever. Or if you do it on-down, and you don't want it to set up, you could have seen this one. I just need to check our store. I'll just... I'm going to build this, and she'll just be done behind the scenes. So, well... One thing I want to say about this is that it's not a non-dependent distribution, you know, I'm not sure how to do it, but I'm not sure that it's... It takes a long time to install that, so I don't have a risk, so I say it's not only changing the language, you know, but at least I do a session, and maybe it's changing the store, and managing the installation process. It's a bit fragile, maybe, but one of the first... It's... It also comes in at the second one. This will allow us to install the most in the first version. I think GNOME Control Center might be good, because if I want to add an input method, I'll normally go to GNOME Control Center. On the same line, if I want to add a new language, so my ideal way where should I go should be about GNOME Control Center. I have a different opinion. I think we should avoid implementing it in a GNOME initial setup and the GNOME Control Center, because GNOME is not the only desktop for Fedora. There are other distops. And also, GNOME is not only in Fedora, it's also in other distributions. So I think that the initial setup is a good place if we are going to stick with Fedora only. And even better, it would be to support it in a package kit, in my opinion, because package kit is a cross-distribution solution. Do you mean the software center? Yes, software center will be good as long as we remember that GNOME software is just one of the front-ends which are possible to build on top of package kit. So if we choose package kit, we are open for GNOME, for KDE, for other desktops, and also for other distributions. First of all, it's a crazy tool to install packages. Some popular languages also can help you to install the landpacks. But it's not that efficient. But it will install every man. So there should be some way and it should come from the original installation. This is the thing for making sure that whenever a Fedora operating system is getting installed, people should get the landpacks installed. They should not be needed to go and do something other than add these landpacks packages. I think that's why we are trying to find some solution here. When at first it was very hard that they will become, it was a bad word for them. And it doesn't mean that the other approaches need to come through something itself. And I think it was kind of attractive that it happens. Because typically, when you're in trouble, there's no problem once you're in trouble finding these solutions. Finding these solutions. Don't you already have to re-select your language during the initial setup? Does it need an extra step on that? No, totally not. In the initial setup, even if you choose your language, the landpacks itself are not available in the live city. But if you do... So your total operating system will get installed with the initial package set. But there is no provision for installing the extra landpacks packages. So you need to do that manually after installing the federal operating system. You can imagine why we do it from multiple places. In the initial setup, you set the language. You then go with labor in the default center and change your language in a later time. You might want that to also fire off this problem. So don't get me right. Actually, when you're in a selective field, so you don't have to give us the deconfusion. So are there any one of these things to fire on that signal? The question is, what is the difference? We are basically... We are tired with my air-con. This is the base. And with the initial setup, the initial setup and the non-conform setup. But when you say the non-conform, that means it's specific to the non-conform environment. I like to call the hero of making that something that package kit handles. And then it's something that, you know, the control center or the initial setup, the initial setup, any of these things, they can tell. And any one of them can tell package kit to do it, but package kit does the work. So the work is agnostic, but you know, you still get it. I'm going to check on that. I'm going to do... I'm just going to do... So we have this KJC for all the Japanese and Bikini for all of us. These are all supposedly the two reason diagrams. And instead of having these, it's still quite hard to find the quality of the kind of work for the KJC. But they're all talented to KJC's Google's, I think, matches for all of us. Okay, we seem to prefer that. China, the KJC, it's free, but it's not free. But then, I need for not for iBus, but for FCI-TX just on terms of some of the Chinese we've made in the past, Korean iBus handle has been mentioned for a long time. I'm just a programmer. So, there's a few things that I'm interested in, but one thing that bothers me, is have long treats in the mentors. The keyboard layout is separate layout. And there's one overall keyboard layout, or it consumes some keyboard layout. Maybe some way to win those. I'm not saying it works okay, but for a civilization, in some cases it might switch to their own brand. The layout you get in your engine might depend on which layout you bring. There's a lot of security here, so I mean, I'm just trying to make it simple. Also, the fact that we're making things that we made in iBus, and also, we're going to have some future in the future. So, to make sense, if it was a story, but other than that, there's no description to it. We don't have it. We don't have it for all different brands. But I think, mostly, widely used input methods, we have already packaging for them, they are available for you. There's a lot of changes we should make. We often didn't have come, it wasn't in iBus. We're doing some work on the emoji rendering. Actually, once the engine came out, we were able to really think of some of the rendering in the emoji. So, probably, it's one of the most known ones we've seen on the coast and to show you some of these. There's some changes we're going to hire recently. We're going to see better very different ways to input. But we've also basically implemented emoji improvement. So, that's the work for iBus and support for... We've also done more work on our stream on Pdps and also it came upstream of Pdps. I think for a long time the players haven't been doing that. We kind of started working on updating the Pdps in the code that we have on script. There's a lot of work and things. Another thing that we want to work on is we want to update some CLDR data time paper size et cetera. There are no messages unless, yes, no this is in CLDR nothing more. No, I said that's always what is in GMC. I'm sure there are a lot of time that's not very much. Mostly, iBus iModule GTK iModule and other things. Also, there's some problem with the cache path I think it is something to do. So, this is the first session which we call so much time and we are discussing it so deeply. From the form side any part of whatever the company do as we are using whatever you see, there's some fonts in the background. Maybe say in Federa, we are doing very good from the font perspective in Federa but things which we are more concerned in Federa is likely for fonts and when I say the default fonts normally the iDM is not that much for the English font because it takes care of the Latin font the English font for the default but the default font for the language is that we take care of the default fonts for the Indian locals for the Japanese locals for Sri Lankan there are many many and the issue right now is long time there has been many fonts came to the I would say 300 books a square and we have not revised the list of default fonts in the Federa that thing I always worry but same time languages are very sensitive people are very sensitive with their languages so if we want to change anything for any particular languages say I want to change anything for the French languages so I need to take some review from the community before I do any changes same way we should now revise our default font list sometime and see if we can select the latest fonts which are the more updated with the English font changes which hasn't happened the second point is not the fonts this is not the same line this font has wider coverage so noto is like noto the objective of the font is that to provide font for all unique what is in the encodings all the languages which are available in the code and it has a wider coverage so I think if I say in the free and open source world this is the only font which is available which has wider coverage so if you have noto installed in your machine it means you don't need to that much worry about other fonts you will get most of the things you will not get a square box then the liberation font liberation font is like though one of the older font which is a vitrious font and I often see some issues that we reported about yesterday those issues used to be very complex it's like one pixel it is rendering one pixel more or it is rendering one pixel less it's more from the empty side and I always find it very difficult to deal with those kind of issues couple of years back I think about half a year back to be folk cross squares font for the liberation tool so right now we have two versions of liberation font there is a liberation one which is a classic liberation it has a white gold hinting function and those are rendering much better and there is a liberation tool which is mostly automated hinting functions are there for liberation tool and there is a diagram few people say they want speed of liberation tool because it's very latest it provides you with a good platform rendering and few people say they like white gold hinting and we did that change around 20 to something but there was lots of users said it has a regression and we switch back to liberation one then but liberation tool is in use right now so liberalization people use liberation tool and same time the workstation that is recently reported issue bug they say they want to package liberation tool with the workstation then fonts coverage so fonts as I said what we try in default installation so if you are expecting that with the federal default installation you will get all the fonts so that is not realistic but we have limited stress so what approach we are following right now is like we are installing the fonts for BMP that is a basic multiple plane of unicode so basic multiple plane of unicode provides the script which are active right now so what we are trying to do is we are at least providing a one font for each script so whenever we see go to a web page or anything and get those points the cold fonts you will not get as good as those but same time there was one approach which is available ideally but it is not working that if if the rendering imports in particular code and that font is not available normally they query the database package data sensing if any other font is providing that so that is a good approach I think we need more testing on that side so please just if you add that I will just make sure that and fonts quality fonts quality so we are not we did a font audit some couple of years back and it imported many files but I think it mentioned so many warnings so why if we can just make that report more concise and see that what things we need to improve it will make it more sense just I think for fonts quality we need to be more realistic the report which I saw had so many things like 4-5 pages report for font and normally if there is so many things we ignored it that is just from font side but if anyone has a query issue or the solution for improvement side I think you can discuss it with me in a some time I am not going to say anything about this but it is an application for tracking and processing modularity we are starting to think now about what that means for writing I am not really sure what that means but it is still kind of in fonts or in font methods or in fonts and so on and so we have that kind of in this report we will come under the few other topics I just mentioned in case if you run into well as it is only related to modularity but for atomic in terms of atomic workstation that is the main one for the atomic as well but they seem to be sort of a workhouse for the atomic workstation but yesterday for atomic workstation session when I was saying the atomic workstation I think I am very familiar we are having a long time to work hells and so on this is a very small thing as well it is quite hard to find the US layout in this slide I want to bring is that in federal automatic testing right now from admin part of site there are many few things that are covered I was checking with Adam yesterday and we used to have that as well but of course when we consider testing so we have an indication on the people who participate and the test cases which are there what we are planning from onwards that we want to put more efforts on the automation so as I said the fonts so I want to have individual font level automatic testing like the rendering is not proper for input level so I want to just click it the input method in some say text window and check whether or not local testing so that is what we are planning in next coming year the main thing is test actually if you see so mostly we are from automatic perspective we are going to make model level testing so for font we are testing for rendering for hapus we have some test cases I think we will have some test cases so once we have model level testing I think with CIFM all things will be made we can reduce the time and also do the accuracy we will start working on a few other topics one is about hygiene versus globalization there are a few other topics there will be crazy ideas there will be fall ban