 Thanks to be here after before launch, lunch. Well, so I've gotten involved in Libreface since 2010 about, and I was immediately sure that it was a really interesting free-office street project, very interesting, which would improve the project at the beginning. But Libreface was created, I was sure that it could be improved easily, and it was the future of the productivity suites. For me, it's a great opportunity to promote free and open source software on the desktop of the user, and the productivity suites is quite important for this, in particular for professional desktop environments, to get a job, in particular when you have a job in the service. So it means that if such productivity suites isn't accessible, it excludes people from the community and the opportunity to have a job, to get a job. And when we know that in the world there are more than 36 millions of people which doesn't have any capability to see, which are blind, like I am blind myself, you have an idea as a member of the people which are excluded when some surface collaborative and productivity suites isn't accessible. But given the size of the code and the Libreface is 7 millions of code lines, so of course it's not really possible to imagine that we can affect someone to track bugs in each new patch and to follow any new bug arrival. It's absolutely impossible because it's just so massive work. However, anyone can help without specific skills inaccessibility, and it is what I would like to explain today for you. Such help would have of course a lot of benefits because it would highlight all the inclusive approach resulting from the freedom of the free software and the quality. All the employees which work with Libreface, in particular when companies migrate to Libreface, would be included in the same organization without any difference. For example, we could, thanks to the fact that the software is free, we could add some plugins, some add-ons to do that. With the same office suite, we get a fully inclusive thing and it's a role of plugins such as AudityTrubri, Audity2Degie, or tools as Gramalite, etc. Then help promote Libreface in front of exclusive tools would be easier if it would fix a lot of accessibility problems. Today it's not absolutely easy. I would say it's difficult. To validate the importance of the difficulty, we tried the following method. We took a training plan for Microsoft Office. We tried, well, these skills that are trained during a training for Microsoft Office. I know that all these skills are possible to be accomplished in Microsoft Office by a blind person because it's accessible from a way or another, but it's accessible. Then we'd say, well, is it the same, is it as accessible as with Microsoft Office with Libreface? Is it accessible, and if it is not, is it just purely not usable? Or is it usable but difficult? Or, well, what exactly is the situation? And we reported bugs from this base. So it's the method you use. And the results is, well, it is a result which is not so bad. We worked essentially on the writer and Impress. And we saw about 50 bugs and a lot of perspectives which make us interested by the fact we could collaborate together to make quite excellent productivity suites. So let's try now. Let's try some situations. I think it's important to show you what kind of problems we could experience and the level of stupidity of the problem, I would say. To do that, I will switch to a document and show you some interesting things. First thing I would like to show you is, for example, the problem we can have when we have the template manager, which doesn't have any labels on the widgets. So when we don't have any labels on the widgets, it means that when we tab with a keyboard, with tab key, and your focus reaches the widgets, the assistive technologies, which doesn't have any label to know what it is, say it's a button. So you tab, you get button. Wow, very good. It's a button. I'm happy. But what is this kind of button? What can I do with this button? I have no idea. And in such case, of course, it's easy to fix. But if it isn't fixed, the product isn't quite accessible. Another kind of example we could try is the spell check. It's a very interesting example. The spell check is a classical thing we want to do when we write a document, especially when you have misspelled a lot of words that I can do. Well, here it's an example because in my dictionary in French, of course, and I inserted in English words, so you imagine the result. Okay, so let's try to run the spell checker. So we arrive on the dialog, and now we are going to try to understand where is the misspelled word. So we mash tab, shift tab, add to dictionary, push button, okay? In your all, in your once, not in dictionary. And here I have all the sentence which is mentioned. So it means that I'm unable to know in the sentence where is the misspelled word. And because the information we have in this dialog is important to have the context, of course, it should also indicate in the display with a capability to send to the assistive technology that the word misspelled, the misspelled word is this and not another word. Otherwise, the truth is not really easy to use when you have a problem to see with your eyes the situation. Well, let's go on this dialog and let's go on shift tab because shift tab should enable to do the, to browse all the dialog, all the dialog. And here the focus is free frozen. So here I tab, I shift tab, and I don't know, but I'm unable to move. I don't know why. And well, for most people it's confusing because of course they don't know where they are. And using that in the daily life, of course, it's difficult. It can be considered like a problem of accessibility, an important problem of accessibility. Other kind of problem always with the screen reader. So the screen reader has the feature that it gets information from the labels in the widgets, but also it displays one part of the information because we don't have any global display of the screen when we are blind. We have just a line, a portion of life, a part of the line. The screen reader has needs, a very feature, needs to have the proper information to be relevant, actually. And let's try now to, well, read the comments that I inserted in these documents. To say I inserted, I should do the demonstration also that when I try to open the comments, to add comments, I can't write the comments, but I no longer can go out the area to write the comments. And when I write, when I try reading the comments, okay? So I use my overall case because overall case is the way to browse in the documents, and in the lines, and in the characters, per characters, and to have a speech feedback on that. And if I do that, oh yeah, because it's not okay. If you have a doubt about the fact I'm blind, you know that I am. It's good. And so I go, oh, yeah. And I arrive here, and the screen reader says, well, here there is a kind of character which is probably a comment or something. I don't know exactly what it is because I don't have the information, but I know that there is something which is particular, which is not the text itself. So the question now is how can I read these comments? Actually, I'm unable to bring the focus on the comments or to display the comments for the screen reader. So actually, I'm unable to know what the comment which is inserted here. So it's another accessibility bug which is easy to show. I don't know if it's easy to fix for this time, but it's easy at least to show and to understand. And it's a finish because I don't want to be so long on that even if I could show other kind of bugs if you're interested. I would be interested to do a slight comparison between LibreFisk 4.26, which is the latest fully accessible release, and this release, the devil release that I'm trying to run here, with the famous sidebar where we have the style and formats bar and the navigation bar. When I, in the 4.26 release, so here, sorry it's in French, but I didn't have the time to switch it to English, I press Ctrl F11 to have the list of the styles. Of course, if I don't press good key, it won't work. So at the end, I have the list of some of the most commonly used styles and more. Plus is more. Okay. And here it's very fine because I have the list of the styles. I can move with a world case. If I tab, I can change style of paragraph, style of characters, character styles, frame style, and I can do everything. I can also do the contextual menu to, oops, I'm not in the good list, but if I was in the good list, I could, here, I can here modify the style, create a new style from an existing one, et cetera, et cetera. At 4.26, I consider that this feature is fully accessible and it's very powerful. I could write a full book with that, a book with an hierarchy, an index, a table of contents, so it works perfectly. It's the same for the browser, the navigator with F5. We can access to it without any problems and have an easy navigation in the documents. Now in the release, well, if I do the same one, well, control F11, I'm used for the same reason actually, so it's in English, it's more simple. I go to more, more styles, and here the focus is still in the documents. It doesn't go anywhere. Well, so, Bruton, because if I press Shift F6, I go to the toolbar, and on the toolbar or the sidebar, there is a style which is open. Okay, style and formatting panel. Okay, so I know I am at the proper place. But now, if I try to do tab, okay, I don't move, or I arrive on a checkbox, an unchecked checkbox, I don't know what it is because there isn't any label, so I don't know what this checkbox, I could try to check it, but, well, I don't know what it is. And if I go to the work case, it doesn't work also. Properties, I have properties, a checkbox and a standard format, but the focus never arrives in the list of styles and never arrives in the dialog where I could choose all I want concerning the style, actually. So it can be considered like an accessibility problem also because I no longer can, absolutely no longer can, change styles easily. Well, my side, if I'm a geek, okay, so for my side, it's not so hard because, actually, if I search, if I tab, if I do a work case, if I do some random tweak, I can get something, I can get some interesting things. But for an end user, a typical end user, it's a nearly impossible operation, actually. So it's another accessibility problem we can mention, I think. So it was to show that those bugs are often due to missing of labels, missing of capability to catch the focus on the carat properly, and to do alternatives, that is, going somewhere with the focus of the cable, and not only with the mouse or things like that. And it had existed in LibreFace, but that's right today, it's less easy. Okay, so now, how could we fix or avoid that? Avoid that, sorry. Actually, there are three possible things. The first one, I don't know it, I'm not a technician myself. So the first one, I know a utility you know, of course, everyone, I think someone here, some people here can talk about it better than I, but there is a tool to generate the GUI, which is Glade, and Glade could enable to very easily fix the problems, in particular concerning the label, because it's easy with Glade to see the widgets, to see which widgets are labeled or not, and to label the widgets. So Glade is the first thing which enables to make more accessible your new widgets or your widgets without any efforts and without any specific skills about accessibility. And that's the most important for me, that these developers could be interested minimum to the question, but especially, I don't want them to have additional work, just they do a clean code, and with this clean code, the application will be accessible. And later it's the role of the screen reader, of the assistive technology, et cetera, and the infrastructure about accessibility to do the work properly. But to do the work properly, the application needs to send the good information, and it's an easy way to send the good information. For people who are somewhat more interested, they can go further to test and fix using a tool which is Accessizer. Accessizer is a tool to show, in GTK at least, and also in Qt, I think now, to show the structure, the tree of the objects which are sent to the accessibility structure. The accessibility infrastructure is called AT-SPI. And Accessizer shows you, well, here's the information which is sent from your application to the bus accessibility, the accessibility bus. And from that, you know what the screen reader will have to deal with, and you know that, well, there is a label, there isn't any label, et cetera, et cetera. So it's a good way to know if some feature is accessible or not. The last thing is to test. To test is not so easy. It's not so difficult. If you just use a keyboard to try a feature, try doing a feature exclusively with the keyboard, for example. And for example, you see that the style and formats, the browsing, the sidebar would be a problem. It's easy to know if you test with the keyboard yourself and not only with the mouse, and you are aware of the problem, and maybe you can think of solutions to fix. Okay, let's see what's the situation, a more complex situation now for the documents. For the documents, it's more complex. I have shown a few seconds ago the problem of the comments which are not accessible. I'm aware that it's more complex to fix, and of course it's another kind of work that maybe for the situation we could imagine to pay someone for fixing that if it requires so much time and so much effort. But, well, it's a bug. Also, problem in the tables, the tables are, or today, in writer, not accessible perfectly because when the focus comes in the table, it is confused. Maybe I have the time to show quickly if I try to come in the table here, table, but the focus, I don't know, I have five lines, but the focus is completely lost when it comes into the table. So we have solutions using tab K, using some random solutions, but basically the focus is confused, is lost when it is in this table. So it's a kind of other bug. Okay, what about Impress now? Impress has surprising results. Yes, I didn't think it was accessible at all, and I'm very happy to say that the slides you are seeing, I did it myself with Impress, with visual control to be sure that it's not so bad, but at least it's what's possible for me to do it. So it's very good news for me. There are some bugs, which could be, of course, fixed and to improve the product. For example, today we cannot insert any shape without the mouse. We can also have problems when we try to add text in transitions. It's a pity because today when you do that, you just have not labels in all the buttons, so the dialog is not fully accessible. So there is something to do, to do better, and if we do better, of course, we will have a quite good list, a quite good utility. Let's be aware that Impress, even of Microsoft Office, is not quite accessible. I mean that it works, but a visually impaired person who tries to do that would need a lot of effort to use it and a lot of visual help to be sure that the presentation is good. Actually, we could do much better with the free software because if, for example, we have predefined models with labels, we get extensive information which is sent to the accessibility bus, we could have the capability for the first time in the world that is a history to write presentations completely alone without the help of no one. And in the professional context, it would be very a good thing for a blind person. And with free software, it's possible. And it doesn't require so much thing only to work on the perspective of how to send more information to label things and to make things with alternative, with most and keyboard, to enable everyone to access to the feature. Well, what about Libreface Online? Very quickly, because as I said this morning, nothing to say of very interesting because it's not accessible at all. Because the tool kit, which is used to do the interface on the Internet, is a tool which essentially uses some picture and images, so it's very difficult to make accessible. Let's explain on the mid-2014 by Jacobo Perez, which did a talk about it and which explains somewhat the problem with Libreface Online. Well, for a second, we could help, of course, and I hope to help fixing bugs, to help financing the fixing bugs. But I think that, as I said, we cannot fix all the bugs and all the future bugs. So it's also the responsibility of developers to label things, to test the feature with cable, to try the feature, to be sure that in at least the most simple situation, the easiest situation, the accessibility is at least expected to work. So I think we can work together with the community so that we have an accessibility which is at least for the most basic problem fixed. And then we can see for, well, the most complicated problem, CS, it's the specific financing and funding because it requires time and specific skills. Anyway, what I hope we can do is focus on specific things, and if Dev needs some feedbacks, of course it can contact us and contact us is very easy on IRC, and we have persons now to help and to test. So if you are ready to go together, it will be very cool for us. Thank you.