 Yeah. OK. Here's the goal. How many? It's time. OK. My first idea here is one idea of the different tools that I require organization. And why we're working that out. There's a whole world of, probably tried to separate organization, a content organization, but we're also working with my industry. I'm trying to learn from them. What are the things that they are doing that we're not doing that would make our work easier? And it comes down to computer assisted organization. So we have the literature that we use to translate just into the organization. It's PRA, which we have. That's very un-controlled organization. We have the source string. We translate it and send it back. So we're reducing open source to a very kind of structure where you're doing organization. We have a lot more in-depth discussions. The question is, if you have put the profile to the translation, this is the profile info or something like that. This is not our translation. It can get also the same. And two people make the same role. Have you got the booking shape or something? Well, I think that can't be an idea. It's not showing us. Just more general about it. OK. So, glossaries in some systems are actually used quite naturally. So, perhaps the idea of that will work. It's first to have a very clear glossary. But then to be able to use it when you do translation. Somebody mentioned this idea before, that when you're going to translate something, the words that are in your glossary somehow show up. You can go much farther than that. You can actually say, please enforce this glossary or give me a ring when the translation doesn't have the words that are supposed to be translated. So, we can control that the glossary is being used. We can control many of these things. Or we can control them online or offline. So, this student offline translation agent is actually not a real program. This is your way. I wish this. Even if it doesn't do this. But the one that I was trying to do was translation memory. One of the main problems that usually the translation of the sources independently meets problems. So, there's very much shame of translation memory or sources that are very useful. You think that in what an issue is to be able to share the translation memory across products. The next thing is process and quality. The translation process can have, as you say, many, many steps. You have translation, you have review. You can have approval later. But what's very important is that you don't repeat it. And with PO files, it's complicated, for example. You have three messages that need to be translated now. And the translator knows that only these three messages need to be translated. But once they are translated, the reviewer doesn't notice. The reviewer sees the 1,000 messages PO file that he has to review. But the reviewer doesn't know the work that was just done before. But the translator, he sees the file with 1,000 messages that are all translated. So, by controlling process, by controlling the work that was done in the States before, you can save a lot of work. You can do a much better organization process. You can do a lot of checking of things. For example, what we're talking for about the glossary. Maybe you can do a different standard glossary. But if the file goes for reviewer, it says, by the way, these three messages do not use standard glossary. Please check them. Please check them. We'll do something. It's your problem. We'll give you the information now on statistics. These messages are disproven. These messages are disproven. Actually, the issue you talked about, translation is not being able to check what changed to the file file, are essentially true, but not always true. For most translations, I use SDN systems, you know, CDS or SDN. Reviewers can check out, I mean, maybe do a checking of the classic translation and then do a checking of the translated stuff. You can easily see what the translation actually means from when you did an MS in marriage and then you actually translated that. So that's, personally, of course, what I'm going to be doing is telling you that it's not now, it might be, so you can actually see it quite easily. The walker has a translator down to update the translation in every release of the message file. So you can actually see if you want to review the translation change, you can actually see it very easily, even if you have a source code in the whole system. The French team, I think, from the PO5, we, when it's just on a date, we send a batch and also the complete PO5. I think having a review of the complete PO5 is also feature. Because the more review you get, the better the translation will be. Yeah, I think that Avia was wanting to enhance, and it's important. I think being able to review what needs to be reviewed is a common need. We solve it by sending batch files. You can solve it by using the SCF or whatever. The important point is we need that feature, actually. Yes. It's important that when there is an update to see what has been updated and, for instance, to have not everything reviewed, it's good. It's good. We have to do it. I think when you open up this from 2.02 to 1.03, then you have five hundred messages in the course of all the files, but you don't want to review all of the open updates. Only those that are getting released. It's important to have some history. Yeah, go on, please. But process gives you a lot of sense when you review the files in your years. I don't want to say why it's interesting because we're mixing files. You can put all these processes inside the file so that when the file goes, it comes back and everything is in the file. You can actually know who to give you that very least message. And that's important. So that brings you to the process management tools in which you have the central but I think it was the same thing that was going on back of this. Something that controls these files and helps movement between the projects to the team leader for that language, the translators, the viewers. So the tool that manages all these that for us is the nature of putting. Documentation in other context is very clear to us that we have to manage that too. That whatever tools we do we have to work for software and content. I see that a matter of changing documentation to be sustainable and using the file files. But there are things that change because when you're doing a documentation you need much more context than when you're doing things like this. All right. But I'm not saying that if you file files we'll be able to spin but actually the tools we use in Debian 4.0 today actually use file files because I'm not good at playing five or six of them. The solution is included. So I think it's maybe half and half. But it actually works because the tools we have don't actually match streams but match paragraphs. You have to content there. So they convert from XML files to profiles in fact. I don't know. What is the point of the translation or documentation of the translation? No, no, no. That's not documentation. So the thing is that whatever you do you have to make sure that the tools that you do also work for documentation. If you try to use a PO editor that PO edits the documentation, it works. But you don't have the content. You don't have the three sentences before. Well, no, I definitely do. That's what I'm saying. You might want to check out the all XML files because they actually work with XML files so you better get the paragraph itself and not the streams. So you can actually translate the whole paragraph as a whole. The only problem with that and that's the big issue is that when it's like change is made of the paragraph you get the translation as a classic and it's not that easy to see where is the translation. Maybe it was a typo change and you get the whole translation as a whole. So it's not actually the... I mean tools, all tools actually work and we've been using that for the full type of XML is XML. And there's some of the models that are actually in XML and are you being used with both files? That's kind of that. And man pages, so they actually work with some minor cabinets. So it's not a manner of maybe at all to be used for that. Yeah. The issue with that is when you go to translation memory you want to always match as many possible sentences. Yeah, that's true. So you can put phrase separators inside the target. So that you can actually match parts of the target and not the whole thing. So if you change something on a small part of your sentence the rest will be matched again with translation memory. But the other thing is to There is something specifically to document translation. And here the main problem of localization is maintenance. You can do it if maintenance is not easy and very easy then it doesn't happen. So process is especially good for maintenance. And the future maybe translation memory machine translation but That's it. Machine assist translation. But the thing is this might work for a few real languages but it will never work for real. Okay. And the main issue here is that we think that we have to use the standards. There are standards now for translation there is no standard of ex-leaf. There are standards now from the organization industry standards association for translation memory information for terminology for segmentation of text for work-hunting so there are a couple of numbers that are coming out specific for this. Actually some of them are visa standards right? The... But they are not going through the standardization process of visa. And it will be only easy because the others are from visa institutions. And there are people who are very happy with it. I mean the abstract meaning is in the visa standards but this I think there is a lot of talk about it or something like that. Ex-leaf is the only one that has theory, the future of organization for both industry and Latin operations. The industry is slowly moving to the big manufacturers including its big funds for translation and they are trying to force all the organization industry for the tool providers to use Ex-leaf just to be free of some of the big guys. And they are running into the huge businesses that are not available to us. The tool manufacturers the best known tool for localizations for traders and now it is owned by a localization service provider who is interested in giving the best tool to his competitors. So it is developing better tools inside and it has stopped development of the tool. You don't want to use standards so it will be better. And this is when we force movement towards Ex-leaf. I think it is very impressive. I am using it now really but you know how different the files and all applications that Ex-leaf are this is good for non-cost users. So the problems for the specific organization they will ask the non-cost to handle them. We can see what is happening in Mozilla. Yeah, with Ex-leaf. And we are trying to take Mozilla to all your clients. So we are working with your clients in Mozilla. You actually even brought a tool to convert Ex-leaf out of some coal, right? Yeah, and I wanted to have it so I did it. I did it and I had previously translated Mozilla on Firefox and everything. Recently I made the tool to convert from XPI to PL and then from PL to XPI to have the translation and we started the translation but some other people had the man with that with it too. So I don't know. It was not in the translation. Yeah. So at the end for us the problem is that you have to maintain conversations with a lot of skills. You get to know how to use Mozilla Translator and other organization tools. So in terms of standards this was my speech for people to do with standards. So I'm going to skip a few. If it's actually closers are shared between translation projects. So these Genome, KDA and some other language teams do share the closers because they are actually the same stuff. I think this argument is a good one. Yeah. It's very difficult for when you want to do a translation project like a language project I may understand it may be difficult to coordinate everything. I see it in many countries that there's somebody doing Mozilla and somebody doing OpenUp is the nature of it. You cross the Mozilla translation list and the OpenUp is your translation list. There are two or three people who are in both projects. So there's still sharing of information. Even if people don't there's not the same people. I wouldn't say there's not sharing information because all of those guys put up closeries and I mean I have a fact called the Devin close in relation to the closery. I've derived from the KDE and Genome project even though I'm not in the KDE or in the Genome project because they put up a variety of closures so I mean there's information for a few projects even though there's not the same people in these projects but maybe that's not clear from the outside. But I mean if they put up a different closery it's quite like that KDA actually points to the KDA and Genome. And I think that you should know that I cross the language. We're seeing new organization effort people are starting new projects and starting from beginning starting from zero and it could be really easy to borrow from other people. People tend to work in isolation in most cases with high cost and low quality. Low quality, many times we have to be in isolation because people start with an experience they might be in the closery but it's not the same thing they use it. So people with that experience and uncontrolled tools tend to produce low quality. That's made a very negative side of the clause I guess we did that. My experience with translation work in V-software I mean it's not bad but you put it very by that I don't know what to mention. I have to put it very by something. It's way too personal I think the long side of the table was not a good set. We have a standard as a good set. That's the fact of the standard and that means that there are a number of pages that work with this standard. On the next day the thing that has to happen somehow we're working with translators more and more that are not technical people so we have to separate the translation part of the translation from the thing about it and this sort of tool this version of the tool any tool that separates that from the other things. So what is to be done is to use common formats make sure that we center the efforts to make only one set of good tools tools that apply the most important techniques of computer assistant translation organization reducing the learning curve for translators. My concern is usually for translators of very small countries of low experience even with computers, right? Yeah, even with computers. And somebody would develop a glossary but everybody else will hate them for doing that because the people who know English very well they don't see a reason why you should use local language. Ah, because it's Spanish in general. You don't have to go to very small countries. Ah, no. That's a very typical problem. Very typical problem. And what is the huge problem in Africa? Africa. It's a huge problem. People don't localize because they don't see the need. They don't. You can see the talk of the guys in Africa you know. Are we using it in French? Yeah, actually an article on IEQ magazine, I was talking about Africa's software and actually there were a lot of shops that were making software and big advantage for them and they got funding from Microsoft and big advantage for them. You know, they were making like software to make to manage a shop. So like, you know, sell stuff in Victoria stuff and finding stuff and the big bonus and that's what they would go but funding from Microsoft actually could localize stuff and make it proper for the guys. So, I mean, maybe the high tech guy that, you know, runs the food south Africa with a lot of customer mobilization but the guy in the shop doesn't. Yeah, of course. Is it what you have, the user mobilization? Yeah, the user mobilization. But there was a study not long ago by a few NGOs in Africa technology NGOs and they actually asked all these people who have labs, computer labs and they would say localization from what? What for? What I summarized last year, I know English sucks. We have to say this to all hackers who want to English sucks. English is not the way to work. It's not the user's language and we have to do this all the time. One of the studies that showed this and we were learning to do a study in which we actually proved this. Actually, I got on this showing when you fit people in local language and English, we wanted to do it in India because India claims that English is very important language in India. And we can show that even in a place where India is very important some people learn, some people don't and it depends on the language that you use. Moving back to the subject it's important that we study projects. The curve is very small, that we know the tools. So we try to do the same tools for all projects. Tools must have techniques for the stream projects. So that we can actually integrate things. For example, now we are we are connecting directly to building open office. So we build open office directly from open office. So we can connect very well to that. We can probably not connect and go see that but I'm sure that we can connect to everybody from the CDS. Why not? Because they don't want they have their own infrastructure, they have their own very, very technical thing. Essentially, we're not using important industries. It's hard to use them for technical Yes, and they have these permissions and sometimes they take nine months before they exit the translation. So there are all these issues that complicate it. We don't use the chain mart. We use different names for the products and we work just as open source. We won't see that. And I think the easiest to localize products is there because we can do independent names but we leave. Even if it's forbidden we don't want the translators to translate and then try to get the infrastructure to do the rest of the work. And then our goal is to provide the projects, in this case like Delia to the instructors that they don't have to develop their own. And this is what we do. This workforce for us. We simplify the process of organization software I'm trying to do quality thinking of the world of quality translation. We develop a complete set of tools for computer system organization trying to do the best possible job with inexperienced translators. And then a methodology to ensure the quality of the translation. So we understand really the implementation process and try to use it in a very flexible way that you can use different processes. But at the same time whenever you decide what your process is to be able to enforce it. And then with the same infrastructure to provide the finished application for the other components. Our conversion features we have we have a number of features for different formats and then in second level for different projects. So from a C that you wouldn't have the properties files but then you have to open the XPI file and prepare things for the organization that will be second level. Other features produced at the PO or XP in the changing. So that we can allow people to work with whatever format they want in the XPI. Yeah. It's a nice drawing of what we are wanting to say in all slides. Modulography actually. You put everything. Formats. You put a little bit of what you were working with BF files and we think that we choose with XPI files. XPI files are similar files. Standard. Standard. That is changing very easily. So it's now 1.1. 2.0. We come out of just fixing very small things. So it's actually quite stable. We think it has a lot of advantages using the same standard for post and for private industry. Same standard for software and for content. You can store many more types of data. Actually, this is quite well known. You can put some data or you can put a lot of translation memory inside the file. You can put glossary information inside the file. So you can do any file that can go to a translator without further intervention without further information. You can put research of tests. So you can put in the file these three messages to not agree with the glossary. You put that in the file as part of the file. So when you go to the states the information is there. You can put counters at the end of the file to say how many messages have this condition. So it's easy to write statistics on the set of files of the project. You can put actually translations. So you can put French in there and then it's in front of the translation editor to say every language that I put on top. The one that is a source or the one that I have for French in the alternative language. I can say that my central system is putting in any translation memory match that has a match of 470%. I just put them there. I don't put them in the target. I put them in alternative translations. And then it's up to the offline or the translation editor used by the translator to say I take these and I don't take these. Or I enforce these because I don't enforce them. But you have the information in the real file and then you can put the processing information. You can define different stages that the file goes through. You can go through enrichment with the translation memory and glossary. You can go through a phase of translation. You can go through and each message, this translation is associated to a phase. This is the very important thing. So when the translator translates only the three messages that need to be translated out of one thousand, they are associated to a phase. And then when it goes to review, the reviewer says, I only want to work from the last phase. Or I want everything. It's his choice. And then he has to review this and he can get the statistics of these five-match three messages that do not agree with glossary or three messages that have problems with capitalization or that have the space of the sentence that the original doesn't have. So we can actually test a number of things in the translation editor very close to the translator within that all the tests should be as close to the translator as possible. And this is the interesting thing about modulating. Tests that we can do for the central system we can also run on the translation editor so that the translator knows right away that he made a mistake. Or no. Or he can think and say no. He liked my translation and I send it back. All this information can go there in the file. We can put the translation in my information in the file all the links as a reference data markup very important that you can say in this message do not translate this part. So for example open up is held we have parts of text that are XML that should not be touched. We can mark this as not touched. We can mark variables as to not touched but they have to be in the translation. So you can enforce that the variable is untouched in the translation. The filters will be the ones that will mark the text for each project. But then the tools are the same for all practice. So within that with all this data we can include the value of translation. By using more information. Then translators can use this information in whichever way they want. But the information is there in this file. And then for the standard standard for translation memory is changed, TNX you may have a database of more businesses than we just had before. But if you want to give it to somebody else you can change it to project then you can use TNX you can use DVH for terminology for glossaries and then coming back to the workforce project so conversion test is the standard for this. That's why I study this. Glossary tools for glossary management. But anyway, you can use any tool that produces PDFs because then we will use the standard for but that's the goal. And then when you want to put the glossary information inside in the file then you have a black box that you can get the X-Leap file you can get the glossary in standard format and then you can output it in X-Leap file which is in reads. You can change that for any other component that does the same thing if you use standard format. Same thing for translation memory if you use standard format the tool that you use for managing the translation memory is Iberma we will do one but if anybody wants to change or they have better one they can use that and they produce DMX and they can use that. Number of tests we can mark the variables inside the text we don't have to have an intelligent editor that knows the format of the variables of OpenOffice or Mozilla they are both marked as don't test this and has to be on the translation. So we can do more and more standard tests we can program these tests that you want to do for example we can test for the specification of the original on the translation but then you have to make it programmable because for example in Thailand there is no operation or in Khmer we use a different sign so you need a number of standard standard programming tests and these tests again they are short of the tests when they are costly they can be put back in files so that when it moves to the reviewer or to the program as a machine manager they know that this is probably a mistake without having to look at the translations we are working on an offline translation method that will try to implement all these tests and then we are working we are also trying to work on product side tools we are interested in understanding for example what Debian needs on the project side how do you do it to your original process to CDS for example yes, how do you upgrade you are probably using GetTex can GetTex be upgraded to fix me for production in several months I've heard that there is somebody doing a web project also to try to do it we have an online translation which is the original output the original idea of output was doing an online translation but then couldn't get the translation back to the system becoming a file server for translatable files so you can either see the statistics and what is happening or download files you can upload files I mean I can actually merge the files from the translator from the one that I have included because one included maybe a newer one but we will use xlit as an internal form but we can export xlit for PO and we can upload xlit for PO but as xlit has more information we will always keep our internal form of xlit because that way we keep our information in the system we are we are now trying to specify this as much as possible so that's why I think we are in a good moment for the STANs what is needed for the different projects I think I have to add another idea part of what we want to do is advance statistics tools goal management bytes management and what you were saying is very important you need somebody to take control of projects to be able to delegate to translators reviewers understand very well I don't want to see the hierarchy but who gives rise to whom so that you can have a controlled localization environment who can upload files who can review who can only translate so there is quite a number of things that need to be defined and by defining goals for the project there needs to be implemented for the project to meet deadlines that open office 2.0.4 has to be on the 15th of June you have to do 15 streams a day at least our idea is to try to have a system most people work offline but they are encouraged to upload their content as often as possible so that the statistics on the system can be maintained a bit like KDE is doing with CVS CVS would be behind the backups they are used to CVS more importantly how do they do it that they push all the translators that are used to CVS on the file that's the idea but we believe that CVS is a hacker tool that's it and the next day you have the statistics and more often that people upload things they better adjust the statistics the point is most important is having them send the material as soon as possible whatever method is used and we come back to our modularity too if they want to send it to the online web system because this is the only thing they can do or may or CVS or SVM they can do it depends on the skills of translators and coordinators I think you made exactly what I've made in the Indian world you are very skilled translators and translation teams mostly the biggest ones there are small teams of people who are less skilled this is my experience with the Vietnamese translation team in the Indian who is made of one millimeter do you know about her we are bringing them in this community that's really nice anyway she has no skill technical skills teach her SVM loans she works a lot hours and hours she works with other people no she's a big one but she's made more than some people and she doesn't understand yeah that's very important is amazing do you know about her do you know that she's very skilled I don't know I don't know if she wants this to be more public yes yes in the last weeks we have been able to bring an organization of open office and other things together also under her okay she works a lot yeah she's impressive so yes we're talking about ways to update information so basically what we want to do is build a traditional marketing system basically what you want yeah actually yes yeah you're mostly defining what we were talking before with different words but I think that it shows that it's not a surprise to see that you have them probably at a higher scale so we can consider they can be one of the interesting projects maybe one of the biggest because we are maybe that makes it interesting to if we need to do the one thing we have to work on so we can take stuff from all the different things the problem with IBM is that it's not like on KDA that you have single CVS but you have multiple projects that maybe use different source code management systems CVS at the end of a star art what are the apps you want to use you have source practice that are from there so the big thing is that you actually want to work without streaming directly instead of going through the devian package because then you get if you try to go this way you might not get the connection to but if you go upstream and then back that would be the best way to do the devian so actually usually the thing would probably be focusing on its first filters that needs to be enforced or not because some teams still have management process and others don't want to use any management process and the ones that do have a different team a spanish team and a lot of the teams here actually the same management process of translation lock you would say so that's the solution I'm going to work on then translate then update translation then review and then that so that's the process that probably needs to be implemented somehow this is what most specifically we want to enhance one or two things the distributed nature of the devian what Javier said contrary to Normandy or Portland we don't have any central point we have like 50 we have like 15,000 packages among these like about 15,000 after translatable materials because we want to translate all the package descriptions from all packages of main pages so most packages have main pages but not all packages have a good framework to translate the main pages you will know that the traditional method to translate main pages is absolutely not scandal so we have some work in the here main page maybe focused on devian specifically like deep package I don't know how many devian native packages only a few because there are more devian native packages but some should not be devian native packages and we often do that they are maintained but there is no point for this the difference is not necessary the most important point is we have many many source of information so devian is not a single project it's more like a big collection of small project the way the devian website organization was is not exactly the same way than devian's total manual organization was and so on and it will probably never be so yes we have to be prepared to have some high model work one of our goals is to be distributive so we will end that year but it will be bigger to be a distributive system we will have different work servers that communicate with each other so you know for devian projects the owner will give you some living for other projects it might be so yes so you can have full service for devian but we will also have for example we will have public service so part of the work is the the spanish the spanish the spanish the spanish the spanish actually one of the things that might be good to do in devian we were thinking one of the things in software one of the things how you implement it and how you architect that one idea is maybe how to put the server for devian which might not be a good idea because each time it seems has its own source and might want to manage kind of micromanage their own server so I think there is a which one it is that has a full server already up and running they talk about it in many ways in devian yeah there is but they mentioned that they started using it I don't know what is the name of this stream which stream I think this stream I heard from this stream yes, how do you see it no, no, no, but she has a translation of which one is stream the stream torrent the torrent they were using the full server for the translation I mean actually a language stream in devian I didn't no, it is I think it is something for one thing I mean one of the problems we have in single standard server is that maybe you have issues with that server that you break all of the translation itself and actually it might have issues with bandwidth you know, hour or whatever and it might make more sense to have dedicated server support on it so that you know they sit down and people are working on this server also because the different teams can probably get resources local resources to provide that service so they can go out again bandwidth at their ISP level on the country their own server which might be better than maybe doing linking to a server that is on the other side of the ocean to have a server locally in a place you can go to although you can go to a team on your own land not probably not because they are country-wide but maybe on your own back home because at least from the Spanish team we can actually get funding to put a server up at a very high level of the patient in Spain that might make sense to us the population we actually already have one server so it's actually out and it might make more sense to have one doctor that translation team can manage but kind of the feed of translation from other points that once you have thousands X-ray files and marketing so what we have to look at with this is that there are two ways of upgrading the organization and you can do it for example the way they're against now which is by language from the CVS so I have to go and do my Spanish I have to for a given file and if I want to upgrade it the great takes place in CVS it doesn't take place in putters and then I import the new file into putters but for other languages and for the systems I have broken up it our ideas will bring the POT files the templates into putters and putters upgrades to the projects because there's no CVS okay, yeah that probably will get issued with upstream lines and with you know that are their own upstream because they won't have handled you know if they want to control all of the files so it might make sense to like what Kucin said to have an output way to say okay Poodle has access rights to the CVS server so don't push up the CVS it's Poodle doing it but in the end it's the upstream CVS or SDN because they don't want to have files outside of that because if they go on it's like an input that the translations are difficult or KDA even that they don't want to have source code program here and you have a package and then you have packages for translations they say they want to give it to you it depends because it's that emerging from that kind of language package this is the talk I had with the frontspot because he begins to worry about me adding more and more languages to the I so it's adding a lot of load to the I so we are thinking about a kind of language package but maybe for the I not for most parts of this I think it's something that we have to think about for this process that's something that we are less than half an hour now way less than half an hour 28 hours are you ready or you want to show us something I think this is the only thing from the standard that always changes that all the tools are outside and all the changes are done to the standard program we would have an API for features in Python and then we actually have money to actually get a number of features so you are funding you are funding we have money for the whole project and we are funded by the Open Society Institute your source foundation and we have another funder as we accepted but we have decided isn't it something that global governments will fund I probably think that in Spain to send the project to the Ministry of Industry and they will fund the project this is something I try to I mentioned to a few people I go to to Bunda in a few moments to attend some kind of launch event and it's interesting because in seconds and that project is government funding and my understanding is it's not it's funded by IDRC I think it would be a good time but they try to get some support from the government they have a government organization this is the Department of Information Technology the government so the people who are there are only government people I mean if people are working on a project working for a government or an organization in my opinion this is government funding the project yeah it might be maybe you can use that because there are a lot of countries or whatever I think this project goes as something that could be people, local people get funding for even if it's not a debting project but it's something that's going to be used by them because I mean at least in Spain I think that it would be very interesting for us to have that funding because that means that the project continues and if we're going to use it we know we're going to be stable we know we're going to do resources we know we're not able to stop we know we have to act for that to be developed and maybe that's something you can also think about that they're going to listen to if they decide it's included they can also be a way to get funding from the our institution by the way the Workforce Foundation I'm going back to Spain on Monday we're going to create a fund the Workforce Foundation in Spain financing the Workforce Project out of Spain we're going to do it spiritually we're going to be in a special organization we'll talk later about that do you want to show us the comparison? this is what you're going to see when you go to the fund then organize a one day trip for this place yeah I'm already waiting for that you're going to go? yeah I'm going on three or four days just after that I have to give a keynote to all the all the official people where? in Tintu you're going to go to Tintu yeah I'll show them the real installer in Zonka and we'll have it in Nepalis too and now we have it in May have you seen it? I'm not sure you put it with installer in Timea this is the latest version oh, you're not talking this is the one with the files for the installer the Timea installer Timea just sent me her Vietnamese files but we didn't have to put them out but the idea is that it's at the end we go to the installer and then we can actually work on time and see the businesses that are not just waiting for the files I've been waiting for four days while we were here Internet, yeah but I have to say that I haven't tried this from any other connection but we're going to have a number of files we're going to have the number of files and does it update automatically from the DRM? we have to send this to the place now and now we have to connect these to make sure that it works can't see the coordinator who happens to be no, no, no cacadize my program for the workforce so we can just come here and download the TO file so we can view the interactive part the interactive message is in that file it's not complete where's the actually we have to review the the score and how it works actually, X4? no, it's version of it doesn't use XDTL we're using the base classes for about two terms specifically before we run all the amount of time in the general status of Twitter for example yes we actually had Nicola look at Twitter and try to package it for Libyan at some moment we're going to use something in Libyan that has to be packaged so yeah, it's not a forced table really? so this is basically where we are now and the goal is having a Libyan package for Twitter as soon as possible it will not necessarily mean that use it natively but then if we have a Libyan package we are in the situation where we have the Libyan maintainer maintainers and an upstream which would happen to be workforce and then we can collaborate we have set up the foundations collaboration that's good this is good we do not have as we mentioned before some performance issues that we are working really hard on but does Poodle use a database package or is it just a browser? we don't yet use a database the problem with Poodle is that it uses PII you see what's that? it's a database engine but it's used for searching strings which is not an important feature so we have to we have done some basic work reducing to one fourth for time that we use for things but we still need to do a lot of work if we don't get a very smooth engine then we cannot get it with the stones Poodle so we are very well at that basically we have the development time for the year we have what is now the table that we are using we are moving towards our next milestone probably June not a question for me it depends on the specific version of vital I think I don't know I think if it works with 2.3 but I have some warnings and I think I do think that we should 4.0H or 4.0H but I don't know but I think it's 4.0H 2.4H but it's not we are very limited in the production production we are talking about the stable engine I think that we should focus on the optimal 4.0H when I speak about stable I'm not speaking about such we can compromise that everything that we want is okay but hopefully we know that we will release H on December 2006 we all know that we will release H we will release on time so yeah we have to add when I say stable it's to be H so if we depend on vital 2.4H yeah it's in H yeah it's in H yeah it's probably it's on the full Python but it's okay it's in H yeah I don't know for an output server and is that on the performance I don't know a team that answered that but I mean that could be that is not as much of a side performance just to give you a very quick idea of the development we are planning the next version of Kool and then export the other ones after that we will place the statistics so that we can show what's happening with these two files then we work on bytes management we call here rollings they find out very well the roles and the bytes operation and doing that and then and then we do the integration and translation memory inside the files that's for us the next step but it's important for us to have a stable version that works internally with 16 files and that can do basically as a database to that for us that would be full of 1.0 okay that would be August right or September that would be before that but that probably would be August I think that by September we would be able to do that yeah there's a paid development team you need help with that we need more people there we would find more people we have to talk to the other director of the team that is doing making they're working in South Africa but we're working now we're studying on boundaries for specific things for the statistics but I'm sure that we could use more developers especially now we're going to have a second part for funding and we'll have money for more developers yeah and where does the developer work do they have a mailing list but do they have a mailing list of something that people something that people in the university would subscribe to and that also you have to subscribe to this there's also an IOS channel yeah there's also an IOS channel so and then we're going to the last statistics management of those and this will be the leading to 2.0 but 1.0 is a stable file server that works in terms of 60 files that can communicate with the CDS and that will be the summary so by the beginning of September we we will have a stable 1.0 which has to be tested and then by the end of the year after we get a stable 1.5 with all the added value we'll start working on distributed on having several tools and then local tools or specific developers tools or specific project tools and with them this actually is more a specification to be able to understand who would want to use the root server and how it has to communicate with the other servers I think we would have to cut it down because probably I don't know if you want to repair their own buff so we have to give them yeah I'm starting to train them I know you're not doing your slides you know it's different for sure but we would love to collaborate if there's a dev team working on a specification who would like to participate we think that it's a good idea that they may be able to understand the specification yeah I think we'll probably match what it would be like so we're not going to actually for some of us it actually doesn't make sense for us to write all this stuff I guess we might not be able to do it so we can require some of the team maybe we can discuss with the people about what they might show and maybe find some specific place which has some connections with Julian and we could assign this as a project and they're not suddenly working and as a second goal I would say that having some of the people of those in Extremadura would be nice we have been talking to them about the with CASAS he has been saying that maybe he can put one of the drivers I would say that's having I didn't want to it's a long time from south are there expansion for us one person that is Spanish is but he's in Sweden but he's maintaining coolant so he's in charge of making sure that he's basically the client of the of the South African team so coolant being developed in South Africa the translation editor he's been interested in developing Napoleon and our client is a writer so that he if you want to have to complain about working and to make sure that there's a lot of work to be done what do you want to do? on the runway do you want to set them on the runway? no, some people left the runway keep it up