 So, well, welcome here. Thank you for coming. This is my speech about pushing writer to its limits and see how does it behaves. This is more than a speech. This is just, let's say, a workshop, meaning that feel free to interrupt me and to contribute and to tell your experiences or whatever you want during this kind of presentation. First of all, let me say that part of my job is editing and preparing technical manuals for industrial machines and, to be honest, also military ones. This means that those manuals are kind of from 300 to 600 pages manuals. So it's quite a challenge to do them, but it's part of my job again. And the funny thing is that I have written black and white on the contract that I want to use LibreOffice, no way, because they wanted me to use a word, but I told them, no way. I have my tool. I use my tool. And if you want my work, you like my work, you have to accept that. At the end, they just want the PDF. And so it's my choice how to prepare that PDF. Said that, yes, I told you that, and it ends on and so on. Then this is the agenda, not updated, to be honest, because I have did something right now and I forgot to update the agenda. So let's go direct and straight to the issues. You know that when you have to do such a manual, really, often you have to include images, pictures. And on most of them, you have to put many call out indicating what the various items are. Imagine that they can have pictures of this room and I want to say, this is a chair, that's a camera, that's a computer, that's Svante. So I have all those call outs indicating, pointing to the various items describing them. What happens is that when you create those call outs and put on the image, they stay there until you save the document. Then when you reopen back that document, not always, but most of the times it happens. I don't know why and when. I mean, now and then it happens that those call outs go behind the picture. And the fact is that if a part of the call out is outside the picture, you may still see it and bring it back to the front. But if you don't see it at all, just like happens many times, you may forget that you have put that there. And so you may publish the document without that call out. That's why I have to give a name to every single object, which is a maniacal job, but I'm almost autistic in that. But jokes apart, it's the way for me to at the end of the job, at the end of the manual, double click on every single object to see if they appear. Because otherwise, I could lose some of them. In this case, I don't know if you can see here. It's just one pixel, but I wanted to enlarge here. As you can see, this is the zoom of this picture. This is the picture. Those are 1, and 2, and 3, and 4 call outs hidden. And this is how they are hidden. And thanks God, in this case, I could see at least a part of them. But most of the times, you can't see them at all. So what you have to do is, one by one, every single image, click with the right button and arrange, send it to back. And this way, the image goes back and the call outs come to front. And anyway, many times it happens also that you can't select them anyway, even if they are in front now, but you click on the image and you have the image selected with the knots, you know? Then the nodes, I mean, then if you click on the call out, which you can see now, sometimes happens that they don't select. So I have to pick up the arrow, define a frame, let's say, and capture that object. In that way, I can all otherwise pressing tab on the keyboard. But the result is not that implied, because I don't know which is the order of the objects that Brighter gave to the document. So pressing tab could bring to the object, I presume. I want to go or could skip another part of the document. I don't know. Yes, please? Yeah, sure. Yes? I know that. But this is simple to say when you have a few objects, but you have thousands of them. Trust me, it's not that simple. And also, not always works, but most of the time. There is some issue also there, please. Oh, wait. There is a microphone for you. No, just because of the recording, you know? Maybe if you group the images, they stay in the same place, in the relative place, one over the other. Have you tried it? To be honest, I can't remember. Maybe I tried, but I'm not sure. I forgot to mention that I used to do this kind of job, let's say, not regularly, not every day, because they commit to me to make a manual, so maybe I work on that for three months, two months, four months, whatever. Then I finish, I deliver, and that's finished. And maybe for six months, I don't do this anymore. So all the speeches at the workshop is based on my memories, because the last one I did anonymize and I'm using for this workshop was delivered on August, well, July, and a small modification in August. So I mean, they're past a lot of time, and I'm not sure to remember all those things. But by the way, and I would first see all the issues, and maybe if we have still time, we could do another round, putting the hands on and trying with the document open, because I have it too. Then again, about those call outs. This is how the final, it is anonymized, obviously. I had to sign 12 pages of NDA for this document, so I can't reveal anything about that. But this is more or less the layout of the document at the end, or part of the document, obviously. So imagine this is the picture of the machinery, and those are the call outs. What happened here is that, as you can see here, there is the anchor, and this call out, in this case, was anchored to this paragraph. This wasn't wanted mostly, but it's not that easy to have another kind of behavior. What I mean is that the ideal situation would be to have kind of this anchor to the image itself, which is not possible because it is not a paragraph. You could link it to the page as an absolute position, but then you may have kind of modifications. There are solutions. I mean, I have to say, it's how many hacks you could do, but work around, thank you. That was the word I was looking for. But it would be interesting and nice to have the possibility to put the anchor here and then move the object without moving the anchor. Kind of holding, let's say, alt or control alt, whatever, a shortcut combination on neutral keys while you drag the object that prevents the anchor from moving. Because what I have to do today is bring the object here to have the anchor here, then click on the properties of the object, right click, and then specify mentally numbers, five centimeters, five and 20, five and 25. Then I take note of that because I have to reapply the very same thing to this one, for example. And that's what I used to do. You laugh, but trust me, is nightmare can become a nightmare, but this is nothing. The worst is the Z-order I mentioned before because imagine that when I finalized the document, you know, the 400 pages and I have to export in PDF, it takes me for those issues, first of all, the Z-order one, hours. Also because we'll see later every exportation takes kind of a quarter or half hour. Let's see, let's move on. So it would be nice to lock somehow the anchor somewhere and then drag the object without moving the anchor. An option could be when you insert a caption to a picture, what LibreOffice Writer does is create a frame and within it puts the image and the caption itself with a specific style. But it is not always possible all, how does it, willing to have. Indeed, in this case, you have a table. So having a frame or we did in a table, I would avoid, let's say that. And especially if you don't have a caption because you don't need it in that situation, that page, I mean, it would be just an overload of objects in that part of the document. Just for the purpose of locking the anchor, I would, I find it too much, by the way. It could be an option otherwise put some in some dialogue to enable or disable the following of the anchor, I don't know. Then, thanks to Mike Aganski, who was so kind to be with me for the whole morning and check all of this stuff, check all of these slides and issues. And he found this bug. This is a link, so it's clickable in the final file. What happens here is that the image goes beyond the margin of the page. And he showed me also work around, so it's not that bad, I mean. But somehow it's something that could be not desirable because from my point of view, the margin is something that you cannot violate. Because for example, in this case, there is the footer and has been covered by the image and it's caption. This case, the frame I mentioned before. This is cross-referred, this is not a bug. It's just, could we solve somehow, not solve, I mean, could we find a nicer way to handle this, but maybe the answer is not. It's just a small issue, I have to, not an issue, to be honest. By the way, I'll tell you. I forgot to mention that when you, at least when I have to work with so huge documents, again, about, in this case, they were 412 pages, it would be a nightmare to have just one single file. And we'll see how and why, because it becomes really slow and really heavy the file. So my recommendation and what I do is split it in several files, usually one per chapter, and then join them into one ODM document master. And that's what I do. But what happens is that somehow and really often you have to refer and put cross-references between different chapters, for example. In this case, this was a troubleshooting. So you have the travel, you have the action to do and then the card, because in this case there are the maintenance cards. And so I had to link to the card, which is on another chapter, so another file. So what I do is force LibreOffice Writer to insert a custom cross-reference. I just insert the name I decided before on the other file. It doesn't list it on that file because it's on the other. So it just inserts because I force it to do, but it's fine, I mean, and that's the way I found and I found also as a kind of documentation to do. I insert this cross-reference and it gives an error, which is right because that cross-reference is not on that file, not on that document, correct? So it's correct to have an error, no problem with that. But what happens is that this error in this case kind of destroys temporarily the layout because at the end this will be just in this case, C027, for example, while at this moment it is correct to have the error, but the cell becomes this tall, you know, and all of them. So imagine that a chapter of maybe 30 pages becomes 50 pages just because of all those errors, which is fine. So talking with Mike, he told me that those cross-references can't have the path of the other file just because indeed when you merge all those files within the OGM, they become virtually just one document. So it is not an external at the end. But because otherwise you could include the path of the other file here in the cross-reference and you may have a function that just checks quickly in the other file if it exists and removes the error. But then will be a problem when you assemble all the files, all the documents. But could be an option to have a code that handle these, removing those path when they assemble all the files. Then I don't know the solution, I'm just showing which are my issues when I work on those huge files. Yeah, otherwise I must say that also this error is really useful to me because at the end of the work I do search for errors because I may have been mistaken. I mean this is not a real error just because I told you it is an external file, okay. But sometimes I do mistakes just like all human beings. And so I need at the end to search for this error. So I need this error because I do look for them. But this is again something I can't do within LibreOffice Writer because if I look for error it won't find this because this is a field and not a text. So either we have a search for fields and there is an extension to do that and I used to use it. Or implementing a function to search for broken links. Or both of them, I mean, okay. This is our suggestion obviously. You may have better ideas to how to solve this. Again, not an issue but this thing. And we come to the tables which is probably, I don't know how to translate it in English. Achille, tallone di Achille. Achille's hill of our beloved writer. In this case you can see that here there is a cell but the border is not there. And if you click here you will see the cursor blinking but if you press F2 to insert a formula for example which is something you could do usually in a cell of a table, you can't. Because so it looks like it is a cell but it is not some strange behavior. And then this is another interesting thing I would like to. When you have merged cells, especially vertically ones, it may happen that if they are long they are split between two pages in this screenshot. This is just one, this is page prac and continuous here. Being just one cell, the text is in the middle of the cell because I decided that way and it's fine. But given that it is split between two pages, would be nice to have it repeated here just to avoid losing the context. Just a suggestion, I mean I don't know if it's doable or not or is it simple or not. As I told you maybe in other occasions, I used to be a developer but not anymore at least. I can't define me developer if compared to the ones here. Sorry? What? Oh, yes, assembly, yeah, sure. 80, Z80. Let me go further. Oh, this is another interesting thing. The table of, yeah, okay, I think it's a contact. And table of contents is, usually they don't give such a headaches because you have a short document just to create the table of contents. But again, imagine that I assemble this 400 pages. And the table of content itself could be 20 pages, 15 pages, for example. What happens is that while it creates the table of contents, its own dimension, its own size, lets the remaining content shift down. So imagine that it states the chapter one is at page four. But when the table of contents is finished, it's 15 pages, now it's not anymore at page four. Maybe it's, now chapter one is at page 19, for example. Just because of the table of contents, so what you have to do is to do two passes per each because each of those table of contents will grow. And you may say, yes, but this is just once. No, but because when you use ODM files, every, at least to me, every time I open the file, those are empty. So every time I have to update, and there is a specific function under the tools menu, update indexes and tables, I didn't have the chance to check if the solved the problem because this is something that this morning Mike told me to be honest, I didn't notice this function. And what I used to do is click one by one with the right button and update index, update index, update index, update index. And again, four times because I have all of those four table of contents. Olivier. No, I think that if... Okay, okay. Table of contents because you can select what kind of headings you are bringing to the table. Yes, but I'm required to. I mean, the table of contents I do is just because they want me to do that the way. Because otherwise I know that I could have, but anyway, the problem, sorry, appears just when the table of contents is more than one page, because when it shifts down the content by just one page, which means that the table of contents is two pages, for example, it brings back wrong values. And on the second round it corrects them. And what happened to me in the past was that I forgot, for example, to do these two rounds and the PDF was wrong, meaning that you may have been clicking on one, you know that the table of contents are hyperlinks, you know. So you may click on one chapter, for example, and it pointed to a wrong page. And they told me this PDF is not working. And I told them it's impossible, I know that they work. Then I understood why and I learned to do these two times. Oh, sorry. Because I use master documents for the guides, and I can tell you that the updates of the PDF and the links of the table of contents on PDFs has been fixed by Michael Stahl about a year ago, and the PDF now is perfectly linked to the right page. Well, I trust you. It may be my fault, but I tell you why. The very first bug I show you, the call out issue, brought me to, and not only that, brought me to use a specific version of LibreOffice under Windows. Because I don't know why, but that version under Windows, which I hate, behaves much better. Let's say that at least I'm able to solve doing that what I told you. Because otherwise, I don't know why, but under Linux, even if I, without the file open, I go there and right click on the document, on the image and send it back, back, back, back, then I save and I prepare the PDF under Linux, doesn't work, I don't know why. That one, that bug. So, and also the scrolling of the pages, and generally speaking the calculation times, I don't know why, but the Windows are a little bit faster. So, that brought me to use a specific version of LibreOffice, maybe old, not maybe surely old, because that works. And I stick to that because otherwise I lost my, now and then I try to use a newer version, but recently I did and I had to go to roll back to the previous version because I had such an issues. So, probably this is solved and I, so this is, wait, this is not any up to date, but by the way, I'm happy to hear this. And by the way, it's not that simple in any case because it's kind of chicken and egg, because you know, the table of contents shifts the contents and so on. I mean, it's not that simple at all. This is not a bug, just a measurement. When I imagine that I work for months on those documents and I prepare chapter one and two and three and whatever, at the end I have to finalize. And this usually takes me two or three days. And one of the reasons is, the reasons are, all the things we have seen right now, all the issues we have seen until now, and also this one. Okay, let me say this and then imagine that these are the timing for opening, saving and rendering the PDF, sorry. And rendering PDF is good one. And on my PC, which is on Intel i5 10th generation, opening the ODM, okay, just the ODM, takes 48 seconds, but the reality is that you can't use this measure under Linux, not under Windows because I was there. You can't use it for real unless it finished its layouting and routines and so on. And this takes two minutes and 10 seconds. That document. Saving takes 32 seconds. Just the ODM, not the single files. But rendering the PDF takes 13 minutes and half, and four more seconds to get back the interface of LibreOffice usable. But imagine that I finalized the PDF and then I realized that there was an error. That's why I usually scroll all the PDF and take note of all the errors, possibly, and I correct them once at all, and then I render again the PDF. But you can imagine that at least three or four times, maybe five times, I have to render because I'm a human, I can not see something. And that hours. And for the more, and then I come to the question. Sometimes it happens that I realize that I made a mistake before the exportation in the rendering finish, and I can't stop it. So I have to wait for it to finish. Yeah, it's frustrating, trust me. And yes, proposal, we were trying to reason about those behavior with Mike, and I know that is not that simple. Let's say that what I wanted to offer, let's say, is a particular figure, not unique, I'm sure, but maybe not that common, because I am a power user of the word processing, in this case, LibreOffice writer, but it used to be a developer. So I don't know the details and all the mechanisms that are behind the behavior of LibreOffice writer, but at least I can guess them and I can understand them if someone like Mike do explain them to me. So I'm not the kind of user who pretends something is impossible to realize or asking for something, you know, the moon. So we, I told him, wouldn't it be possible, for example, to speed up the layouting and maybe the rendering of PDF to split the document in chunks and he said, no, it's not possible. And I stopped him telling, wait, there are the hard page breaks. When you insert the hard page break, somehow you reset, meaning that from that point on, you can consider document kind of another chunk. This is not fully true and he explained me why, but it could be an option to do this way and render in parallel such chunks of the document and then at the end, have a kind of a check if everything is fine or not and have, how to say, a round of verification, validation of the rendering. But in most cases, maybe that nothing breaks and you may have achieved a parallel tasking, a parallel layouting of the document. Please. So it's more a message than a question from Timur. We have a site asked LibreOffice to first discuss a solution to a problem and a tool, Bugzilla, to search if bug is reported or to report it if not. Advanced usage of LibreOffice for complex docs is a good opportunity to see if we have all covered in Bugzilla and that's the best way to start resolving. If we talk about thousands of bugs we uncutured on, it will be hard to follow. This is the first in an after-sale point remark from Bogdan, Gabrielle should use new version as portable. There were improvement in speed in latest versions. Oh, the portable idea is a good one and I could go for it indeed because I may have more than one installed and try them in parallel indeed, which is something that otherwise in Windows, for example, is not that simple. There used to be a tool to SU something, I can't remember. By the way, thank you for the suggestion and about Timur, yes, he is right, but if you go seeking for open bugs about tables in Writer, they're kind of hundreds. And I love you guys, but I have a job, I have a family and I can't waste all my time to chase bugs. I say, I do my best and sometimes I try to do that, but I also have something else to do. At least I try to reporting these kind of things somehow, which is again, I don't want to consider myself this guru of Writer, but I can also bet that not so many people use as Writer in a so heavy way, in a so extended way and in power. And again, being also a former developer, so having kind of formamentis mindset who can somehow understand and try to understand the behavior of the software instead of just complaining. Let's go on. Unless you unlock, well, the ODM files, the master files, they used to treat the child files as a child documents, just like sections, external sections of our document and that's fine. And you can unlock them because by default, when you do that, the ODM file, those sections are locked. So if you try editing somehow, it gives you right protected content, cannot be changed. And it's okay, that's fine. And you may unlock it. What I don't know because I just, I didn't, I have to admit that I didn't think to that. This morning Mike told me, but what about unlocking that sections? Oh, you're right. Indeed, and we tried. So in that way, you may modify the document within the ODM because what I used to do is find an error in the ODM, for example, or in the PDF, then open, since it doesn't allow by default to work on that, allow, sorry, open the right document, the right chapter I have to correct, correct it, save it, update the ODM file, okay, and then render or whatever. So you may unlock the section, but I don't know if it's a good idea, meaning that first of all, I don't know what happens if you save because being within the ODM, if you save, I imagine that the ODM file has been saved. I don't know if the child files are going to be saved as well. Maybe I didn't try because again, Mike told me about this option, let's say to unlock the sections. Anyway, wait, I think the rest of the document needs to, no, save in the original file, how long would it take indeed? Because at the end of the game, I have to also to optimize my times. So it's about to understand if it's better. Given that, let's say it saves the child document, it depends on how much time does it take. Maybe it's quicker, I have to try, I have to try, I'm sorry, I didn't. And then I thought also, going back to the way I use it, imagine that I correct something on chapter two, then I go to the ODM file, I ask it to update, it takes an age to do it, it updates, but why does it take an age? Because it renders again the whole document, which is fine. But I was wondering, could it be possible to not render what's not being affected by my modification? Imagine that I just changed the word. So no new lines, no new paragraphs, no shifting, no modification at all for the rest of our 111 pages. Why relay outing all of them just because a single word changed? So that was the question. Again, I may be wrong, and I'm happy to be corrected. Again, it is a workshop, this is not the holy tables. And so, let's move on. If you have something to add, yes please, thank you. I guess here. Well, it's about the previous slide, where you are showing the export times of PDF. I just remembered that for 7.3, there was an attendee about that regard, and I think it was Lubos improve the sporting of PDF. So probably you should try with 7.3, 7.4. This is 7.4. I didn't mention, but it is 7.412, if I'm not wrong. We may check it, it's the version installed in my laptop right now. I see, okay. But thank you. I love to hear such news. I mean, they are always welcome. You know, imagine me. Again, two days, finalizing manuals are two or three days of work. Then there is the autosave issue. When a document gets huge, kind of I had also single files of 40 megabytes, for example, it made to the whole manual, because of pictures and vectors and so on, and we'll see. When you save, you saw it may take half a minute, one minute, maybe more. Then I have to disable the autosave, because otherwise, or I should, let's say, put it once every 30 minutes, I don't know. But the problem is that when it decide, it's much better to me deciding when to save than being interrupted suddenly, just because it's time to save. Which is fine, I mean, nothing against this function. It's not a fault, it's not a bug. But the problem is that when the file becomes huge, it stops and you can't do anything in the meanwhile. And maybe you were doing something sensitive, kind of you may forgot what you were doing. I mean, I'm quite old, so this kind of things happens to me. So, the idea could be having the autosave function in background. It's not that simple, so with Mike, we've been reasoning how to implement that. And I thought, as like Nick around, let's say a kind of a shadow copy of the document. So, you know, you have this shadow, and you save it in background, and in the meanwhile, I could continue my job, my work on the main file, the current file. Or maybe if this shadow is not that easy to achieve, I don't know, maybe I need an idea to have constantly updated two copy of the files. Everything you do is done, let's say, twice, so you have always have a copy of the file. And when it's time to autosave, then one got frozen, and how to say, written to the disk, and you continue to the other copy. And then you have to find some way to resynchronize them, which could be a third copy, and then you destroy the second, and so on. Then I've been told, to me, I don't think it happened, maybe sometimes I can't remember, but I've been told that it also happens that if a file is really huge, and imagine a PPTX of 400 megabytes, they exist, yes, they exist, and you want to clean them up, because you know and you see that there are elephants. But, oh, I see 40 minutes here on my clock, 40 minutes. By the way, imagine that you want to clean up them. The problem is that you can't do it because the autosave starts, and before the autosave finish its work, another autosave, let's say, call starts, and so you can never work on the document, but I've been told this. Probably I've been experiencing, but not so much because I'm not such a, how to say, intense user of Impress. Then this is about the navigator. I love the navigator. I find that this is the killing feature that distinguish writer from word or other products. I love it, definitely. This is something that I couldn't make without. And I told you that I, just like a maniac, I name every single object just because I have to check that they are all visible at the end. And also because, yes, I'm a nerd and I love to name objects just like when I used to develop, okay? So the suggestion is to have kind of auto naming which is already in place, but something little bit smarter and possibly more configurable with some kind of parameter. For example, my, I have a convention for naming convention where images starts with yeah, IMG obviously. And when I do add caption to the pictures, it creates a frame would be nice. For example, my convention is FRA underscore and the name of the picture. So you will see here FRA, F-R-A underscore, IMG or one time, okay? But it could be interesting to have such a behavior automatic. So every time I add a caption to a picture, for example, it creates a frame and it gives a sensible name. What else? Yeah, sometimes, but I can't reproduce easily. Sometimes it happens that if you double click on an object on the navigator, it brings you to that object, not the opposite, meaning that you click on the object and not always, I can't tell you when, but not always the navigator goes to that object highlighting it, not always. Again, it's not that simple to reproduce because those documents are huge. Oh, I forgot to do this, sorry. Compressing any pictures at once. This is a really nice feature, when you insert a picture, maybe from a camera, it could be kind of 14 megapixel and those 10 megabytes, for example, and especially people who is not used to the word processing, they leave them that way, which is what brings the document to be an elephant. While when you have decided the size on the page of the picture, and that's final, you can use this compressed feature, which is great, really rocks. You just define what's the DPI resolution you want to have, and that's it. What is missing is, as far as I know, an option to do it once for all the images instead of having to do one by one. And this is something that exists in printing press, for example. Then, this is a strange behavior, and given that I don't know how to anonymize vector images, I don't know how to provide a file to test it. What happened is that vector images provided by the customer, probably coming from Autodesk, AutoCAD, and what Mike calls spaghetti vector, because they maybe are too heavy. At the end, what happens is that a single document could be 14 megabytes if I convert them in raster, because I'm forced, or 46 if I leave them as a vector. And in the PDF, it's 250 megabytes versus 60. This is also because I've seen that if you, for example, use the very same vector, many times, it is included many times in the file. While I know that, thanks to Mike, that raster images, they calculate a hash. Wait, okay. Okay, oh, thank you, oh, sorry. They calculate a hash, and so if the software sees that it is the very same picture, it doesn't include many, many times. But with a vector, it is not a fact that we will see later, later, soon. Not mentioning how slow it becomes to scroll the page when you have vectors on the page. It comes a lot, this way, kind of, trust me. Always on I5 and whatever. So the question is, what about a pre-rendering image? I mean, we know that this is a vector. I'm fine, I'm happy. I wanted to have the vector because of the quality. But when I work in the document, you can also replace a raster. And to be honest, I've seen on the file that it creates a raster as well. By the way, we'll see. And, nothing else? Yeah, what? Sorry. Yeah, I understood. So what? Yes, yes. Hello? Works. Works, works. So what exactly do you mean by the vectors? So is it like the SVGs or is it the pictures like that you insert like directly in the writer or like what exactly do you mean? Okay. Usually, for example, let's say SVG, but mostly WMF, which is not a format I love, but I saw that that's the best way to import vector. Meaning that there is no time to show it now if you wish we can do it outside. But by the way, sometimes importing vectors, SVG especially, into writer, some of the objects of the vector becomes blue. Our beloved sun blue. And yes, and thus I found that saving it in WMF and then re-importing that way is much better. So I meant this kind of vector, not what you draw. Yeah. As you can see here, to avoid these, because you have seen I had that 250 megabytes of PDF, I had to solve somehow because the customers were telling me it's unsustainable. I can't deliver 250 megabytes of PDF. Come on. So I tried to find a solution and one of them was to link those images instead of embedding. Because in that way, I have just one file once external and not many times the very same file. So the improvement here could be to, the reason option in the general options to specify if you want link the images to be the path to be relative of absolute and you may choose there, could be an option to have, this is per single file, could be an option to have per single link, or both I mean. And this is a demo document I've just done 10 minutes ago with, oh sorry, yes, this is a vector, vector. And these, I can't remember, maybe this is a raster. By the way, you can see here that those two are repeated. I mean, this is the vector, this is the vector again. And this is a very same picture. To be honest, I did copy and paste. So I did not insert two times the picture, I just did copy and paste. And it is included two times. And then we can see here the path of the file which is relative in this case, okay. One minute. Yep, but I think I've finished. Yes, I did, I have. Then finally, how to anonymize documents. I've seen a specific wiki page if I'm not wrong. By the way, there is a procedure to anonymize documents. And it suggests to use a macro. Can't remember the name, but I have saved it in the directory. And they say you should open your document, select all, you know, control A, copy, create a new document with the macro inside, paste, and then run the macro. But first of all, in this way, you lose the page formatting because you create a new document. So the page style, the default one, not the one you used, and so the layout is gone. Then you have this macro which at the end replace all letters with X, okay. Which is not optimal also because in regular writing, they are not all X. So you may have I, which is thinner. So the occupation of the text on the layout is not the same. So the layout goes away anyway. And at the end, I found that just doing the search and replace with the regular expression slash capital S does exactly the same thing without having to include a macro which prompts you every time you open the document. Hey, there is a marker. Do you wish to run it? So, yeah. Please, what about images? How to anonymize images? That's a problem, open to you. This is my solution, but it's not that convenient. Pack the file, convert and repack the file. So thank you for listening.