 So I'm going to take this simple document, I'm going to open it, and if you open it in a typical PDF viewer, this is on Linux, but it doesn't really matter, versus LibreOfficeDraw and we'll take the current release, I'm going to show you the side-by-side in a second. The blue frames are not part of the rendering, they're just to distinguish the different applications, and also when you open it in LibreOfficeDraw there's a grid, so that's not a bug, that's just fine, but let's have a sense of how these things look, and this is when the system I'm running it on has all the fonts, so this is not about missing fonts. So these are just our bugs, okay, so what problems are we seeing here, there are issues with positioning, issues with spacing, issues with font selection, we're not choosing the right fonts, there's overlap of words, and that's just the immediately visible ones, there's also the fact that we're not reconstituting the document structures, so like Chalme was saying before, there are stretches of text, but not paragraphs one after another, and other structural features. Yes? Can I ask a preliminary question, do you have the same rendering results if you open the PDF from later using the... So they would be extremely similar, yes. So that's one aspect of motivation for even having this discussion, of improving LibreOffice as a PDF editor, but that's even just as a PDF opener, then another aspect of motivation is this kind of archetypal exchange that you might see on the bug tracker on our bugzilla. So some user opens a new bug and says, please improve some aspect of functionality involving how we open PDFs or edit PDFs in LibreOffice, and then someone from the QA says, oh, but you know, we're really not going to be doing that because LibreOffice is not a PDF editor. So if you're asking for better support for editing PDFs, that's out of scope. And then I join the discussion, and then I say, but come on, LibreOffice is a PDF editor, what are you talking about? And the reply is no, so LibreOffice is not a PDF editor, we support importing PDFs, we support editing the structure document that you get from importing the PDFs, and we also support exporting it to a PDF, but it's not, and then my answer is, but this is what makes it a PDF editor, yes, Mike? It's not the same person every time, plus I'm going to present these arguments in a more convincing manner in a second, this is just an archetypal exchange which motivates this talk, so I hope I haven't, well, at the end of the talk you'll tell me whether I've offended you or not, and this goes on because you can repeat this exchange several times. And then the third part of the motivation is, so I'm not really going to give a talk about winning an argument that I've had with someone on a bug page, there's actually an even more substantial motive, but I won't tell you what it is right away, although Italo actually kind of knows if he's still here, but yeah. So this will be our mystery that I hope you will figure out as we progress in our discussion. So let's try to, I hope I can make the fair argument against considering LibreOffice as a PDF editor, and if I haven't, you'll have to excuse me because I'm biased. So PDF is not a format that's intended for editing, it's a final present or presentational format, mostly or essentially, and other than the tags which we heard about an hour ago and are very often or most usually not present, there is no semantic decomposition of the contents of a PDF, we don't have indications of things like sections or paragraphs or breaks to the next line, and we don't have information about a lot of the constructs that we have in our structured documents, so things like tables or columns or document objects and groupings of things, and what's the header, the footer as opposed to the main part of a page, and what's the footnote, etc., etc., and also things like what's a numbering as opposed to just text or how text is aligned, we do have positioning information, but there's nobody saying that that means this text is justified or this text is on the left edge of something, and certainly not things like tabs, and also not really any style information, of course the individual element that's in the PDF is styled, so to speak, but nothing tells us, this styling is the same styling as we have elsewhere in a different run of text, so all that information that we need and an editor has and uses and needs a PDF basically doesn't have, of course the PDFs do carry all the specific information that is required to actually render the document, and like Khaume mentioned, if you use that mostly, and because of this, and if you use a different kind of filter that doesn't extract structure, just renders it the way, like in a presentational sense, say using the PDFium filter, then you can get very good rendering of the PDF inside LibreOffice, but then you can't do anything with it, it's just like it's similar to an image, and so you can either do that or you use the input filter which gives you a bit of the structure, but not a lot of it, so that's one reason we shouldn't consider it as a PDF editor. Another reason is that the import and the export are separate from each other, and they're imperfect, and they don't produce something that is an authentic representation of the actual structure of a PDF. We make it into something else as though we would take sort of a sketch, some artist would take a sketch of what's in the document, then we would work on the sketch, and then we would save it to something else, or like photocopy the sketch, and that means that whatever problems we might have with the import filter and the export filter, they expound each other and make everything, so every such problem becomes even more serious. Another reason no less important is the intentional reason, that is it's not supposed to be a PDF editor. We've never decided that I know of to offer a PDF editor, so PDF editing capability, we got it through having an import filter, but we did not design a draw or writer or any other module for editing PDFs, and actually some contributors think that it's a bad idea to even invest effort in enhancing these capabilities, so maybe it's an inappropriate use of project resources to work on something or to improve something that we never decided we want, and here is a semi quote, not of Mike, that we actually need to do, yes spelling error, we need to manage the expectations of our users who now might be mistakenly believing that the Libre Office is a PDF editor, because it isn't, and if they expect it to be then they will be disappointed. So like I said, I hope this was, I related the argument fairly, but definitely check out the relevant bugs if you want to read it, read the actual proponents of this position, which is not mine. And now, so I want to think together with you about this question and put it in perspective, through several prisms, one of them is functional, that is how relevant are we or how relevant is Libre Office to the use cases that people do PDF editing in, then also something more formal, so what is it, should we consider Libre Office as a program which meets the definition of an editor, and finally mention public perception, so what does the public think or what does the internet think about Libre Office as a PDF editor? I'll start with the functional part, which is perhaps the most educational, so let's talk about cases in which a typical office suite user or typical user would want to edit a PDF document. Of course this is my opinion, perception, experience, and I've not conducted like a proper survey or anything like that, but I think you'll probably agree with me that if these, and I have like four or five cases, there might be another one that you might mention or you might claim that one of them is not as significant, but here is the way I see it, so perhaps the most common or very common use case is getting a form in the PDF format, so some organization could be the government, could be the TDF, could be a company, has a procedure in place which requires us filling a form and the form is not available interactively, it's available statically as a rendered PDF document, and the document is done, so it doesn't have form controls which the PDF format supports, it's just a rendered, like a rendered piece of paper or the file equivalent of a piece of paper and maybe it's even a physical form that somebody scanned, although then you might want to extract the image but never mind that, so that's the PDF that we got and it's very, oh missing font, yeah, so the font here is supposed to be different, anyway, so we've gotten this file, we need to submit the form, so we need to edit the PDF and of course if it's a smart PDF with form controls then we wouldn't use an office suite, we'd just use a PDF viewer, so what kind of editing would we do in this case, we would need to add text within boxes or on lines, we might need to add like individual digits when you want to fill out your ID number or phone number, we might need to insert a signature or another image from somewhere, we might need to be able to check a box or mark it with an X or fill it with black or some kind of color, we may want to strike out some text, we may want to circle some texts or maybe we might have a free form area in which you want to insert all sorts of contents, so that's one use case and I claim that it's quite likely for a user to expect to do this kind of work in LibreOffice and again now we're in the part of the presentation where it's my opinion, so if you disagree, you're of course entitled too. A second use case is when we want to do markup or add commentary on the PDF, so somebody sent us a file, somebody sent me their slides for this conference and I want to provide all sorts of comments but I don't want to edit their presentation, so I want to add some text, I want to mark certain areas, there is some overlap with the previous use case. Doing this is also, this is not extremely common but it also happens occasionally and it's less clear, we want to use a proper editor for this because the PDF specific tools whether it's commercial like Acrobat or even PDF viewers or like viewers plus often have some features for doing annotation, so it's not clear that people would turn to something like an office suite for doing for this use case. Another use case is that you have a document that is supposed to be the final form but you notice that you need to make some changes to it, it's your document or a document of say a colleague of yours but maybe there's a typo in it or a phrase that's missing or the graphics is misplaced or misaligned or something is missing and you can't use the original document, so you can't take the original document, make the change, reproduce the PDF, maybe you don't have access to it or you don't have access to some of the fonts it uses or other artifacts so you have no choice but to take the PDF and make an edit to that itself. This is another very common case of PDF editing and this is something we might expect to want to do in LibreOffice, if the import filter were good enough then making this sort of changes, possibly we would consider an office suite or something like draw a writer for doing. And another use case is when you want to combine pages, so we have maybe some existing PDFs or we have content from some of the LibreOffice modules or other applications and maybe we have some images and you want to combine all these things as different pages within a single PDF, maybe you want to concatenate them and maybe you want to cut out some of the pages and include other pages or you want to do some rearrangement and when does this happen? So if you're writing a letter and you want to add attachments or you've worked on some complex document in several applications, maybe you have a page in calc and a page in writer and an image that you worked on somewhere and maybe you it's the form filling case but you actually don't need all the pages so you want to drop some of the pages etc. And this is another use case that's extremely common and here it's not clear to me whether an office suite is where we would want to do this because this is less about the actual editing of the content and more about the arrangement and again there are like custom apps or utilities which are better tailored to do this but if LibreOffice were really good in importing content then maybe it would still make sense so not entirely clear and the final use case is when okay so I'm going to skip one of these because we're short on time and the final use case and the least relevant for us is when you want to make precision alterations so you want to make changes to the specific structure represented in a PDF and even though you're changing things the rest of the contact must be maintained pristinely and precisely and this is if you're doing some kind of professional type setting then you might need to be able to do this maybe you want to do some programmatic generation of documents or other similar cases and this is a niche use case which is really not fitting the way an office suite or specifically LibreOffice works with documents because the kind of structure like the native structure of the PDF is just not something that we work with so we definitely would not want to work with LibreOffice on something like that so functionally we have some use cases that are more relevant and some use cases that are less relevant and this is independently of how well we can import PDFs so that was the functional part and now for the formal part then so if we look at the dictionary and we look for the definition of an editor then here are a few definitions of what an editor is and if I combine these three definitions which you really don't have to memorize there are the parts of the definition that are important so it needs to be computer program that works on data and it needs to be able to create or change the data or correct the data or read what's stored and make changes to that and so technically if we take these parts of the definition then yeah we are a computer program we can create PDFs we can open PDFs although not perfectly and we can make changes or corrections but not really to the original PDF but to something that's not the PDF it's a it's a it's the result of a somewhat destructive process and plus we have so many issues and the things are introduced by the the import filters that it's it's hard to to say we really can probably make these changes so actually my argument for meeting the formal definition is not such a strong argument for for Libre Office as a PDF editor so I regretfully have to concede that that part of the argument or the discussion but now now is the more the more interesting part so public perception so first let's do let's search on YouTube I searched for that and I did this a few days ago as I was preparing the presentation I searched for how to edit a PDF and the top result is how to edit a PDF free by this guy and here is the like the image that you get this from 2020 so we've approved a bit maybe and here's what he suggests first he suggests Microsoft Word because you supposedly have already paid for it or gotten it somehow so you can use that then he suggests this web application which is it's mixed free and like it's a freemium kind of arrangement but there are some free features there and the third opportunity is Libre Office draw and these are his three suggestions so that's like the top result and the by far the most the most popular now what happens if we ask chat gpt and by the way you can ask chat gpt anonymously if you follow the link that's in my slides how do I edit a PDF so you get a bunch of texts it's you know very conversational and it tells us the few methods you can consider to edit PDFs and these are a few applications some of these are online and we're not on here at all so but google docs is and we're not that that bad compared to google docs and I don't know about preview so I asked it another question because that's all commercial software so how to edit a PDF with FOS and now it said this to add the PDF with FOS you can follow these steps install an editor and then use the editor so which editors does it just suggest we're number one so Microsoft somehow likes us I'm a pdfsem inkscape and scribbus and we won't go into comparison of all of these but this but we're the like the top FOS suggestion from from chat gpt and and then let's also do like a plain web search on how to edit a PDF file and we'll use unfortunately I'm using google because it's the most popular but you should really not use google for anything at all because they spy on you so let's not use google but most people do that so what does what does google suggest so google suggests several things Microsoft Word and and some online apps and LibreOff is draw so it's it's the main apps it suggests are office weeks our main competitor and us um uh so I know we're a bit short on time but I started late so I'm going to steal some of the some of the time from the next talk so you would you thought that I might now put these perspective together and and reach a conclusion right so I'm not going to do that at all um because think of what we've seen here LibreOffice is perceived as one of the top if not the top PDF editor that's free and open source software so what which means that people find it and download it and use it for non-niche and more common PDF editing scenarios and and that means that even though we've never made the decision for this to be the case we sort of have an an extra variant of a module so and especially in in in draw so it can be either impressed for presentations draw for drawings or the PDF editor module which is draw for drawings and effectively we have that and and it doesn't really matter whether LibreOffice is a PDF editor or not what matters is that at least in terms of FOSS or free software it is the PDF editor and that's that's a that's a public fact but what why does this matter it matters because um uh uh think about some of the challenges that we have when we're promoting LibreOffice so we need to make people aware of of of the suite of LibreOffice um we need to give people reasons to prefer LibreOffice to Microsoft Office we need to give people reasons or motivations to install to download and install LibreOffice even if they haven't necessarily chosen Microsoft Office over us and we also want to to build a good technical reputation for our office suite and the fact is that that all of these are massively helped by those PDF editing use cases because millions of people um are may notice and probably are noticing LibreOffice not as an office suite but as a way to edit PDFs even if they don't care at all about switching from Microsoft Office or from anything else and um uh and that's a that's a consideration when choosing which office suite to use because you have at least on in one of the features if we could be better than Microsoft Office then then why not another um uh point in favor of Libre and uh and then also if you've already installed this thing because you wanted to work on a PDF you wanted to edit a PDF then maybe you might just you know try it for other documents and it's since it's already on your system and finally if you know if you've been impressed with how we're handling PDFs and it's such a nice such a nice piece of software maybe it's it's good overall and but of course and and I'll close with this that we're not even though that's a potential boon for us we're not really there yet because um uh um because there isn't that much awareness um that that we we're promoting of of LibreOffice as a PDF editor we're not the top contender if you remove the the the free software qualification um uh and Microsoft Office opens PDFs with a similar or better quality to us so that's embarrassing let let's not be in this situation we're not encouraging people to download LibreOffice as an editor of PDFs and finally the impression that people might get is that I'm I'm so annoyed by so these PDF import bugs that I'm seeing when I try to use um uh draw as the PDF editor that I doubt whether the rest of LibreOffice is really all that good so the the boons are also like detriments if we don't don't um handle this or or invest effort here and so with not very significant but not terribly extensive coding effort we can convert all of these detriments to the benefits from the previous slide and with this I will end my presentation and thank you very much