 So, let me share again. Window. Right on, now we see. That's my screen. That's how it looks when I share my screen. So, as a fallback, you can send me the presentation and I click through it. Or this one is a draft. And maybe why Maria is trying to find a solution. I can entertain with a little bit introduction. Maria joins the team a couple of months ago and we were wondering what she can do. She was interested in Calc and thinking about to aggregate a little bit, say, tickets on Buxilla to see what people want or how people want to enhance Calc. This is, well, we do have this kind of aggregation that is called a meta ticket. And it is still not simple to guess what a user really want, what they need and how they scenario. So, we came up with the idea to run a survey similar to what we did from the design team a couple of years ago for draw. And this was a very interesting survey as it revealed a large number of options. Options, what users do this draw. And we were a little bit in the hope for a similar result. So, we run this one. We run a survey on our survey platform with three pages. The analysis was done with R for the quantitative data and we did the quantitative data based on quantitative is the idea that you are not really aware what you want from an application. So, if you're getting asked what you like on jitzy, how should jitzy improve? It is not so easy to answer. But if you ask, would it be nice to have a better presentation mode? Maybe people are, it's easier for people to reply. So, we did this combination of questions and now it's Maria's show. So, I think it's better speaking than me. Sure. So, I just send you the like updated file because I added a couple of slides. I shared it with you with telegram. I'm still having problems with, it does not let me share. When I select the window, it's the message shows failed. Yeah, so I just added a couple of slides. So, we started with a summary. So, we surveyed 1,230 people to learn more about the users of Calc as Haika mentioned. The results were not surprising, but the survey did give us an idea of the most significant pain points of Calc users and where efforts to improve the software should be allocated. As a part of the survey, we've administered a system usability scale or SAS questionnaire, which is a standardized questionnaire used to measure usability of a software. The SAS score generally ranges from 0 to 100 and the overall score for Calc was 71.4. This is considered as an above average score. Top reported pain points for server participants included usability, especially around pivot tables and charts. People mentioned that these features were more difficult to use or limited compared to Excel. Another pain point related to usability was feature discoverability. Participants reported that it is sometimes challenging to find features they need. Survey results also indicated that Calc users want to see a better compatibility with Excel and one way to address this issue is to introduce new functions such as XLOOKUP. Another reported pain point was speed and performance, especially while handling large data sets as well as stability of the software and UI design. So we can go to the next slide though. So here we have an overview of characteristics and demographics of our respondents. It is important to note that the survey was advertised on subreddits dedicated to Linux, so our data is queued towards users of Linux. I think the other graphs are pretty self-explanatory so I will not go into more detail to describe them. So yeah, slide four. So not surprisingly, most participants reported to employ Calc for personal and business use and the majority of our participants reported to use Calc almost every day or rather frequently. Okay. Also not surprisingly, Excel and Google Sheets were reported as some of the most popular alternative data-related softwares. As you can see on this graph, 30% of server participants reported to use R as well and another major category of 10% was Python. And on the right, you can see a bigger list of different other tools that people use, but the number of people that use them is relatively low compared to Excel, Google Sheets, R and Python. So these graphs show the most, I guess, common data sizes reported participants of our survey and the most common file sizes as we can see from these graphs is under five megabytes or under 1,000 rows of data. However, some people report using Calc with rather large data sets such as over a million rows or about 1,000 megabytes. So the most popular reported use cases for Calc are data storage and tracking, finance and accounting, as well as data analysis and calculations. So I just named the top three features on this graph. Server participants reported to use Calc for personal finances or budgeting as well as business accounting. Some of the data people reported to track include health-related data such as blood pressure measurements, expenses, time and inventory. We can go to the next slide. So in terms of formulas and calculations, the most widely used functions are mathematical, logical and date functions. Array and database functions are less common. So this graph shows how important certain Calc features are to respondents and the gray bars show the percentage of respondents that are not familiar with a particular feature. For example, quite a few server respondents are not familiar with scenarios or goal-seek. Not surprisingly, formulas and calculations are reported as some of the most important features. As you can see from this graph, and Heiko points this out. What's interesting is that some features such as pivot tables and data validity that have been not familiar to some users are rated as quite important by users that are familiar with that particular feature. Now we can go to the next slide. To me, now we get to probably the most interesting part. So we get to the system usability scale or SAS questionnaire. And SAS, as I mentioned in the introduction, is a standardized questionnaire used to measure usability of a software. Our survey results indicate that SAS score for Calc is quite high. It is 71.4, which is considered above average as indicated by other similar studies. For example, the average SAS score from 500 studies conducted by Jeff Sauer, who is one of the prominent usability practitioners, was 68. So having a score of 71.4 is quite good. However, here it is important to note that although the overall SAS score for Calc was 71, the average SAS score for Calc beginners was slightly lower. It was 68. And this score or result is consistent with some feedback we've seen from the survey. We've noticed that several participants reported that Calc is a little bit difficult to use when transitioning from Excel. So if a beginner that is used to Excel tries to use LibreOffice, they may face some difficulties. We can go to the next slide. So we asked participants, what do you expect from Calc in the future? Most answers evolved around improvements such as better compatibility with Excel, usability improvements, especially around pivot tables and charts, design as well as speed and performance, especially when working with large data sets. And this graph lists other additional features or improvements that people mentioned. And I just named the top ones. So this is the first part, and then this is the second part. So like learnability was also one of them. And learnability is tied with usability and free to discoverability. I think we should mention that what 8% means that is a bit unclear from the scale. This category is an extraction from an open question. People respond. I think LibreOffice should develop on the usability. It should be faster and I would appreciate if pivot tables were easier to handle. This person would receive an entry for pivot table, one for usability and maybe one for performance. And from the total number of subjects, the percentage reported here is the number of people, the total number from total number, who are reporting in categories that we extract from the open question post hoc. People say quite a lot if you ask them freely. The challenge is to aggregate everything into meaningful categories. That's something that Maria did with great effort and took a lot of time and this actual result here. Sorry for interrupting you. No, no, it was very good addition. Thank you. So I think now we can go to the next slide. So this slide again shows the list of improvements that people suggested. And this particular slide will include, as Haiko will go further, some quotes from participants that we extracted from the survey. So for example, in terms of keyboard shortcuts, one participant reported that copying, cutting, moving of cells should be simplified. And the ribbon and toolbar interface should also be improved with most used functions. In terms of health documentation, participants reported that a lot of information, especially for non-trivial things, is not detailed enough or is obsolete. And additional formatting should not fragment when rows added. And removing duplicates from a data set is much more cumbersome than in Excel. Scrolling provide an option to, yeah, provide, yeah. So improved function wizard. Going a little bit faster. Perhaps better than I speak. It's a couple of examples. You can read it or glance of it to get an impression how people report here when they get asked what they expect from the future. And that's what they report and how we collect the bigger categories out of it. They correct the styles within a cell. People want a lot of formatting within a cell. That's quite, well, not really surprising, but Calc is used, also used as a kind of a layouting tool for tabular data. Speed is quite often reported here. It's not sorted. The response is just to point out from, I was going through the reports and these are just examples that came up here. Well, yes, conditional formatting is something that both us people quite a lot and they talk about pivot table, precise response. Solver needs to store data and they ask for wizards. The chart wizard is something that can be improved. Of course, a lot can be improved. The documentation is something that needs some love. Speed again. Here's an example of this size and that's all, Maria, your turn. Yeah, so basically these were just examples based on what people reported and basically most improvements can fall under the six categories and we can go to the next slide. So, yeah, so based on the comments we've seen, here's some recommendations for usability improvements. So one would be to simplify UI to make the software easier to use for beginners or still offer more complex UI with advanced functionality on demand for advanced and intermediate users. Improve vulnerability by educating users about the workflow with the wizards, documentation and examples. I think this is especially important for pivot tables, formulas, charts and Python macros as these seem to be the features people have most problems with. It also could be that some of these are most commonly used and that's why we saw more comments about these features. Also, improve feature discoverability. One good suggestion we've seen regarding these issues is a global search functionality. And lastly, improve compatibility with Excel. For example, introduce Excel functions that may be currently missing in columns such as AxleCup function. So click to exit. Thanks, Maria. I think we have time for a few questions. Yes, indeed. Thanks a lot. And thanks for the work. I'm not seeing any questions in the chat room, but I guess we could take them live here in Jitsi if anyone has something to say. Well, if not, then we'll take a short break before the next talk and it's the final one in this room. Thanks for listening. We will report about the results in the design block and of course share the presentation. So maybe if more questions arise, the design people are available in Telegram, in Matrix, in IRC and everywhere in the world. Just ask. And now as we close the questions, of course we have a question. It's from Rafael Lima. And he's asking, there's a lot of work to be done in ImprovingCog. Could it be done with some sort of grants to fund these improvements? And yes, obviously some of this could be done by money taken from the Document Foundation budget. And this is decided by the Engineering Steering Committee. So anyone can propose budget items and they will be evaluated by the committee and then presented to the board for tendering. So everything is possible. And it's a question about prioritisation from Shinji. I think, Maria, you wanted to put these results together with what we have in the bug tracker, the meta tickets and generate some kind of a preference list, priority list. Is this a plan? Yeah, that was the idea for the survey to see where people report most problems and then take the results and prioritise the bug list. So for example, if people struggle with pivot tables, maybe the first thing to do is to look at bugs related to pivot tables or charts, for example. I got a question in private by Flywire. A reasonable sample of users, of course not. That's what we tried to show in the first picture here. Sharing again this one. It would be a great world if the majority of users is using Linux as operating systems, green one. And Windows is down to let's say 30%. I'm not against Windows, by the way. No, just kidding. It shows that our sample is not representative for the whole users of the profits. But that's a problem that we always have. We cannot speak to the real user. We have to use some social media, some communication channels where the usual suspects are always addressed. And that's the nerds, people who are more advanced users, more interested in this technical stuff. And that's something we see quite often. But there's no other opportunity to get a large number of users other than making the application to call home, unfortunately. I also think that if we would do like a representative sample, we would need to compensate users, and for that we would need additional budget. Slide nine. That's a Likert scale. Likert scale means you ask with a number from one to five between the low or the minor anchor and the high value. In this case, it was how important is a feature to you, not important at all, one or very important. And the subjects will ask, using formulas and calculation, how important is it? And most people say, absolutely, I use formulas every day. Of course, you do so. When it comes to pivot table, many users who report on this question reply with either very important, the highest category, or the one below five or four. That's the two greenish, blueish, whatever color it is. But at the same time, you have a good number of people who choose the extra category of, I'm not familiar with the feature. I have never heard about pivot table. I don't know what it is. So I cannot say if it's important to me or not. That means here, you cannot read it. Probably this 20% of users of subjects reported no, I don't know, pivot table. That's quite a lot. The percentage is deducted here from 100. So when it comes to scenarios, 50% of users of subjects, sorry, do not know what a scenario is. 20% or so when it comes to pivot table. But the remaining 80%, they said this absolutely, that's important. Similar over and post graph. It is a beautiful graph. It took me a week to produce it. Just kidding. Why not? We can do a similar survey with Impress again. Perhaps although it's a writer, of course, why not for all modules answering such a survey and take some time and we cannot occupy people for half an hour with, yes, one by one. That's the right way as you suggest on IRC. And after some time, I think we can continue this way of questioning to the user, to asking users how they like the application, where they see the most relevant additions. Yeah. Could you address a different audience? I cannot. This survey is based on reports from 1,200 people and I don't have access to that many average users. I cannot just ask as a supermarket people, how do you like the pivot table feature? I cannot make the application call home. I cannot run a survey starting from the application. So a bit difficult. It would be nice. And I'm afraid Ilmarie is cutting off the session now for the next speaker. Yeah, actually we are still waiting for the next speaker to enter the room. You want me to entertain the audience? Yes, some freestyle rap is always welcome. French. Yeah, French, preferably. But yeah, because the next talk is not pre-recorded either. So it will be live. And I think Italo is maybe having some technical issues. Flywire asked me on IRC, maybe service could be included with new releases. That's what I tried to answer. I believe if you use Lipper Office, you are interested in this free and open idea of software. And you explicitly do not want software to generate data for some weird organization behind. If you know that TDF are the good guys, yes, maybe you want it. But basically you use an application to have your privacy and to not get in contact somehow. Sure, we can have the discussion. And we had it in the past. It was about user metrics. To get real data from how people access unicomments. It's a functionality behind every command. Or everything, what you click in the UI, that's a unicomand and we could, in theory, collect this information and send it to some service that collects all the interaction from users. Well, there was a tender a couple of years ago to improve this. Meanwhile, we removed the functionality from the program to store the usage data. It's hard to imagine for me that we can restart this effort and get real user data. And if it comes to surveys, nothing is unimpossible. Of course, we could include some link to a survey in a release. But do we want it? Yes, another info bar. Well, it's a good question and I cannot answer. I would love to get better data, more data. It's always fun to analyze. It was not so much fun with Maria. She did the analysis of open questions. And I think you can imagine that readings through all the different reports from users is not always perfect English. And to aggregate everything into some categories, bring it together and collect the user knowledge into something that can be reported. That is quite a lot of work. So, big challenge and a lot of something that the kudos applause should go to.