If you ask me, who cares how something looks so long as it works? And speaking as someone who's done at least some limited user interface design... It's tough. There are so many little things that the average user is never even aware of that the developer has to account for it's not even funny. For example, most people take it for granted that if a textbox is disabled it has a gray background and you can't click inside. The reality is that the developer has to explicitly set the background color if the control is disabled and then remember to change it to a different color if it's enabled again. In theory the developer could set the background color to whatever he/she wanted, but typically we follow convention which is to gray it out. And just because you disabled some textbox doesn't mean that the text you had typed into it will go away, you have to do that explicitly. The developer has to do everything right down to setting the tab order, any and all keyboard shortcuts, even trapping specific keys (like the enter key) to trigger submitting a form if you hit enter on a text box.
It's also a not insignificant undertaking to take an established product like LibreOffice and completely rip out the existing UI and redo it. I develop and maintain a small set of VBA macros for work. Recently I decided to redesign the entire interface more or less from top to bottom because the previous method was really getting to be a lot of overhead to maintain. I had a series of radio buttons and it was getting increasingly difficult to maintain the proper state of each radio button. I had another form where I made the mistake of trying to reuse the same textbox for multiple functions and it just got to be really confusing in my code because of the name of the control often times didn't match up with what task I was performing at the time. So I would end up creating a second variable just to provide a more convenient name for the textbox
Even accounting for all the experience I had gained from developing the previous version, so having some idea of what not to do, and the fact that I had long since broken out the control logic from the UI logic, it still probably took me around 2 full days to just get the UI logic in place. This is just for a macro that has people make a couple selections in the form of checkboxes and a list select control. If you've ever worked with VBA you know how limited your options are for UI controls. My macros are only maybe 500 lines of code all told, so miniscule in the face of LibreOffice. I'm also nowhere near as skilled a developer as many of the LibreOffice developers are, but redesigning the UI on LibreOffice would be like a 6 month long undertaking, at least, and you'd have to put a lot of other development work on hold. Plus there would undoubtedly be bugs that escape testing. After around 100 or so lines of code, it's pretty much a given that there will be bugs that will get missed in testing no matter how much testing you do.
LibreOffice may be kind of clunky looking, but it works wonderfully where it counts. Form should follow function, not the other way around.