LibreOffice Gets An OpenGL Rendering Back-End

[intellivision]

Ah, the joys of the Apache license (or even the LO MPL, which would likely allow that!)

Actually, the MPL is essentially the same as the LGPL with a few changes:
- Static linking is allowed
- It can be made compatible with any copyleft open source license the author chooses through the 'secondary licensing' provision e.g. GPLv2 only, AGPLv3+, CDDL if the license is ammended etc.
- The 'further restrictions' section has a reduced scope, IIRC you just have to provide the source in an easy to access way.

[brk0_0]

LibreOffice needs more than tablets. It should make using Linux on regular workstations easier and cheaper.

Nowadays, LibreOffice is hard to use and has a confusing interface. It also has a bad end-user documentation, in the sense that a regular user has no means to enter the official website and follow simple and visual tutorials to get things done. Nobody is going to read the current docs to know how to make a new kind of graphic. What about official video tutorials? What about easy-to-follow visual tutorials? Whole new site (I think that is on the way, looking at the new home page)?

Still, the interface sucks a lot. Even MSOffice is better than LOffice and OOffice right now. It's simpler, more straight-foward, more eye-pleasing and more intuitive than open alternatives. We seem to be stuck on 90s to new users. LOffice should realise that they are NOT the only Office Suite on the open-source side, and be brave enough to turn the table. We need a new interface. A brand new. With new (but intuitive) visual, the same (if not more) features, that hiddes less things from user and is less cluttered at the same time. This is the only chance we have to gain maket-share.

We need enterprises' eyes. We need to show them that we are not behind MS, that we have not only a cheaper Office Suite, but a better one. Big enterprises have money to spend on MS things if they want. And they matter to us.

We need to realise that we are making the same work twice (OOffice and LOffice) and make them different: one remais more classic, following the same dev logic (implement the new .doc very badly -> debbug -> fix some compatibility issues -> 'optimize' -> clean), the other would be focused on bringing new stuff to open source (like hardware accelerations, new interface, online features, more power-features that MSOffice doesn't have). I personally prefer LOffice to be the second one.

But, since I am just an user, not a collaborator, these are only toughts. Not even suggestions.

[Saverios]

Speaking for myself, I tend to agree that OpenGL rendering seems somewhat superfluous at this point. I was very excited when LibreOffice 4 came out, only to experience quite a lot of not consistently reproducible bugs (especially in Writer). I always update to the latest version, but it's been what, more than a year now and Writer's reliability, either in terms of stability or in terms of consistent document formatting has been quite disappointing, especially when it comes to numbering, bullets, indents, which I believe is really basic stuff. Every time I try to change the indent of bullets or numbering I think the whole PC will explode. And, yes, for some reason a lot of clicks are ignored. I did my PhD thesis in Writer, only to find out that every other day I reopened the document pictures had been moved around, and sometimes even lost due to loading errors, regardless of the format I saved the file in.

I am not saying this to devalue the hard work of LO developers, far from it. I think overall they have done a great job and I consider the whole suite to be a viable alternative to proprietary products (which, of course, have their own sets of issues), at least for the majority of users. But I fail to see how faster/better/trendier graphics rendering is going to improve reliability and user experience. And I don't think users expect their word processors to be very exciting.

For the time being, if I have to write something that involves anything else than plain text, I am switching to LaTex. It may have a steep learning curve and sometimes simple things are a bitch to get right, but at least I know that when I re-open my document, everything will be there as I left it.

[brk0_0]

Speaking for myself, I tend to agree that OpenGL rendering seems somewhat superfluous at this point. I was very excited when LibreOffice 4 came out, only to experience quite a lot of not consistently reproducible bugs (especially in Writer). I always update to the latest version, but it's been what, more than a year now and Writer's reliability, either in terms of stability or in terms of consistent document formatting has been quite disappointing, especially when it comes to numbering, bullets, indents, which I believe is really basic stuff. Every time I try to change the indent of bullets or numbering I think the whole PC will explode. And, yes, for some reason a lot of clicks are ignored. I did my PhD thesis in Writer, only to find out that every other day I reopened the document pictures had been moved around, and sometimes even lost due to loading errors, regardless of the format I saved the file in.

Agreed. Office Suites should be reliable, powerful, fast and easy. LOffice, today, is not the best choice in none of these aspects. Other programs used in office works are far better, let's just look at Google Docs, MSOffice, LaTeX...Users should see LibreOffice not as a cheap alternative, but one that matters when doing your job. One that helps, that do things others don't. But LOffice just isn't this program yet.

I am not saying this to devalue the hard work of LO developers, far from it. I think overall they have done a great job and I consider the whole suite to be a viable alternative to proprietary products (which, of course, have their own sets of issues), at least for the majority of users. But I fail to see how faster/better/trendier graphics rendering is going to improve reliability and user experience. And I don't think users expect their word processors to be very exciting.

Not VERY exciting, but at the end of the day, they shouldn't blame the text editor (or any other tool) for not getting the work done. This is a good tool, the one
that only helps you, not get on your way.

[GreatEmerald]

Still, the interface sucks a lot. Even MSOffice is better than LOffice and OOffice right now. It's simpler, more straight-foward, more eye-pleasing and more intuitive than open alternatives. We seem to be stuck on 90s to new users. LOffice should realise that they are NOT the only Office Suite on the open-source side, and be brave enough to turn the table. We need a new interface. A brand new. With new (but intuitive) visual, the same (if not more) features, that hiddes less things from user and is less cluttered at the same time. This is the only chance we have to gain maket-share.

No. LibreOffice needs none of that. That is the job for Calligra, and they have been doing quite well at it lately. LibreOffice is fine with the menu interface that works exactly as people expect it to work ? like MS Office 2003, and not the mess that came with later MS Office versions. The fact that LibreOffice doesn't take revolutionary steps is exactly its strength, because those who use it are sure to not have things changed around for no real reason, when the previous scheme worked just fine. "LibreOffice Human Interface" would be the biggest disaster you could ever think of, as it would alienate the existing user base, prevent people migrating from XP to adopt LibreOffice, and be a useless learning curve for those who want to use it despite the fact.

[liam]

Did no one notice the part about gl text rendering? I'm assuming this is referring to behdad's work doing the same with egl.
That is potentially a really big differentiator between Linux and any other OS.
Completely accelerated text rendering has alluded everyone, apparently, and behdad seems to have cracked it.

[BlackStar]

Did no one notice the part about gl text rendering? I'm assuming this is referring to behdad's work doing the same with egl.
That is potentially a really big differentiator between Linux and any other OS.
Completely accelerated text rendering has alluded everyone, apparently, and behdad seems to have cracked it.

Haha, that's a good one. Windows has had fully accelerated text rendering since 2009: DirectWrite. It shipped with Win7 and uses the GPU exclusively.

LibreOffice should *really* focus on its user interface, it's so ugly that I prefer to use Google Docs instead.

[brk0_0]

No. LibreOffice needs none of that. That is the job for Calligra, and they have been doing quite well at it lately. LibreOffice is fine with the menu interface that works exactly as people expect it to work ? like MS Office 2003, and not the mess that came with later MS Office versions. The fact that LibreOffice doesn't take revolutionary steps is exactly its strength, because those who use it are sure to not have things changed around for no real reason, when the previous scheme worked just fine. "LibreOffice Human Interface" would be the biggest disaster you could ever think of, as it would alienate the existing user base, prevent people migrating from XP to adopt LibreOffice, and be a useless learning curve for those who want to use it despite the fact.

As taking this kind of revolutionary step is the job of Calligra, being conservative is also job of OpenOffice. The problem is: Calligra is a KDE application, and have only a beta Windows/OSX version, and no mindshare at all. This is a problem.

Changing is not evil as you put it. And changing does not mean change just like MSOffice did (even if I like the new look, I know it have many problems). Look at Blender for example. It's interface changed, the users had to adapt and, at the end of day, we have an easier-to-use software that is still very powerful. Unlike you said, people look at LOffice/OOffice interfaces and find it very cluttered and actually very different from MSOffice 2003 or older. So, again, their are at a limbo: they are not a drop-in from MSOffice 2003 and doesn't have a nice, easy to use interface.

If there is a learning curve either way, what would the average user prefer at the end of it? Something that works just like MSOffice 2003 (more than 10 years old) or one that's actually better and easier to use?

[Krysto]

Now they need to finally update that early 2000 UI and we're set.

[GreatEmerald]

As taking this kind of revolutionary step is the job of Calligra, being conservative is also job of OpenOffice. The problem is: Calligra is a KDE application, and have only a beta Windows/OSX version, and no mindshare at all. This is a problem.

Changing is not evil as you put it. And changing does not mean change just like MSOffice did (even if I like the new look, I know it have many problems). Look at Blender for example. It's interface changed, the users had to adapt and, at the end of day, we have an easier-to-use software that is still very powerful. Unlike you said, people look at LOffice/OOffice interfaces and find it very cluttered and actually very different from MSOffice 2003 or older. So, again, their are at a limbo: they are not a drop-in from MSOffice 2003 and doesn't have a nice, easy to use interface.

If there is a learning curve either way, what would the average user prefer at the end of it? Something that works just like MSOffice 2003 (more than 10 years old) or one that's actually better and easier to use?

Blender isn't a fair comparison, because it had a really mind-boggling counter-intuitive interface (dragging the top to reveal options, for example). LibreOffice doesn't have such problems, and generally has the same things in the same places as in MS Office 2003. If there is a learning curve coming from MS Office 2003, I am yet to see one (one of the people I maintain an office computer for didn't even notice anything when, during the update from XP to Win8, MS Office 2003 was replaced with LibreOffice).

As for new users, they might prefer something that is better and easier to use, but what is better and easier to use for everyone? If you do sweeping changes like GNOME is doing, there would be a lot of annoyed people. Like I said, better leave those things to Calligra, see what they come up, and if it works for everyone, perhaps slowly migrate towards that. The fact that Calligra is new and not everything is finished yet isn't much of a problem, all projects have to start somewhere. They've been making good progress so far.