Announcement

Collapse
No announcement yet.

KDE's Calligra KOffice Fork Sees Its First Release

Collapse
X
  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    Originally posted by Awesomeness View Post
    Yes, they can. And it's much more complicated than to extend an easy maintainable code base like Calligra’s. LO is so immensely huge and messy, when Nokia was still on the MeeGo train, it chose to adopt Calligra for its Office Viewer app on the N9, even though at that time Calligra was still in alpha stage. LO is nowhere as feasible for that use case and on LO work quite a lot full-time employees whereas Calligra is a real community project.
    just to make myself clear its not a LO vs Caligra thing.

    i just want 1 (one) suite that is feature complete, cross platform (including mobile) and can compete with the rest.

    i feel the same way for the million multimedia players, web browsers, email clients etc

    Comment


    • #17
      Originally posted by 89c51 View Post
      just to make myself clear its not a LO vs Caligra thing.

      i just want 1 (one) suite that is feature complete, cross platform (including mobile) and can compete with the rest.

      i feel the same way for the million multimedia players, web browsers, email clients etc
      Different people, ideas, preferences, goals, skills, programming languages, desktop environments - it's not easy to convince someone to our preferences.

      Comment


      • #18
        Originally posted by 89c51 View Post
        just to make myself clear its not a LO vs Caligra thing.

        i just want 1 (one) suite that is feature complete, cross platform (including mobile) and can compete with the rest.

        i feel the same way for the million multimedia players, web browsers, email clients etc
        Just wait for awhile. LO is working on a mobile version withstand the gtk Broadway backend. However, don't expect it in less than a year.
        I don't see the point of a KDE office unless it is simply a toolkit exercise. Building an office suite is crazy amounts of work. Having a code base that is 20 years old isn't necessarily a bad thing if, as LO is doing, unused functions are continuously searched for and pruned. LO has a very large and active community, and has excellent support for the major formats (which, I would imagine is the hardest thing to achieve), but even they don't have ideal format support. To my knowledge these libraries aren't shared between the two projects, and that is a huge shame since it is the absolute most important aspect of an office suite.

        Comment


        • #19
          Originally posted by 89c51 View Post
          just to make myself clear its not a LO vs Caligra thing.

          i just want 1 (one) suite that is feature complete, cross platform (including mobile) and can compete with the rest.
          No such thing as “feature complete”. I have likely a very different definition of that than you. IMO the only missing feature of Calligra Words is RTF export, although vastly superior app load time compared to LO more than makes up for it.

          Originally posted by liam View Post
          Just wait for awhile. LO is working on a mobile version withstand the gtk Broadway backend. However, don't expect it in less than a year.
          I'd say a year is overly optimistic. To get an actual mobile version, real modularization is required: A task that already failed several times.


          Originally posted by liam View Post
          I don't see the point of a KDE office unless it is simply a toolkit exercise. Building an office suite is crazy amounts of work.
          So? Most work is already done. Calligra didn't start from scratch but – at least partially – dates back to 1997 (KPresenter was started then). KOffice was mostly held back by text rendering bugs in Qt. It took a motivated Nokia employee to finally fix that bug in Qt 4.8.

          Originally posted by liam View Post
          Having a code base that is 20 years old isn't necessarily a bad thing
          I don't know how your math skills make an 1980s code base 20 years old but whatever.

          Originally posted by liam View Post
          as LO is doing, unused functions are continuously searched for and pruned.
          That’s a claim often repeated by LO devs but it only is a small fraction of the true story.
          LO’s biggest problem aren’t the unused functions, but the crappy functions that are actually used and not replaced! Its VCL toolkit needs to go ASAP but there are absolutely no attempts at all to completely rewrite the GUI using a modern non-buggy toolkit within the LO community.
          The funny thing is that Apache OpenOffice will get a completely new GUI IBM developed years ago for Lotus Symphony, although LO fans already declared OO dead.

          Originally posted by liam View Post
          LO has a very large and active community,
          …that achieved almost nothing. Where is the VCL-less GUI? Why are all LO components still in one huge “soffice.bin” process? Why is KDE integration still broken?
          Both projects – LO and Calligra – were initiated in late 2010. During pretty much exactly the same time LO devs were struggling removing unused code whereas Calligra devs built a completely new word processor from scratch!

          Originally posted by liam View Post
          and has excellent support for the major formats (which, I would imagine is the hardest thing to achieve), but even they don't have ideal format support.
          Fact is nonetheless that Calligra has better MS OOXML import support than LO – a feat achieved in a shorter time than LO even with Red Hat’s and SUSE’s full-time employees.


          IMO it’s a testament to the quality of the code base when a handful of mostly hobby developers achieve more in Calligra than the army of full-time devs achieve in LO during the same time.
          TDF will need to invest lots of development time in cleaning up LO’s rotten code base to make it easily extendable.

          Comment


          • #20
            Originally posted by liam View Post
            Just wait for awhile. LO is working on a mobile version withstand the gtk Broadway backend. However, don't expect it in less than a year.
            I don't see the point of a KDE office unless it is simply a toolkit exercise. Building an office suite is crazy amounts of work. Having a code base that is 20 years old isn't necessarily a bad thing if, as LO is doing, unused functions are continuously searched for and pruned. LO has a very large and active community, and has excellent support for the major formats (which, I would imagine is the hardest thing to achieve), but even they don't have ideal format support. To my knowledge these libraries aren't shared between the two projects, and that is a huge shame since it is the absolute most important aspect of an office suite.
            Calligra suite is written, I believe, in C++ using Qt and KDE libraries. It employs an interesting concept of "flakes" to share components between applications.

            http://en.wikipedia.org/wiki/Flake_%28KDE%29

            The functionality of Flake is divided up between Shapes which display content, and Tools which manipulate content or the user interface. Different Shapes can be created to support different kinds of content, for example the text-shape in KWord would support .txt and .odt formats while the Kchart shape would just support chart related document standard like .odc. Shapes are packaged with a set of tools to manipulate that kind of content and UI elements that expose the functionality to the user. This provides an application with all the features it needs and also allows for easy embedding of Shapes in other applications. Shapes can load other shapes when needed, for example when images are in text documents the image shape will be loaded to handle the images.
            This means that a "table", for example, in Calligra Words is the same thing, and uses the same code, as a "table" in Calligra Sheets, Stage and Kexi.

            Calligra has a "core" of functionality, and a few different GUI's are built using that same core. At this time there are three such GUI wrappers.

            http://en.wikipedia.org/wiki/Calligra_Suite

            Desktops

            Calligra’s main platform are desktop PCs running Linux, FreeBSD, Mac OS X, and Windows, of which Linux is the best supported system.

            Smartphones

            Calligra Mobile is a version for smartphone-like formfactors. Its main purpose is to serve as document viewer for devices running Maemo or its successor MeeGo but simple editing features are also available.

            Tablets

            Calligra Active was launched after the Plasma Active initiative to provide a document viewer similar to Calligra Mobile but for tablet computers.
            This is obviously a case where a new codebase and saner, cleaner structure is far easier to maintain and extend to different versions.

            This is so important, in fact, that it DOES provide a "point" for the Calligra Suite. Very much so. Maintainability and portability of Calligra Office are streets ahead of OpenOffice and LibreOffice. It is not "crazy amounts of work", the development team is only about 87 members strong.
            Last edited by hal2k1; 04-13-2012, 12:37 AM.

            Comment


            • #21
              Originally posted by liam View Post
              Just wait for awhile. LO is working on a mobile version withstand the gtk Broadway backend. However, don't expect it in less than a year.
              Yet the Calligra developers were able to develop an entire new mobile version, Calligra active, in a few months, and one developer as a trial created a partially-working port to android in a few days.

              Originally posted by liam View Post
              I don't see the point of a KDE office unless it is simply a toolkit exercise.
              The fact that they are able to create an entire new mobile interface in no time at all should show one of the points: having a flexible, modular, easily-to-develop-for code base. That is something that LO doesn't have and may never have.

              As someone else pointed out, the fact that it needed a 3-month GSOC project just to get a multi-line cell editor in calc shows just how fundamentally broken the LO code base is.

              Comment


              • #22

                I'd say a year is overly optimistic. To get an actual mobile version, real modularization is required: A task that already failed several times.
                I am simply parroting the expected date given by, I think it was Meeks. As has been said, getting the entire suite running on a mobile platform will be...difficult (however, I also think that for the near term, at least, it would also be overkill since it would only be really usable with a keyboard/pointer).


                So? Most work is already done. Calligra didn't start from scratch but – at least partially – dates back to 1997 (KPresenter was started then). KOffice was mostly held back by text rendering bugs in Qt. It took a motivated Nokia employee to finally fix that bug in Qt 4.8.

                Wait, Calligra is based on code that's FIFTEEN YEARS OLD!!! That simply cannot be

                I don't know how your math skills make an 1980s code base 20 years old but whatever.

                Fair enough. So say twenty-three years (you didn't specify when in the 80's). That is a whole seven years older than some of the Calligra code.
                Since this should all be POSIX, I don't see where having an older codebase is NECCESSARILY worse. All things being equal, yes, I would agree, but optimizations are happening since this isn't new information you are giving.

                That’s a claim often repeated by LO devs but it only is a small fraction of the true story.
                LO’s biggest problem aren’t the unused functions, but the crappy functions that are actually used and not replaced! Its VCL toolkit needs to go ASAP but there are absolutely no attempts at all to completely rewrite the GUI using a modern non-buggy toolkit within the LO community.
                The funny thing is that Apache OpenOffice will get a completely new GUI IBM developed years ago for Lotus Symphony, although LO fans already declared OO dead.

                At least you didn't call me a fanboy
                Meeks pushed a pure gtk3 effort back in Oct (http://lists.freedesktop.org/archive...er/000302.html), though I haven't noticed much since. Otherwise, I don't disagree, but, IMHO, the interface isn't a huge priority.

                …that achieved almost nothing. Where is the VCL-less GUI? Why are all LO components still in one huge “soffice.bin” process? Why is KDE integration still broken?
                Both projects – LO and Calligra – were initiated in late 2010. During pretty much exactly the same time LO devs were struggling removing unused code whereas Calligra devs built a completely new word processor from scratch!
                I'm not quite sure why you are asking this question since you've already said the answer. The codebase has needed vast amounts of work. Office suites, as I said earlier, require fidelity, and that can sometimes be hacky (depending on the standard), but it always requires ungodly work to make sure formats are treated the same across platforms (not easy, obviously, when dealing with something like Word's ooxml). Toolkits (unless they actually implement these standards for you) don't really help beyond getting up a skeleton with a bit of skin. It's the stuff inside that takes so damn long, but it's not terribly visible. If you want to criticize LO for anything it should be for not using robust enough format testing. That is the thing that bugs most users.


                Fact is nonetheless that Calligra has better MS OOXML import support than LO – a feat achieved in a shorter time than LO even with Red Hat’s and SUSE’s full-time employees.

                How do you judge this? The only thing I found that compared the two's ability to deal with formats was a slashdot comment that said that Calligra's equation handling is a bit broken.

                IMO it’s a testament to the quality of the code base when a handful of mostly hobby developers achieve more in Calligra than the army of full-time devs achieve in LO during the same time.
                TDF will need to invest lots of development time in cleaning up LO’s rotten code base to make it easily extendable.

                I think that it is extremly arguable that they have achieved more, but I'm not going to take away from the work they've done. At the same time, do they support ALL of the same formats as LO, and to at least the same degree? If they have, then they are geniuses without parallel in the history of computing and Turing bows to them.

                Comment


                • #23
                  Originally posted by liam View Post
                  I am simply parroting the expected date given by, I think it was Meeks. As has been said, getting the entire suite running on a mobile platform will be...difficult (however, I also think that for the near term, at least, it would also be overkill since it would only be really usable with a keyboard/pointer).
                  It is not difficult but basically impossible as long as neither VCL is removed nor LO modularized.
                  Did you know that when Sun back then to open up StarOffice’s source code, the GNOME project wanted take that code, split it apart and base its “GNOME Office” on it? It was a huge failure: It simply couldn’t be achieved with that messy source code.
                  Yes, the source code improved since then but it is simply a fact that modularization and VCL removal could still not be a achieved.

                  Originally posted by liam View Post
                  Wait, Calligra is based on code that's FIFTEEN YEARS OLD!!! That simply cannot be
                  Well, yes, but you are missing a very important piece of the puzzle: The big KOffice 2.x rewrite. It took the then-small team years. In fact Kivio’s 2.x/Qt4 port took until a few days ago when it was released as Flow in Calligra 2.4.
                  The KOffice 1.x source code began to rot and instead of going the OO/LO way to work around fundamental problems, it was rewritten. Heck, since then even KWord began to rot and was entirely rewritten as Calligra Words.

                  Originally posted by liam View Post
                  So say twenty-three years (you didn't specify when in the 80's).
                  And could not look it up on Wikipedia yourself?
                  StarDivision was founded in 1984 and the first version of StarWriter for the Zilog Z80 was released shortly after.

                  Originally posted by liam View Post
                  Meeks pushed a pure gtk3 effort back in Oct (http://lists.freedesktop.org/archive...er/000302.html), though I haven't noticed much since. Otherwise, I don't disagree, but, IMHO, the interface isn't a huge priority.
                  That is just a GTK3 theme renderer for VCL. It is not a new GUI for LO written in GTK3.
                  And no: If TDF is serious about mobile versions, VCL removal must be a high priority, unless TDF is only targeting future high-end platforms with multiple gigs of RAM and many CPU cores.


                  Originally posted by liam View Post
                  How do you judge this? The only thing I found that compared the two's ability to deal with formats was a slashdot comment that said that Calligra's equation handling is a bit broken.
                  It is so simple: I have installed both and simply opened documents in both.
                  Granted, I only compared Words with Writer and only with a handful of documents I could find.


                  Originally posted by liam View Post
                  I think that it is extremly arguable that they have achieved more, but I'm not going to take away from the work they've done. At the same time, do they support ALL of the same formats as LO, and to at least the same degree? If they have, then they are geniuses without parallel in the history of computing and Turing bows to them.
                  No, it is not arguable. How many entirely new file formats did LO adopt in the last 1.5 years? Pretty much all file format support was inherited from OO/Go-OO. I was simply referring to the achievements in the last 1.5 years when both projects were founded, not the achievements both predecessors (OO and KOffice) did.
                  Calligra devs wrote an entirely new word processor because they found KWord to be broken, they wrote a tablet GUI (Calligra Active) and in the master branch actually import support for many new formats was introduced: CorelDRAW for Karbon, MS Visio import for Calligra Flow and MS Project, Planner, and Primavera for Calligra Plan. How many new file formats were added in LO in that time?

                  Comment

                  Working...
                  X