Pages of the blog

24 October, 2017

History and some causes of LibreOffice bugs

He took the small bundle from Karlsson and held her tenderly in his arms.
"Don't cry, there's a good baby", he said.
Astrid Lindgren "Karlsson on the Roof"
You certainly know that the LibreOffice includes several modules, such as Writer, Calc, Draw... All these modules are united into one application and can not be installed separately. But most of all you can be confused by the various development (though, in truth, the abandonment) of these modules. In order to understand better why this diverse interest of developers to different modules grows, the appearance and correction of errors and the inadequacy of the documentation, I suggest first to dig into the history of the appearance of LibreOffice. And we will begin from the very beginning.


In the far-distant 85 year of the XX century, the German company StarDivision GMBH, created by 16-year-old programmer Marco Berris (Marco Börries), developed an application called StarWriter. And it was worked under MS-DOS. A normal office suit was not available at that time. As suit StarOffice was released only in 1994 and included already StarWriter, StarCalc and StarBase for Windows 3.1. In 1995, it already had version 3.0 and StarWriter, StarCalc, StarDraw, StarImage, StarChart and could work under MS-DOS, Windows 3.1, OS/2, Solaris SPARC, Power Macintosh. In 1996, version 3.1 was released, with a distinctive feature of Linux support. Somewhere between 95 and 98 years, the suit was packed with the personal manager StarSchedule, the mail client StarMail, the feed aggregator StarDiscussion and several other components, whose normal names I cannot find now. Later all these components were removed from the package by Sun Microsystems. And, by the way, the suit was proprietary, it became free later. It is difficult to assess the popularity of this office package at this stage. Of course, if you set a goal to dig, then you can dig a lot. But I think we will be interested in another.

In the autumn of 1999, StarDivision was acquired Sun Microsystems. At the same time, the source code was opened, and starting with version 6.0, StarOffice was released on the basis of OpenOffice.org. Since 2005, Sun announced its rejection of SISSL, and the code became available only under the LGPL license. We must pay tribute to the developers of Sun Microsystems, who contributed the lion's share in the development of a free office suite. With the opening of the source code, it also began to form a free community around the suite. People who were interested in the development of documentation, design, code base began to apper. And, to be honest, in my opinion, LibreOffice teams still can not repeat this high day of the creativity. OpenOffice.org was released from 2001 to 2010 under the control of Sun Microsystems. And the hallmark was that there was a plan for the version, and the versions were released on readiness. Then there were several corrective releases. But still, one of the main features was that the major releases were tested by the community, and key mistakes were corrected before release. Of course not everything, but nonetheless... And this feeling of a holistic, complete and stable product disappeared in LibreOffice. For the sake of completeness, it should be mentioned that in Wikipedia without references to sources it is said that 25 million StarOffice licenses were sold. It is not entirely clear whether this was only during the existence of StarDivision, or Sun Microsystem is included here too. But according to research FlashCounter Statistics Service in early 2010 in Germany, OpenOffice.org and its derivatives were installed on 21% of user computers. That is, more than a fifth of the user's computers. A few forks appeared during the existence of OpenOffice.org Of the most important forks for history, we can mention the following three: Go-oo, which is essentially just a set of patches for improvement; NeoOffice, the goal of which was improved integration with the OS X environment; IBM Symfony was created as another commercial version of the office suite.

In 2009, Sun Microsystems was absorbed by Oracle, to which were also transferred the rights to brand names StarOffice and OpenOffice.org and their code. For the OpenOffice.org project, this period is more like a death cramp: a hard press community from the side of Oracle, in this reason the community collapse and extinction of interest from active members. In fact, for a year Oracle was able to destroy everything that was achieved. Oracle released its StarOffice, renamed the Oracle Open Office. At the same time, two versions of OpenOffice.org are released. But already in September 2010 the creation of the LibreOffice project was announced.

If you like to look for dirt in a human relationship, you may find that any open community, however much it glorifies the bright ideals, in fact is just a community, with all its shortcomings. These shortcomings also hinder it, also like to any other community. We are interested in history.

The official version of the LibreOffice project is Oracle's pressure to the OpenOffice.org project. In this regard, the The Document Foundation (TDF) was created and the fork from OpenOffice.org called LibreOffice was organized. By the way, the organizers of TDF asked Oracle to transfer the rights to the OpenOffice.org trademark to the fund, but the company denied it and gave the code and trademark to the Apache Software Foundation, further destroyed the community. The LibreOffice project was supported by many opensource projects. But to be honest, then, most likely, this support was more a kick to the hated Oracle than a real support to a free project. In fact, the first versions did not differ from OpenOffice.org.

From the very beginning, the LibreOffice project took advantage of the strategy that Google used to promote its browser Google Chrome. If you first releases of browser Chrome, maybe you remember that the browser just madly was changing the versions the first six months or the year. It seems to me that at first it was generally once in 2 weeks, and I'm not talking about bugfix releases, but about main releases. And I remember my opinion when I watched this: "How fast it develops, how intensely... It's the technology of the future!" Believe me, I did not dig in the versions, I just watched how the numbers change in versions and I made my conclusions. And I think many compare so, if some product releases versions more often, it seems to us that it develops faster. The LibreOffice team consciously or unconsciously used this principle to immediately position themselves as a rapidly developing solution. In fact, the amount of actually entered code in the project does not increase from this. And as a programmer, a documentation team member and a permanent user, I see that some components have not changed since Sun Microsystems, but not because it's unnecessary. The main changes are made in Writer and Calc. And these changes are not complete, especially at interface level. Even if you take only Writer, you can find that when working in normal mode and in Master document mode, we see different design of the panels, menus and different behavior of the pop-up windows. And IDE Basic, for example, has not been touched since the days of OpenOffice.org. Of course, in part this is due to the fact that the project does not have a clear concept. Now its development looks more like the evolution of a biological organism: each developer makes some changes which can be called "mutations", and eventually there are "useful" or "harmless" mutations, but there is no holistic development plan. This is the principle of meritocracy in action – who made more "mutations", the one is right. The integrity and usefulness of these mutations is considered only from the point of survival of the species. The second reason can be called psychological. Significant versions of LibreOffice are released every six months, and every six months there is a cut-off in the developer's head – "ready". Instead of stopping, looking around, creating a concept and a plan and trying to implement it in the next year and a half, our developers hurry up to make, because, in six months, they need to make a new, something meaningful. And "we finally put everything in order" will not work, a new one has to be made, because it is not bugfix release. This attitude to the project creates a load on many of the accompanying teams, the documentation and localization teams suffer the most. It's no secret that today, the documentation team lags behind for several years (years, not versions), the localization team is busy with the translation of the interface and they do not have enough time to translate the documentation. Russian-language help has not been translated since the days of OpenOffice.org, and those changes that were made a couple of years ago, were carried out with the assistance of the German company CIB Software. The only team that is happy with the rapid change of versions is the marketing team. Because, even if "the train does not go", they can still "swings the train cars and announce stops."

If to be honest before the end, it should be pointed that the development of LibreOffice is not only by individuals, but also by commercial companies. Greatest contribution is made by the British company Collabora, German CIB Software and American Red Hat. At the same time, Collabora developed Online LibreOffice and the version for Android, returning the code to the project. In addition, if you look at the TDF advisory board, you will see many interesting names: non-profit organizations are Free Software Foundation, GNOME, KDE e.V .; and commercial are Google, Intel, AMD, Canonical, RPA RusBITech. And to date, the code base, at least in popular components, has gone very far from its progenitor OpenOffice.org.

But with all this, mistakes are accumulated by a huge wave, the interface is not unified, and there are absolutely incomprehensible small defects that poison the working day in LibreOffice. Yes, and poor compatibility with proprietary formats. And it would be logical to freeze the new main versions release and put everything in order, to release a full-fledged, complete office suit, but...

Why do I love LibreOffice? Why, for all the shortcomings, I use it as my main office suit, and from LibreOffice Calc I get out "just to have tea".

The reason is that, to date, LibreOffice is truly a giant in the world of open source software. But even if we do not take into account openness, and for many users it does not matter, important only to free (do not pay money), LibreOffice could implement a lot of convenient concepts that greatly simplify the work of an office employee. Mistakes – yes, they are. But almost any bug that can now be found in the application has a fairly simple bypass. And it is more important for a person to have a professional tool. And the more professionally you are able to use the tool, the less significant mistakes are for you, but the very concept of working with the tool is more important. And in this, as I said, LibreOffice has no equal.

Source and additional information: