Volunteer spotlight: Yavor Doganov

Yavor Doganov is the translations coordinator for gnu.org and the author of GNUnited Nations, a new project which hopes to make management of translations significantly easier. He lives in Bulgaria.

Questions were asked by Matt Lee, FSF campaigns manager.

Matt: Yavor, you've been working on a new system for handling translations on our website. What's it called, and how does it work?

Yavor: Yes, it is called GNUnited Nations (or GNUN for short). The name is, as you may guess, a pun of United Nations (UN).

GNUN works much like GNU gettext works for programs or like the GNOME Documentation Utilities (gnome-doc-utils) work for translations of GNOME manuals. For years, translators had to duplicate the HTML markup and examine changes as diffs between revisions of the original article(s). This is extremely tedious and error prone, which is why the GNOME Project has developed its own package to handle translations of the manuals (whose source is DocBook).

The source of almost all gnu.org articles is XHTML, so GNUN converts each article to a PO template, and each translation has its own PO file. The benefit is tremendous, because if the original article changes, each of its article.LANG.PO files is updated automatically, and subsequently all of the HTML translations are rebuilt. The translator can easily identify the changes using a PO editor and updating many articles is possible with little effort.

The overall benefits for gnu.org are also worth mentioning. Since translations automatically follow the markup of the original articles, it would be possible to update them at once when we move to a new standard. If a team becomes understaffed and undermaintained, its translations will not rot but will get rebuilt automatically following the changes in the originals. Of course, this means mixed native/English language but we will turn this slight annoyance to be our ally: we will identify teams that need attention and will try to attract new volunteers.

Matt: What are the biggest challenges with the translation of gnu.org?

Yavor: I think GNUnited Nations will solve most of the technical aspects of the problem. We will enhance the system to make it work for people who are not comfortable working on the GNU/Linux console, and will implement some sort of web-based automatic statistics.

But this is only part of the job, the eaiser part. The biggest challenge in gnu.org translation is the lack of volunteers, or more precisely, the constant lack of devoted contributors. This should not be surprising to anyone, though. In the Free World, people usually work on what they like and they switch to something else once they lose interest. The most appropriate example for a job that is always exciting and rewarding is being a programmer. Most programmers work on what they enjoy, and they switch projects as time goes by (well, sometimes). Programming is facing challenges every day, and finding (clever) ways to solve problems.

Translators are, in my humble opinion, in the other camp. Translating programs is more or less a monotone activity, and people usually burn out after 3-5 years. Translating documentation is substantially harder. Translating the GNU philosophy is the hardest thing ever for a translator, because it is a difficult activity that requires a lot of thought, understanding and responsibility. Awareness about the ethical issues that the free software movement set out to solve is still very low, which leads to the lower number of translators contributing to gnu.org compared to, say, GNOME.

Matt: Have you looked at HTML 5 at all?

Yavor: No, I dislike anything related to HTML and read only what is necessary to complete the tasks I'm doing. Perhaps even less. I find it ironic that W3C develops standard after standard, each determined to solve entirely all problems. And we are doomed to maintain all standards forever, since no one will use a browser that cannot render a page in HTML 2.0. The only thing I know about HTML 5 is that the committee rejected Ogg Vorbis/Theora which (provided it's true) doesn't make me feel excited and filled with hope.

Matt: If someone wanted to get involved with translating gnu.org, what is the work like? What kind of skills do they need?

Yavor: Well, the work is enormous given the fact of how many articles are available to translate. Even if we count only the essays, and only the most important essays, it is still a titanic effort for a new language. But don't despair, GNU was not built in one day! If the amount of work had scared away the countless number of GNU maintainers and contributors, we wouldn't have our beloved free system today.

A translator should read and understand English well and have good knowledge about our philosophy and the various issues raised and discussed in the articles. It is not a requirement to be a fluent speaker -- in fact, many translators cannot speak English well but they understand it perfectly when reading. That's the important part. Of course, time is always necessary. A relatively short essay might look like a job for an hour or two, but it often turns out to take days or weeks. Sometimes it is necessary to do some research, or discuss problematic phrases, and this takes time.

Translation teams play an important role by establishing the most appropriate terminology for the specific language and of course by presenting the philosophy of the free software movement to readers who do not understand English (and very often, to those who understand English but would not bother to read an essay in English).

The most valuable "feature" a translator could have is her dedication, and her determination to keep going; her firm belief that translating these essays is a job that is useful for the society. Everything else, like small technical skills, the usual translator's "sense", etc., will get settled by itself, one way or another.

Matt: And finally, what's your favorite piece of free software?

Yavor: GNU Emacs, undoubtedly. Emacs is special in so many ways that it is impossible to explain its nature to someone who has not seen/used it. It is most definitely one of the Wonders of the World but unlike them it keeps surprising you every day. I use Emacs for absolutely everything; I also believe it has helped me a lot in my decision to entirely reject non-free software some years ago.