I consider that translation tools have three main functional areas, technical interaction capabilities, teamwork support capabilities, and translation support capabilities.
The technical part includes:
- supported file formats
- synchronization with the source repository
- automation support tools
- the interface possibilities
Support for teamwork, which could also be called “community animation” includes:
- Monitoring changes (by a translator, on a project…)
- Follow-up on updates pushed by projects
- The display of the state of the situation
- The ability to enable or not the review and validation steps
- Assistance in discussion between translators (from the same team and inter-languages)
- Assistance with discussion with the project maintainer
- The possibility to communicate globally on the platform (news, etc.)
Translation assistance includes:
- A clear and ergonomic interface
- A limited number of steps to find your project and start working
- A simple ability to read the flow between translation and distribution
- Access to a translation memory machine
- The enrichment of a glossary
When we look at these few points, there are finally no big differences with a source code management platform on the first two perimeters. I even suspect that the last point is largely valid for source code. However, the data handled is quite different and users are often much less technically sophisticated than developers, although they are much more numerous.
Now let’s look at the ones that are the best for me, the GNOME platform:
- Their site contains both the organization of the teams and the translation platform. We see who is responsible and the roles in the team. Everything is concentrated in a few screens.
- It’s easy to find what to work on, but you quickly realize that you’ll have to use files that you download to your computer and then send back once modified. It’s not very sexy, but the logic is easy to understand.
- We send the file back, it potentially sends an alert to the mailing list. The following steps are known and the translation can be easily discussed at the global level (but not comment on a sentence).
This may sound surprising, but it is the best translation platform with 297 languages and a clear percentage of progress, both on basic sentences and advanced menus and documentation. Coupled with a predictable Gnome release schedule, everything is there to work well. It is because they have focused on a tool that promotes community work.
If we now look at the practices of the Debian translation team, which has been doing a good job for years translating unimaginable content for Fedora (especially news), we see that they have a highly codified translation process, based exclusively on emails with manual push in the repositories. They too have put everything into the process rather than the tools, and despite the considerable energy that this seems to me to require, it seems to have been working for years, while being among the leading group of languages!
My perception is that the challenge of a successful translation platform is not based on the ability to make the unitary (Technical, Translate), but on the ability to structure the translation team by helping it to support its processes, this is what gives it sustainability.
It is the production processes that are the most important to structure a team, by putting them together correctly, each newcomer easily understands how it works, registers, appropriates them and then explains how it works to the next ones.
To build a sustainable community, it is therefore necessary: first of all a tool that helps collaborative work, then only to be interested in usability.
To quickly obtain translations on a project, all you need is an ergonomic interface.
This explanation will explain my frustration with the Zanata tool, which is technically efficient, interface efficient, but so poor when it comes to helping to structure our community.