[josm-dev] JOSM-Tested

[Dirk Stöcker]

On Sat, 12 Jun 2010, Frederik Ramm wrote:

>> Actually that is a very bad example, as the original translator was me and
>> the word you talk about was token. Your translation is much better than
>> mine was, but I did know what I translated and tried long to find a
>> good translation,
>
> Sorry then. But you agree that making good translations in the JOSM
> context will in all but the most primitive cases require an OSM background?

Yes and no. Really skillfull translators know when to stop and only do 
suggestions or skip strings when they miss context. Also when you have no 
idea, then looking at already translated strings with same words helps to 
find correct context.

I more tend to the idea of having many people translate possibly with 
errors and let review kill these errors. A "fix a string" is usually a 
smaller step to do than 4000 strings are missing (Well at least people 
think so - when you detected that fix a string actually means fix 300 
strings, then its too late - you're already half through :-).

Contrary to commercial translation for OpenSource you do a step by step 
work, which means quality also varies for one translator. That's why for 
me the tools support is very important:
  - good search mechanism
  - proper documentation of the translations style
  - suggestion mechanism to help for new/modified strings
  - contact and user interaction
  - messaging and notifications

Launchpad has several drawbacks in these areas, but at least the general 
structure is usuable and allows acceptable workflow.

Ciao
-- 
http://www.dstoecker.eu/ (PGP key available)