Clean Sheet?

[Jonathon Blake]

Mikko wrote:

> a) While I recognize that a style guide is useful, I don't see why you seem to think a dictionary is evil

I'm not saying a dictionary is evil.  I'm saying it indicates that one
is not fluent in both languages.

> If you are trying to tell me that not remembering the translation of a word, means the person isn't fluent in one of the languages, 

I am.  That little datapoint means that they don't have a fluent
vocabulary in both languages.

> That sounds pretty harsh. 

Do you want an L10N Project to teach people both the language, and how
to translate, or just how to translate?

>Additionally, believe it or not but people are capable of learning
from their mistakes,

I know people can learn from their mistakes.  

The reason for eliminating 80% of the volunteers for translation, is
because they have to learn both the language that they are
translating, and how to translate.

> More people working on translations means more translations. An improved review process mean more accurate translations. 

a)  Both of those can be used, without an L10N Team using a web based
translation tool.

b) An L10N Team can  integrate web based translation tools into the
project home, if they think that that improves translation quality.

> As I indicated in my previous reply, you are comparing apples to oranges. 

> Wikipedia's articles are of remarkably good quality when 

a) I was using Wikipedia as a demonstration that more eyeballs does
not mean less errors.

b) Wikipedia's lack of quality has been extensively documented.

>you are comparing apples to oranges. 

The original assertion of this thread is that L10N projects should
switch to Rosetta, and that software project teams should use Rosetta.

I've been outlining why web based tools do not always make sense,
pointing to other things, as a demonstration of why web based tools
aren't always the best solution.

I haven't touched on specific licencing issues, yet.

> There is no reason why people should be able to freely modify reviewed strings. There is no reason why we shouldn't lock completed, reviewed translations

You already have that in non-web based translation tools.

xan

jonathon
-- 
Monolingualism is a curable disease.
                               Carlos Fuentes