On 13/08/2006, at 5:29 PM, Steve Langasek wrote:
On Sun, Aug 13, 2006 at 04:12:15PM +0930, Clytie Siddall wrote:
On 13/08/2006, at 4:43 AM, Steve Langasek wrote:
Using non-ASCII in the original English version is likely to
some problems with PO files handling, os I'd recommend changing
with caution even if it's certainly correct as the English
have stolen that word from my language..:-)
How about a recommendation not to use words from another language
unless you can write them properly?
Given that such words can be found in English dictionaries, that's
not a very useful recommendation.
As a linguist, I am aware that incorrect forms of words can be
variants in some use.
As a linguist, I and 200 million of my native English-speaking
Steve, I am a native speaker of English, and I lecture in English.
Otherwise I would not make recommendations on English.
reject your claim of authority over what constitutes the "correct"
of a loan word whose source form includes diacritics not present in
According to the dictionary you quoted, "voila" is only a variant of
the primary form "voilà". Primary forms are preferred over variants.
I'd still strongly advise either using the correctly-spelt form of
I disagree; I think we should only use correctly-spelled forms of
I accept "voila" as a valid variant of "voilà" in English;
the only spelling of the word which is guaranteed to be compatible
files for all destination languages due to encoding concerns.
or not using it at all.
Which would be a fair recommendation, but such an interdiction
much weight if you don't have consensus on the question of what is
an acceptably correct form.
I'm agreeing with your dictionary quotation. It would only waste time
if we compared different dictonaries and got into research papers.
Consensus on language is a moving target. ;)
Translators are not expecting incorrect forms in the original
have enough difficulty with the wild variety of vocabulary and
in original strings, without having to cope with incorrect forms of
loan words (entire words borrowed from another language).
Quite frankly, I consider that the translators' problem, not the
maintainers'. It is quite reasonable to constrain source English
with style rules concerning consistent use of vocab, forms of
UI references because these are rules that benefit the primary
the string: the user. Subsetting the language for the benefit of
translators, OTOH, is a misoptimization which impoverishes the user
experience and deprives the translators themselves of opportunities
I don't think making translators look up variants of loan words is
useful. In my experience of translation projects, loan words in
general are often misunderstood, causing an incorrect translation.
The GNOME developer's choice to spell "né" as "ne" confused nearly
all the translators, and wasn't recognized even by the French
from Clytie (vi-VN, Vietnamese free-software translation team / nhóm
Việt hóa phần mềm tự do)