W3C home > Mailing lists > Public > w3c-translators@w3.org > October to December 2007

Re: translation Circus

From: Charles McCathieNevile <chaals@opera.com>
Date: Mon, 31 Dec 2007 17:29:19 +0100
To: "gareth edison" <gareth.edison@googlemail.com>, w3c-translators@w3.org
Message-ID: <op.t37j25ibwxe0ny@widsith.local>

On Fri, 28 Dec 2007 13:17:23 +0100, gareth edison  
<gareth.edison@googlemail.com> wrote:

> Good evening everyone here at w3c translations,
>
> As a long time supporter of the W3C project I would like to voice my  
> opinion...

Which is a fine thing...

> What was supposed to be a good idea is turning into a fiasco of link  
> hungry webmasters posing as translators who are translating documents
> no one really wants or needs.

It is questionable whether nobody wants or needs the documents - having  
worked with professionals in a wide variety of languages I tend to think  
that even a bad translation is as useful aa someone badly interpreting a  
document becaus they don't undertand the original. After all, it is easier  
to correct a document whose existence you know of than to correct a  
misunderstanding in the mind of someone you never meet even virtually, and  
with whom you share no common language.

It is pretty clear that people are doing this for linking karma. So long  
as they notify their translations, and accept that these can be checked  
for quality by anyone who wants to and is able to explain what's wrong, I  
would prefer to have the translations and pay link karma as a quid pro quo  
than not to have the translations.

> I fail to grasp the importancy of translating documents into languages  
> like Turmen, Uzbek, Azerbajan, Kazakh, Belarussian, Ukrainian, Estonian,
> Latvian, Tatar, Georgian or even Armenian.

Well, if you happen to speak such a language, which quite a lot of people  
do, then it is helpful to find documents in that language.

I have had the opposite problem. There is an excellent accesibility  
testing tool called Hera. Open source, multilingual, wonderful and all,  
But the developer happens not to speak much english. (He did do a  
good-quality translation of a major W3C specification to a top-5 world  
language some time ago, by getting help with parts he didn't understand  
clearly). In order to allow a community whose languages include english,  
spanish, french, portuguese, danish, romanian and a few others but who  
have *no* common language, we accept translations based on other  
translations. It's not perfect, but it means that we get better quality  
localisation than having none at all. (We also encourage the entire  
translation community to verify and cross-check and propose changes even  
to the original text where there is a clearer way of expressing it).

> Imagine Indian webmasters translating these documents into some of the 50
> different dialects of Tamil or Sanskrit

I work for a company that does imagine that, and tries to find competent  
people to do so, ecause it turns out that it is useful to real people in  
the real world.

> My question is, where will W3c draw the line?
>
> The Turkish translation below is just one of the results of people
> translating documents into languages they are not familiar with.
> This document was clearly translated into Turkish from a Russian  
> translation which is spoiling the high standard of work required by
> W3C in order to produce quality translations.

I have edited and continue to edit W3C specifications. Let's not fool  
ourselves - these are documents written mostly by engineers whose skills  
are not in communication through language, and by and large the quality is  
"not bad". I agree that we should strive to maintain as high a level of  
quality as possible, but we don't need to treat thes things as some kind  
of holy text - it is an attempt by a bunch of people to write down  
something they do or plan to do in a way that makes it easy for others to  
copy it.


To return to your point about link-hungry webmasters and minor documents -  
it does upset me that people are clearly chasing simple thigns to do for  
the links, but not so much that I don't want the translations. (Anyway,  
the requests are as useful a source of information for those who are  
developing systems to identify people gaming search algorithms and the  
like as the translations are for such people). But the reason there are  
documents approved for translation and those that are jsut out there is  
that many of the documents W3C publishes are not considered especially  
high wuality (though some such documents are in practice), but are made  
available because it is important to get the information out.

> Wouldnt it be much wiser to allow ONLY *native speakers* to translate
> documents for W3C instead of people producing translations
> which they cant read themselves?

Not really, in my opinion. People in the real world use automatic  
translations of things they find, so we are unlikely to end up worse off  
than that by offering something that is based on that principle with a  
small number of really glarin errors corrected. At least if we see the  
document we can suggest improvements, even if we don't have the time to do  
a complete high-quality translation ourselves. (For the same reason, while  
a translation of a translation is sub-optimal, I see no reason to ban it  
in practice).

A more open community discussing the quality of translations, and using  
that to provide feedback on the clarity of original documents, would  
positively benefit W3C - and if what people get from that is a 1%  
advantage to their real job as a spammer, well good luck to them. (I still  
hate spammers and work on systems to make their life miserable, but if  
they improve my life I don't see why they should not be allowed to get any  
beenfit from that improvement).

> Maybe W3c should start only allowing main
> languages instead of sub-divisions of these languages like the
> many Russian dialects as mentioned above.

Along with most other respondents I think this demonstrates taht you don't  
know much about the languages you listed (estonian is related to finnish,  
but nothing at all like russian, english is closer to german than latvian  
or estonian to russian). That probably explains why you would even make  
such a terrible suggestion. I think it is best forgotten as quickly as  
possible.

Cheers

Chaals

-- 
Charles McCathieNevile  Opera Software, Standards Group
     je parle français -- hablo español -- jeg lærer norsk
http://my.opera.com/chaals   Try Opera 9.5: http://snapshot.opera.com
Received on Monday, 31 December 2007 16:29:13 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 20:27:40 UTC