W3C home > Mailing lists > Public > www-validator@w3.org > August 2011

Translations of the user interface

From: Jukka K. Korpela <jkorpela@cs.tut.fi>
Date: Fri, 19 Aug 2011 10:29:58 +0300
Message-ID: <4E4E10F6.2070905@cs.tut.fi>
To: www-validator@w3.org, r.umlauf@googlemail.com, jobst@barrett.com.au
19.8.2011 2:21, Jobst Schmalenbach wrote:

> Ok, I bite

Bite what? You seem to be commenting on a message that was posted in 
German and you reply in English, the discussion language of the list, 
but you don't explain what the question was.

Well, the question is not crystal clear, but it seems to be that Mr. 
Umlauf asked why the validator's user interface is in English only. 
That's a good question, and the answer is that nobody translated it.

Other W3C services like the W3C CSS Validator
are available in several languages, because
a) some people made it possible to localize it (this typically involves 
a lot of work to globalize it so that translations can be fed in)
b) some other people translated the texts into their languages.
As far as I know, at least part b) is purely volunteer work, and I 
suppose part a) was at least partly carried out on a volunteer bases.

So the more concrete answer is: people just didn't volunteer to 
translate it.

In my opinion, the markup validator's messages are largely highly 
technical in nature, using e.g. SGML terminology, which is rather 
difficult to translate. And the few people who understand SGML terms 
there days would probably have difficulties in recognizing them in any 
other language than English, as there are no "standard" translations for 
them (except perhaps in a few languages).

For example, considering the validation report for the page
(which seems to be what Mr. Umlauf had difficulties with)
the first message is:
"cannot generate system identifier for general entity 'h'"
I don't think it would help much to translate that into German, for 
example. Perhaps one user out one hundred among validator users really 
knows what "general entity" means; if you translated "general entity" 
to, say, "allgemeines Ding", make that close to zero out of a hundred.

The explanations below the error messages are often (and in this case) 
much more understandable than the messages themselves. But people 
generally don't read them much, perhaps thinking that small font size 
indicates lack of importance (or that length of the text scares them 
into thinking that the text is even more difficult than the message 
itself!). Moreover, it would be difficult to find competent volunteers 
to translate such large amounts of texts.

The simple error message would, in the particular case I mentioned, 
simply say "Replace each '&' by '&amp;'". But making the validator issue 
such messages would be a major change, and a difficult one.

So if you ask me, there would not be much point in translating the 
markup validator. Things may change with HTML5, as it is no more SGML 
based or XML based but has its own basis and terminology. If HTML5 and 
especially HTML5 checkers (confusingly called "validators") become 
reasonably stable, a translated user interface for an HTML5 checker 
would make much sense - if at least some minimal quality checks can be 
imposed on the translations, including the translations of basic terms.

Now, to avoid some misconceptions...

> That means the machine is located in America, why should w3c respond in
> any other way????????

If the service were localized, it should make a good effort in selecting 
the user's preferred language instead of wildly guessing language from 
IP address.

And that's what W3C CSS Validator does. It uses the Accept-Language HTTP 
headers (indirectly settable by users by changing their browser 
preferences), and it contains explicit links to different language 
versions too.

> Further, the content of the website you are checking is in English, too.

That's not relevant. People often work with web pages written in a 
language other than their own primary language.

Yucca, http://www.cs.tut.fi/~jkorpela/
Received on Friday, 19 August 2011 07:30:20 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:31:07 UTC