The Unicode Standard and ISO

Marcel Schneider via Unicode unicode at
Sat Jun 9 14:01:39 CDT 2018

On the other hand, most end-users don’t appreciate to get “a screenfull of all-in-English” when “something happened.”
If even big companies still didn’t succeed in getting automatted computer translation to work for error messages, then 
best practice could eventually be to provide an internet link with every message. Given that web pages are generally 
less sibylline than error messages, they may be better translatable, and Philippe Verdy’s hint is therefore a working 
solution for localized software end-user support.

Still a computer should be understandable off-line, so CLDR providing a standard library of error messages could be 
appreciated by the industry.

Best regards,


On Sat, 9 Jun 2018 18:14:17 +0000, Jonathan Rosenne via Unicode wrote:
> Translated error messages are a horror story. Often I have to play around with my locale settings to avoid them.
> Using computer translation on programming error messages is no way near to being useful.
> Best Regards,
> Jonathan Rosenne
> From: Unicode [mailto:unicode-bounces at] On Behalf Of Philippe Verdy via Unicode
> Sent: Saturday, June 09, 2018 7:49 PM
> To: Marcel Schneider
> Cc: UnicodeMailingList
> Subject: Re: The Unicode Standard and ISO


2018-06-09 17:22 GMT+02:00 Marcel Schneider via Unicode :
On Sat, 9 Jun 2018 09:47:01 +0100, Richard Wordingham via Unicode wrote:
> > 
> > On Sat, 9 Jun 2018 08:23:33 +0200 (CEST)
> > Marcel Schneider via Unicode  wrote:
> > 
> > > > Where there is opportunity for productive sync and merging with is
> > > > glibc. We have had some discussions, but more needs to be done-
> > > > especially a lot of tooling work. Currently many bug reports are
> > > > duplicated between glibc and cldr, a sort of manual
> > > > synchronization. Help wanted here.  
> > > 
> > > Noted. For my part, sadly for C libraries I’m unlikely to be of any
> > > help.
> > 
> > I wonder how much of that comes under the sad category of "better not
> > translated". If an English speaker has to resort to search engines to
> > understand, let alone fix, a reported problem, it may be better for a
> > non-English speaker to search for the error message in English, and then
> > with luck he may find a solution he can understand.
> Then adding a "Display in English" button in the message box is best practice.
> Still I’ve never encountered any yet, and I guess this is because such a facility 
> would be understood as an admission that up to now, i18n is partly a failure.


- Navigate any page on the web in another language than yours, with a Google Translate plugin enabled on your browser. you'll have the choice of seeing 
the automatic translation or the original.


- Many websites that have pages proposed in multiple languages offers such buttons to select the language you want to see (and not necesarily falling 
back to English, becausse the original may as well be in another language and English is an approximate translation, notably for sites in Asia, Africa and 
south America).


- Even the official websites of the European Union (or EEA) offers such choice (but at least the available translations are correctly reviewed for European 
languages; not all pages are translated in all official languages of member countries, but this is the case for most pages intended to be read by the 
general public, while pages about ongoing works, or technical reports for specialists, or recent legal decisions may not be translated except in a few 
"working languages", generally English, German, and French, sometimes Italian, the 4 languages spoken officially in multiple countries in the EEA 
including at least one in the European Union).


So it's not a "failure" but a feature to be able to select the language, and to know when a proposed translation is fully or partly automated.

More information about the Unicode mailing list