I worked in Italy and I can't say everyone was over joyed with being
required to work in English; however,
I was quite surprised by how many said one of the benefits of this job was
the chance to improve
their English. Most also realized that this was not a just "American's
forcing their way on them attitude": an
Italian support person opening an APAR in Italian was not very likely to
see a quick fix from a component
whose developer is the Tokyo research lab. They realized the fact was
just that English was the common
denominator for the company.
I have never played around with this aspect of LOG4J but it has categories
which is what is usually what
you set with the classname. In theory however this could be set to
anything and there is no requirement
to have one per class. It also possible to filter by these categories too.
"Assuming" what I have just said
is true why not have a category for NLS enabled messages like :
protected static Log nlslog =
LogFactory.getLog("NLS"+ Message.class.getName());
protected static Log log =
LogFactory.getLog( Message.class.getName());
Message that might be a benefit to an end customer could use nlslog.debug
(...) that could be translated
and those message that are just for devloper debug tracing: log.debug
("English only here");
This is only going on assumptions I read a while back when looking at
LOG4J.
Rick Rineholt
"The truth is out there... All you need is a better search engine!"
[EMAIL PROTECTED]
Russell Butek/Austin/IBM@IBMUS on 03/15/2002 09:48:28 AM
Please respond to [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
cc:
Subject: Re: RE: cvs commit:
xml-axis/java/src/org/apache/axis/wsdl/toJava Jav
aComplexTypeWriter.java JavaStubWriter.java
Curious. We have different viewpoints. I worked in Germany for a few
months and the IBMers there were a bit upset about the English-centric view
that was forced on them.
We disagree, obviously, but we can decide on a future direction for the
debug statements. If folks think log.debug strings should stay English,
I'll refrain from i18ning them in the future. But ONLY those strings.
Russell Butek
[EMAIL PROTECTED]
Rick Rineholt/Raleigh/IBM@IBMUS on 03/15/2002 08:18:38 AM
Please respond to [EMAIL PROTECTED]
To: [EMAIL PROTECTED]
cc:
Subject: Re: RE: cvs commit:
xml-axis/java/src/org/apache/axis/wsdl/toJava Jav
aComplexTypeWriter.java JavaStubWriter.java
I agree here with Glen. I have never seen in my time in IBM where debug
trace was I18N.
I have also worked overseas and for technical positions it was understood
that English was
a requirement. I even once asked some of the developers at the time why
they didn't get
product documentation in their native language and was told that technical
translations were so
poor that it was almost useless. Besides the trace information is for the
most part useless
unless you have a reasonable understanding of the code in which case as
part of the translation
effort will the code comments be also translated? :-) It also adds
additional cumbersome when
trying to debug to have to look up in message tables to match the trace
statements with the
outputs. Which also leads to developers ultimately not putting trace
statements in whatsoever.
>I don't agree (obviously, since I keep i18ning them!). IBM's got customer
>support sites all over the world. Most of the problems they support will
>be resolved by them and not even get to a developer back here. When they
>tell their customers to turn on debug, they would really prefer to work in
>their own language.
>Russell Butek
>[EMAIL PROTECTED]
>
>Glen Daniels <[EMAIL PROTECTED]> on 03/14/2002 10:16:30 PM
>Please respond to [EMAIL PROTECTED]
>To: "'[EMAIL PROTECTED]'" <[EMAIL PROTECTED]>
>cc:
>Subject: RE: cvs commit: xml-axis/java/src/org/apache/axis/wsdl/toJava
>Jav aComplexTypeWriter.java JavaStubWriter.java
>
>
>You know, if you ask me, it's going a bit far to internationalize
debugging
>trace messages....
>
Rick Rineholt
"The truth is out there... All you need is a better search engine!"
[EMAIL PROTECTED]
