Re: Accentuated characters in references (now very OT)

[Lane Sharman <la...@san.rr.com>]

"Guillaume Mathé" wrote:

> Original message from: Lane Sharman
> [...]
> (something i forgot about your first reply: yes i see your point, it's nice and clean, but that's 1 label per object. Here we have 15+ labels per object... Not
> counting Servlet-specific labels, site browsing labels, etc.... Moreover it's written to use servlets+velocity as well as swing+swing for client-server
> operation. Still, i welcome any other idea)

Generally, my approach for localization is to make each attribute of an object localizable. Further, the collection of the attributes is what gets proxied across
the network to the swing, velocity and database services. java.lang.Object and reflection are inadequate or subject to misinterpretation. Thus, any
viewable/storable object carries with it in my universe a map of FieldData and MetaData. I documented this because I thought in case I got hit by a truck, my team
would at least understand where I was coming from:

http://engineering.acctiva.org/jldap/

Guillaume, I tried to write templates so that, when undergoing translation, the variables were immutable as to spelling, accents, etc and that the literal content
was the only content that required to be changed. It seems that you have implied that a template writer in France would write $Client_Prénom and the english
translator would be free to write $Client_First_Name. I do not do things like this but on thinking about it, it certainly makes the templates and text documents
more readable in the native language of the translator.

This makes the keys localized as well as the values. I would still abandon the use of ResourceBundle and use a direct dictionary scheme for the keys as well as
the values.

-lane

Regards,

> Guillaume Mathé

--
Lane Sharman
http://corporate.acctiva.com/lane