[Seaside] Re: no HTML entities for 8bit characters in UTF-8 encoding

Colin Putney cputney at wiresong.ca
Sat Jun 26 03:58:58 CEST 2004


On Jun 25, 2004, at 4:46 PM, Avi Bryant wrote:

>
> On Jun 25, 2004, at 1:35 PM, Colin Putney wrote:
>
>> A while back I read a fascinating article on localization, but I 
>> haven't been able to dig it up again. The one thing that stuck with 
>> me though is the idea that localized strings should be functions, not 
>> table lookups.
>>
>> So I suspect the right way to go about it would be to use no string 
>> literals at all in the source code, but have method calls like:
>>
>> Localization youHaveKnives: 2.
>>
>> Strings that don't need to be parameterized can be unary selectors. 
>> This lets all the different languages deal with their own 
>> idiosyncracies correctly, while still giving a clean interface to the 
>> programmer.
>
> Yeah, I think I said much the same thing when we were discussing 
> Babel.  IIRC, what I proposed then was something like
>
> String>>localized
> 	^ self
>
> Symbol>>localized
> 	^ Localization perform: self
>
> Localization>>localized
> 	self subclassResponsibility
>
> So you could still provide strings directly, or symbols to be looked 
> up, or full objects that would do arbitrary things to come up with the 
> right string.

Well, no. That's the opposite of what I suggested above.

Frankly I think that's too much flexibility. The whole idea of what I 
proposed above is *not* to have strings in the source code. If every 
user-visible string is a message to a single localization object, it's 
an easy bottleneck for translators to intercept, but still flexible 
enough to handle the vagaries of diverse languages.

Colin



More information about the Seaside mailing list