Best practices for dealing with mnemonics in XLIFF?

[kbier@sybase.com]

Hi there,

I hope this is the right mailing list to ask these questions. 

 We have started looking into providing our 12 (well, likely more :-) 
different software resource formats as XLIFF for our vendors, which 
involves writing our own conversion for most of the formats.

A discussion has come up internally on how to best deal with mnemonics
in 
our resources ( not shortcuts but  the mnemonics like the ampersand
in 
"&File" that are turned into the underlined characters in
the GUI etc.)

We are looking at the two alternatives below. Does anyone have any 
recommendations on which representation method to choose? 
Method 1 seems to be the commonly used method, method 2 - proposed by our
engineer - seems to be rather uncommon. 

1. use the standard XML representation as character entity (&),
e.g. 

 <trans-unit id='58.1' resname='DL_RK_FILE' xml:space='preserve'>
        <source xml:lang='en'>&amp;File</source>
        <target xml:lang='de'>&amp;Datei</target>
   </trans-unit> 

2. move the mnemonic letter into a separate trans-unit, e.g.:
 
  <trans-unit id='58.1' resname='DL_RK_FILE (mnemonic)' 
xml:space='preserve'>
        <source xml:lang='en'>F</source>
        <target xml:lang='de'>D</target>
   </trans-unit> 


THANKS a lot!

Kerstin
Sybase iAnywhere