This is TAN to the principal discussion on uncertainty, but i would just 
like to check that I understand - or don't understand what Tim Finney wrote:

Tim Finney skrev  14.07.2005 23:19:

>(8) <abc> The letters are added by the editor
>
><choice><sic>klmxyz</sic><corr>klmabcxyz</corr></choice> (6.4.1)
>
>(9) {abc} The letters are regarded as mistaken and rejected by the
>editor
>
><choice><sic>klmabcxyz</sic><corr>klmxyz</corr></choice> (6.4.1)
>
>  
>
As far as I can see it would be difficult (or in cases maybe even 
impossible) to use this encoding if you wish to keep the information on 
exactly which letters have been added by the editor in case 8, and which 
have been deleted in case 9. Maybe we should have something like:

(8) <choice><sic>klmxyz</sic><corr>klm<edadd>abc</edadd>xyz</corr></choice>

and

(9) 
<choice><sic>klmabcxyz</sic><corr>klm<eddel>abc</eddel>xyz</corr></choice>

Espen Ore