[sword-devel] osis2mod

[Chris Little]

Troy A. Griffitts wrote:
> The other benefit of using the basic SWORD API when processing XML is 
> that we might find bugs when using it for processing the input (which, I 
> believe DM might have found with his 3 tag example, though I always 
> thought > had to be escaped to >  Can someone check the W3C spec?)

Only & and < always have to be encoded with entities.

' and " only have to be encoded with entities inside of attribute values 
demarcated with the same type of quotation mark.

 > never has to be encoded with an entity. The only reason it is one of 
the 5 universal XML entities is for parallelism with <. Since > cannot 
occur in either element or attribute names, there's no circumstance in 
which it could be misinterpreted as the end of an element.

--Chris