Question on XLIFF editors

Title: Message

Hi 
all,

 

First of all, please accept my apologies for my 
absence at yesterday's meeting. I had an unexpected last minute visit to another 
office and was unable to make it back in time due to adverse (to say the least!) 
weather conditions.

 

I just wanted to post a question to the group to 
get some feedback. We are currently using the Xerces parser to process XLIFF 
documents but it doesn't provide us with any feedback as how attributes have 
been encoded. As such, what comes in as single-quoted attributes (e.g. <g 
id='1'>) may end up as being double-quoted when we write it back out (<g 
id="1">). My feeling on the matter is that since both are valid XML and 
represent exactly the same thing, it really shouldn't matter if this happens. 
However, I'm concerned that someone out there might be using an XLIFF parser 
which requires one method or the other and that it will break if the quotation 
paradigm is changed.

 

Is there anything in the specification or 
anywhere else that says that an XLIFF editor shouldn't change this setting? I'd 
like to think that since an XLIFF document can go through numerous editors that 
valid XML should be the only pre-requisite but I was wondering what other 
developers thought of this.

 

Thanks for any feedback.

 

Regards

 

David Pooley
Software 
Architect
SDL International