RE: [xliff-inline] Reference to opening code

[Yves Savourel <ysavourel@translate.com>]

Hi Christian,

> I am not sure that I understood all of them. I am especially 
> unclear about the "distinct ids for opening and closing codes"
> dimension.
> You - amongst others - suggest to go from
> <sc id='1'/> some text <ec id='2' rid='1'/>
> to
> <sc id='o1'/> some text <ec id='c1'/>
> Is this correct?

Not quite.

What I was trying to say is that I don't see the need for an attribute in the closing code that points to its corresponding opening one. The knowledge that a given opening code and a given closing code are together can be effectively carried by having the same id value.

XLIFF should not dictate what the value is, only whether it's unique or not, etc.

I would propose: <sc id='1'/> some text <ec id='1'/>

Now, you are going to ask me, what about uniqueness of the id values?
Well, uniqueness is very relative. We need uniqueness within the source or the target content to identify a code. Take "<pc id='1'> some text </pc>", the paired representation of the same thing. In that notation we apparently do not need a separate id for the closing part. Why do we need one when we unique distinct elements? The element type itself tells you if it's the opening or the closing part.

My example with "o1" and 'c1" is simply an illustration of what some internal id could look like, if an implementation would need to associate a true unique id to objects holding code information.

I haven't thought over all the implications of using such mechanism yet, but the few tests I've done so far don't show any problem with it, that's why I'd like to see other input on this.

I hope this clarify my thoughts.
-ys