difference of interpreter usage

Hi,

I created a Qt-prototype to process result of interpreter. In my java-app I use the interpreter as external executable called via thread/process ...
As Axis integrates the interpreter into the executable, I tried that variant too.

What I like very much is the line prefix of sequence and N-word when calling interpreter externally. That variant is easier to build (no problems with nesting dependencies in linuxcnc), but is much slower. But it offers full control over parsing results.
Internal interpreter uses a lot of c-functions, that don't have any relation with interpreter, so the only usable information is provided by function arguments.

What I've seen so far - processing of nurbs_feed is possible with internal interpreter usage only. External interface does not provide enuf informations ...

What about unifying both variants?

As a test, I added linenumbers as n-words and result is, that sequence has finer resolution than linenumbers. Some GCodes have one linenumber but result in various sequence numbers. So one of both numbers got lost with internal usage.

Any thoughts/comments are welcome