A SWIG for the library documentations ?

[Borcis]

How difficult do you think would be an engine that with minimal tuning by 
"macro stylesheets" would provide an automatic equivalent in target computer 
language syntax, to the original documentation of a ported language library? 
while keeping human language as close to unchanged as possible?

First targets would obviously be the swallowing by python of the available 
TCL/TK and Java library documentations (because of tkinter and jython). Then I 
guess the chase should target javascript, by first requiring javascript 
implementations to get jvm-based if they aren't already (what's in a name ?).

Now if jvm implementations of javascript can't be had, then let the python 
church shout loud that python-jython is in fact the *real* javascript, while 
the ecmascript variant is a fake :)

Important point : the above counts as a minimalistic but non-trivial and 
moderately useful analogic model version of the challenge that really is the 
machine translation of human languages. So it should count as *appropriate 
training* in the form of a humble simplification, to the power of solving that 
harder problem !