On 08/22/2013 05:42 AM, Sebastian Kranz wrote:
Dear Liviu,
thanks for your email. After Pavel's reply I took a look at that
function and the MathExtern.cpp code. I agree that an improved
MathExtern.cpp would probably the best solution. However, I have
basically no c++ skills, so I fear I am not well fitted to contribute
along that route.
I am still not sure about one thing though: Given that parsing Latex
is mainly heuristic, it may turn out that it is substantially easier
to develop an external Lyx-To-Maxima GUI written in some scripting
language like R or Python instead of extending the direct Lyx
integration in C++. It just seems to me that a try-and-error search
for a good heurstic can be performed much quicker in a scripting
language. Furthermore, an error in the GUI would not crash Lyx
completetly, but just the GUI, which one can restart without loosing
Lyx data. Hence, one could be much more relaxed in using a version
that just works in 98% of cases. Having some more Lyx commands, like
"copy as Latex" would make such external GUI easier to use, however.
Anyway, I guess I will continue working a bit on my "R" solution and
if it some day is stable enough and adds a worthwhile benefits to the
current Lyx CAS abilities, I will write a decent documentation so that
others can try it out.
Obviously, you can pursue whatever path you wish. But you would need
only VERY minimal C++ skills to work on the Maxima output.
Richard
