[Ktechlab-devel] The biggest technical issue

[celelibi at gmail.com]

2008/6/4, Peter Jørgensen <fifafrazer at gmail.com>:
> Hi
> Im not into the code myself, but if the whole ktechlab engine has to be
> rewritten, wouldn't it be a good idea to do it in python instead, and only
> use c++ extensions for the very speed-demanding parts. Then the application
> could easily be made cross-platform, which would make it available for a lot
> more users, and there is no free alternative to ktechlab for windows or mac.
> Just a thought.. It may be stupid to write everything all over again, but I
> can't tell, as all I know is that the engine of ktechlab should be
> rewritten, as Alan Grimes has told us.
>

Hum, well. I think ktechlab is slow enough for now. But it would need
some tests to see C++ is really faster than python. (NB : perl forever
!)
I think that very speed-demanding parts should be written in C, not C++ IMHO.
I think C++ is a good choice (not QT but that's another problem), it
could be quite cross-platform if the code have a good design.

Unfortunately I'm not really into the code but I think redesigning the
way new components are added would involve some big modifications in
the code. Therefore I think a new version would be created in order to
"rewrite" the whole code (maybe with some big copy/past carefully read
three time).

But as I said, I don't know how big this challenge is because I'm not
into the code.
I think that if this is done, I will read every line commited to
understand how it works.


My 2 cent.


Celelibi