[Baypiggies] Is it problematic to embed python in a Qt C++ app?

[William Deegan]

On 1/15/07, Tony Cappellini <cappy2112 at gmail.com> wrote:
> Do any of you remember Jimmy Retzlaff's presentation (at Stanford) on the
> python/wxPython app he wrote for his brother's company?
>
> There is quite a bit of fast graphics manipulation going on in that app, and
> it was all written in Python ( Jimmy, correct me if I'm wrong). He had even
> illustrated the fact that earlier cuts of the program were prohibitively
> slow, until he optimized the critical areas.
>
> Drew Pertulla's video clip organizer app (sorry, I forgot the name of it),
> is another fine example of Python not being the bottleneck in a
> high-performance situation.
>
> It would be great is examples like these would proliferate to help disspell
> the myth that Python is too slow for serious development.

To be fair this is true of TCL, Java, and just about any
interpretted/JiT'd/P-codish language.

A company I worked at in 97 used tcl/tk for its EDA gui ( think
polygons representing each an every layer/transister/via/etc in a
chip, many many many many polygons).  The company was acquired and the
new parent company decided TCL/TK was too slow.

Then our GUI guy demo'd his mostly TCL/TK (calling c++ to traverse the
database) GUI and the result was those espousing a pure c++/C/X gui
were publically embarassed...

Same story for a Java based gui at another company.. the question was
c++/QT vs Java,   with the Java product customers ask if the demo is
canned because they don't beleive such a gui would be able to update
as fast as it does..

Most times the issue is the algorithm, lack of caching, and programmer
skill.  Computers are soo much faster than the human eye at this
point. (And have been for some time.)

-Bill