On 21/02/2007, at 6:39 AM, Daniel Stenning wrote:

> Is implementing this in RB too slow ?
>
long list of heavy data crunching in Python elided...

1) I don't get the choice - Python is our language of choice, albeit  
often calling into FORTRAN. C or C++ backends. Anything as  
proprietary as RB would not get a lookin (OK, weird corner-cases like  
Matlab sneak into some projects).

2) whilst it has some good points, the RB IDE isn't up to the  
standard of the Komodo environment I'm using, as a pure programming  
IDE. We also have people using Eclipse, vi, emacs and several other  
editing environments of choice.

3) I'd be flat-out amazed if the RB dictionary implementations, in  
particular, were as fast as CPython (just to be pedantic - there's  
also Jython that runs in the Java VM and the MS-sponsored IronPython  
for .Net). The entire runtime object model of Python is basically a  
cascading set of dictionaries of properties and functions that have  
had far more people than on the RS engineering team working solidly  
on improving them for many years.

4) Without going into lots of detail, for more declarative styles of  
programming and data crunching, Python is a better language than RB.  
Lookup 'list comprehensions' and generator functions for starters.
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>

Search the archives:
<http://support.realsoftware.com/listarchives/lists.html>

Reply via email to