Python Productivity over C++

[Gordon McMillan]

Cameron Laird wrote: 

>On the other hand, I'm still searching for the best
>way to speak about the "well you Python people only
>find those errors at run-time" objection.  I find
>it a challenge to communicate effectively to
>newcomers how much better it is to have a language
>that encourages unit testing as Python does. 

Sometimes you can shock them a bit by claiming they're using the compiler 
as a crutch. When the tools are that different, you need to change working 
styles; something the static-typing dependent don't seem to consider.

>Python's developmental productivity is so much
>greater, in my experience, as also to defy easy
>explanation for those who haven't already used it.

Interesting, though, that this thread (at least as much of it as I've 
tracked) talks about the libs and the syntactic simplicity. Good stuff, no 
question; but not order of magnitude stuff.

One point worth reiterating is that where the app (in a C++ design) calls 
for lots of related classes, you will probably only have a handful in 
Python. Of course, this means discarding classical "OOP" thinking, and 
desinging more in the "component" style. But this can yield an order of 
magnitude all by itself.

The real killer is when you are working on something you don't have a model 
for. Long before you've completed your UML for your best guess of how to do 
it in C++, you can be on your 4th or 5th functional prototype in Python. In 
other words, almost done.

In fact (as I'm sure you've heard me say before), even if I have to deliver 
in C / C++ or Java, I write it first in Python. The few times in the last 3 
years that I've skipped that step, I've regretted it.

- Gordon