[FRPythoneers] In search of a member...
fschmid at archenergy.com
Wed Jul 18 11:17:11 MDT 2001
Sean Reifschneider wrote:
> On Tue, Jul 17, 2001 at 05:40:07PM -0600, Quent wrote:
> >It seems the cost/benefit trade-off people are making is between
> >execution performance and time and ease of development, so they're
> The point I like to make is that you might as well use a high-level
> language like Python to develo the code quickly, then once you have the
> prototype or proof of concept working you can do profiling if the code is
> too slow. Instead of guessing where the performance issues are, get some
> performance numbers and concentrate on the hot-spots...
> If you don't guess where it's slow, you won't guess wrong...
I agree with Sean's point of making sure analyzing performance rather
than guessing. Think about what your code does and where it spends
time. Take the speed of your storage and potential network connection
into account when you do this. Then you can test code snippets...
However I don't think writing Java code is significantly slower than
writing Python code. I do both and find that depending on the
application either environment may be fit. Some Java VMs show
outstanding performance (speed, reliability) and there are systems out
there that run java natively (TINI computer and others for $50 and up
http://www.ibutton.com/TINI/hardware ). Python on the other hand has a
smaller footprint than Java on systems that need a VM.
Please don't take this as a product endorsement, just food for thought.
More information about the FRPythoneers