[plt-scheme] Scheme productivity data point

On Nov 9, 2009, at 12:27 AM, Neil Van Dyke wrote:

> Noel Welsh wrote at 11/08/2009 03:43 PM:
>> The issue is total time to complete the job. Neil is saying that w/  
>> Scheme
>>
> What Noel said.
>
> Disregard the non sequitur about the about the size of the output  
> files; that was quoted to help support the validity of my file I/O  
> comparisons, but then I decided I didn't need to support those in a  
> blog post, yet accidentally neglected to remove that part from the  
> quote.
>
> The nature of the work done, however, is relevant, since not all  
> tools are necessarily good at all tasks.  PLT Scheme happens to be a  
> good general-purpose toolset, IMHO, but my anecdote alone doesn't go  
> far to supporting that.

1) Clearly, this argument applies to *any* high-level language with  
good libraries and low coding overhead.

2) I think it would be illuminating to find the crossover point.  That  
is: how many times would you have to run this job to make the C++  
version better?  That depends on the size of the job itself, but if it  
took (total guess) 1 minute in C++ and 2 minutes in scheme, and you  
assume that the computer can run 24 hours a day after the code is  
written, then you have:

2 days + n*(2 minutes) = 5 days + n*(1 minute)

looks like about 4,320. How many times did you actually run the program?

Obviously, there are a bunch of guesses in here.

John

-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/pkcs7-signature
Size: 2484 bytes
Desc: not available
URL: <http://lists.racket-lang.org/users/archive/attachments/20091110/7081e10b/attachment.p7s>