lines of code per functional point

[Bryan]

in document, which is titled, "Java and programmer productivity", there is a
table on page 4 which shows LOC/FP for several languages. unfortunately, python
is the only language "dashed out" and has no value. i'm unable to google and
find a source that shows this value for python, but i did find other sites that
show java and c++ at 53 LOC/FP which is the same as this document. is it safe
to assume that python's value would be similar to perl's value of 21 ?

http://www.abo.fi/~kaisa/FN.pdf

thanks,

bryan

 

[Harald Armin Massa]

Bryan,

at the end of the paper there is a reference to:

http://page.mi.fu-berlin.de/~prechelt/documents/jccpp_tr.pdf

In chapter 5.6 on page 19 of this publication you can find Figure 10,
Displaying program length in comparison.

I read the graphics (looking at the yellow boxes) that most of the
python programs are quite as small as the smallest 60% of perl
programs; where "small" ist LOC.

All progs in that study should have dealt with the same problem, so
FP(perl)=FP(Python), and you should be correct with LOC/FP(Python) <=
LOC/FP(perl)

Please also see that they were evaluating Python 1.5.2, which missed
some very "density improving features" as there are esp. list
comprehenstions.

Harald