[TO]What's the big deal with EJB? [Re: PEP scepticism]

[Alex Martelli]

"Nicola Musatti" <objectway at divalsim.it> wrote in message
news:3B45C48D.413A5E at divalsim.it...
> Ciao, Alex.
>
> Alex Martelli wrote:
> [...]
> > Haven't done *identical* projects in C++ and Python yet, except
> > for Python extensions prototyped in Python then coded with Boost
> > Python, and for those (small extensions) it seems I'm about 4
> > times more productive in Python than in C++.  I have no personal
> > experience doing the same task in Python and Java, yet.
>
> How do you measure your productivity?

For a given task (sometimes estimated, in turn, in function points,
but that's irrelevant here, since we're not comparing different
tasks!), productivity = K / (number of hours I spend on the task),
for some constant K.  The number-of-hours part is easy for tasks I
perform for work since I'm supposed to report on that, therefore I
measure it (as Senior SW Consultant I'm a "shared resource" for
the firm, so my time is debited to whatever projects have used it;
there's no formal concept of "billable hours" but it comes close).

As I reported on other threads, I've done identical tasks in C++
and Java, and other identicals tasks in Python, C++, and C, so I
do have a reasonable measurement of what happens to my own
productivity dependin on language.  Of course, the _second_ time
I program the same things I may well be more productive because
of learning effects, and I have not controlled for that -- the 15/20%
advantage for Java over C++ may partly be due to learning, and
the 4-to-1 advantage for Python over C++ may be actually higher
since in that case I did the Python first.  (And it's in one specific
field -- extending Python; in another, writing simple COM servers,
Python and C++/ATL took me the same time when I tried [but the
servers had SIMPLE logic indeed in that case]).


Alex