Foot in mouth disease

[GerritM]

"Derek Thomson" <derek at wedgetail.com> schreef in bericht
news:vR1y9.45$8%5.8453 at news.optus.net.au...
> GerritM wrote:
> > "Derek Thomson" <derek at wedgetail.com> schreef in bericht
> > news:3dc3dc4e$0$12762$afc38c87 at news.optusnet.com.au...
<...snip...>
> > Do you by any means have the data at hand (kloc's for Java, kloc's for
> > comparable Python; ORB and/or IDL compiler)? I am highly interested in
any
> > substantiated evidence with respect to language
efficiency/expressiveness.
>
> Nope, I was just replying based on my feelings for an average task,
> which was the tone of the OP. Think about how difficult it is to do the
> equivalent of a list slice assigment in Java, for example.
>
> For some harder numbers, I compared JacORB (a Java ORB) to Fnorb (a
> Python ORB), and Fnorb is about one half the lines of code of JacORB,
> just for the core ORB itself.
>
> Now, that's for an ORB, which is low level enough (marshalling and
> unmarshalling data on-the-wire, and connection management), that it
> doesn't play too well to Python's expressiveness. But still, that's
> quite a difference.
>
> As the application becomes higher level, I'd expect the gains to be much
> greater.
>
> --
> D.
>
I share the same expectation. It would be nice to have some more factual
comparison of larger application(s). The shootout site
http://www.bagley.org/~doug/shootout/ does a great job in comparing "small"
well defined problems. In general Python scores well in expressiveness
(=small amount of loc), but not as much as factors 2 or higher.

It would be interesting to know if our expectations are realistic for large
applications, or only hopes. I have seen many oversold technologies, which
don't pass a more factual comparison.

regards Gerrit