speed/efficiency of INCLUDERC vs. actual recipe?

Message has been automatically signed by PGPsendmail 1.4,
available from ftp.atnf.csiro.au:pub/people/rgooch


-----BEGIN PGP SIGNED MESSAGE-----

My .procmailrc and all its includes are getting awfully large and I'm afraid
that, for the amount of mail I receive, it may be struggling to keep up.
So, in my goal towards higher efficiency, I ask this question:

Is it faster to put a INCLUDERC where I'd otherwise put the actual contents
of the recipe?

I.e., is this:

:0
*^^TOlow-traffic(_at_)mailinglist(_dot_)com
{
INCLUDERC=foo.proc
}

faster than:

:0
*^TOlow-traffic(_at_)mailinglist(_dot_)com
{
        :0:
        *^subject: .*blah
        folder1
        :E
        /dev/null
}

for the case where the message is _not_ to 
low-traffic(_at_)mailinglist(_dot_)com?

(You understand, the recipe between the {}'s is much, much longer and
hairier.)

Thanks for any hints.

P.S. Maybe a good grad student project: a procmail compiler where a gigantic
     DFA could be built from all the recipes.  This might speed things up...

- -- 
mailto:dummy(_at_)cyberpass(_dot_)net                             In sunny 
Berkeley, CA
http://www.cyberpass.net/~dummy                        waiting for The Big One
     >> Get used to it: breeding time is over! <<       
       >> He who fails to learn Unix is lost. <<       (510) 464-4604


-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQB1AwUBMzsUsNCFOCNWKXb9AQHOpAL/RECbqabTFVnIHZ4T+EvCpOysnZ8ceMdE
4AhZ/6FCvA2vw/sK+k1Ryk8aimXvnXMiR1Vf1p7oBbLSsB9ugdZxzmu8h7WjtERT
gn/20ROjRO3osli7Z6HPyU/ZtXFS+kgg
=Tmq7
-----END PGP SIGNATURE-----