Date:         Thu, 27 Dec 2007 09:41:57 -0500
Reply-To:     Allan Streib <streib@CS.INDIANA.EDU>
Sender:       Vanagon Mailing List <vanagon@gerry.vanagon.com>
From:         Allan Streib <streib@CS.INDIANA.EDU>
Subject:      Re: An interesting repair sequence
Comments: To: James Henry <jch@JCHGIS.COM>
In-Reply-To:  <3EC4A6AE3116436E96CC915857829F02@jhenryPC> (James Henry's
              message of "Fri, 21 Dec 2007 16:36:32 -0500")
Content-Type: text/plain; charset=us-ascii

James Henry <jch@JCHGIS.COM> writes:

> Question: does the computer on a 1.9 "learn" and set up a "standard"
> state that it then adjusts from so that it seems to run better over
> time? I have heard of this happening on other cars, and would have
> the evolution of better behavior as driven more miles.

I doubt it.  We're talking about technology that's 20+ years old.  All
that the computer is doing is lookups in ROM indexed by a few
monitored inputs (O2 sensor, throttle, RPM, temperature, etc) to find
the correct operating parameters for the current conditions.

I think the effect you are describing is actually the reverse, the
owner "learns" the behavior of the car and adjusts his driving
behavior to get optimal performance within those limits.

A brand new engine will tend to run a little better after several
thousand miles, that's not the computer learning anything it's the
seals and rings getting properly seated and commonly called
"break-in".  Conversely an old gunked up engine can sometimes find new
life if it is driven aggressively for a while (the so-called "italian
tune up") but things have to be basically in good order for that to
work.  It sometimes happens when there is a change in drivers and the
former driver was conservative and the new driver is aggressive.

Allan