Re: Part III - pdQscan - A Working Example

> 3. It seems to have two kind of copies, one forward and one backward ?

That's it, each copy makes two more.

>
> Does it do anything else ?

Some years back there was an attempt to make papers resistant to spl-
wiping with a strategy called 'wipe uncle' which performed a dat-wipe
of a copy one generation back.  It was not very successful.  The style
that works is to use a mov-wipe on a twin, where the mov statements
are self-sup****ting like imps.

Initially four copies are launched, making four descendent branches
capable of rapid doubling.  The branches are separated by 2000, and
this will stay true for a long time.  The main danger to a replicator
is being spl-wiped by a scanner, pdQscan defends itself by doing its
own wipe of a brother branch at distance 6000 using a 'mov'
statement.  The mov's are designed to be self-sup****ting as well,
somewhat like imps.  pdQscan is pretty resistant to scanners that only
do a single pass through core and then go to a dat-clear like a lot of
The Machine and Zooom! derivatives.  If pdQscan is still replicating
when the scanner switches into dat-clear mode it will almost certainly
win.

Try searching rec.games.corewar for 'wipe uncle' or .2by4k'.

P. Kline