Re: Don't rewrite your go program!

["Jon Diamond" <jdiamond@xxxxxxxxxxxxxx>]

But after a time (unspecified) any program becomes essentially
unmaintainable/changeable/improvable at reasonable cost (effort). Therefore
there will always come a time at which a rewrite is really necessary to
proceed further.

Of course, if your program has taken a long time in development originally
then the rewrite time to get back to the same level may be longer than you
originally estimated, perhaps even an order of magnitude longer.... The
world is full of unsuccessful rewrite projects of application systems which
have been rewritten over a number of years. Most of these have never been
publicised for obvious reasons, but they exist in many commercial
organisations not just the governmental ones you might expect.

Perhaps the people mentioned got discouraged by this elongated timescale and
just stopped redevelopment? Perhaps there just didn't appear to be
sufficient incentive to complete the task?

----- Original Message -----
From: David Fotland <fotland@xxxxxxxxxxxxxxxxx>
To: <computer-go@xxxxxxxxxxxxxxxxx>
Sent: Monday, September 06, 1999 2:58 AM
Subject: Don't rewrite your go program!


>
> It seems that rewriting a go program is a sure way to fall out of
> the top group, perhaps forever.  Look at the programs that have
> been rewritten:
>
> Alan Scarf/Microgo - new program rumored, but never competed again
> Bruce Wilcox/Nemesis - new program competed once, and finished poorly
> Mark Boon/Goliath - new program never completed
> Chen Zhixing/handtalk - new program weaker than old one
> Ken Chen/Go Intellect - new program weaker than old one
>
> It seems much better to continue improving what you have than
> to start over, however much you want to.
>
> -David
>
>
>
>