From:	Mike Elkins <MikeE@*********.COM>
Subject:	Re: VR 2.0
Date:	Fri, 29 May 1998 13:20:13 -0500

>Given good programming techniques, it seems reasonable to me that
>one should be able to 'upgrade' existing programs that you own.


If you have the source code, sure. But, speaking from a decade of
experience, there are a lot of times when it just isn't worth it. Here's
how I would model it: upgrading a program can save you a lot of time by
reusing the design points you've already got, but it is inherently more
difficult than when you have a free rein to implement everything exactly
how you want it. Every existing program has a "maintainability rating",
commonly called cruftieness. When you upgrade a program, you don't
need to redo the old design points, but you must add the cruftieness
rating to the TN for your programming test.

Program start with a cruftieness rating of 1.
If you didn't write it, +2 cruftieness.
Each previous upgrade: +1 cruftieness.
(Optional) every two failing dice of previous programming tests: +1
cruftieness

reducing a program's cruftieness requires a programming task with a
target number of 5 and design points equal to 10% of the existing
program per point of cruftieness.

These rules aren't play tested, but something like this may be what you
want. I'd apply them to SOTA upgrades too.

Double-Domed Mike
--Dictated using Dragon NaturallySpeaking, look Ma, no hands!