[identifiable e-mail addresses have been deleted to frustrate mining by spammers]

From: William L Anderson
Sent: Sunday, May 23, 2004 10:58
Subject: Re: Num.java
Security Label: Signed



Dennis E. Hamilton wrote:
> Wow!
> 
> OK,
> 
> 1.	We have moved from comments to e-mail.  That is interesting.

I didn't see an easy way of keeping the code separate and easily 
manipulated. I also thought a move to a wiki would be better, but 
wanted to wait for the compilable interface first.

> 
> 2.	Although my note was intentionally crude, I see that our automatic
> induction of meaning has already led to something interesting.  I had not
> omitted the result type of pred() and next(), although I can see how that
> could be seen as a mistake by me.
> 
> 3.	Here is an update based on your version.

I thought about saying public Num next(); but my choice of int 
reflects my own imperfect understanding of Java, and maybe OOP.
> 
> 3.1	It is interesting to wonder whether there is some advantage in
> changing the names of pred() and next() to something else, although that
> has only a temporary advantage.

Agile methods would counsel to use full words like "successor" and 
"predecessor".
> 
> 3.2	Likewise, I am not happy about isOrigin versus is0, although changing
> the name is useful.  I am not certain about "origin" is all.

Yeah, this is very interesting. I thought I had a clue about what 
is0() was about. isOrigin() seems confusing to me in that it makes 
me think this is more complicated than I had thought.

This is a very interesting example of assumed contexts and meanings.