Re: [gnugo-devel] use of LOSE in reading.c

[Gunnar Farneback]

Teun wrote:
> > Sorry, but we're quite happy to use 0 in this context.
> 
> Who is we? Doesn't include me apparently.

You're right that I should only have spoken for myself. I was thinking
of me and Dan who have done most of the work on the reading code. When
these symbolic codes were first introduced, Dan wrote a message (dated
May 25, 2001) saying:

| These changes have been discussed in this list since 
| Monday. Look for the thread ``A reading nuance.''
| 
|     Old scheme           New scheme
| 
|          1                 WIN  = 3
|          2                 KO_A = 2
|          3                 KO_B = 1
|          0                        0
| 
| There are symbolic names for the nonzero return
| codes. If failure had a symbolic name it would
| be LOSE but I thought it was better just to
| call it 0.

> What is the point of the #define LOSE in gnugo.h then?

Not much at all, as far as I can tell.

/Gunnar