Re: A Crime by a Board of Old Imposters

David Richerby wrote:

> > I think that is a strange position to take. I feel that the GUI is also 
> > just an external I/O device, like the human operator

> What you're saying is true but impractical.  I agree with you that the GUI
> is a distinct piece of software that relays status from the engine, which
> plays chess, to the operator, who interfaces with the world.  However, if
> the rules are to be based on whether the engine or the GUI claimed a draw,
> for example, there must be a clear distinction between the two.

I think it would be a great idea to formalize this distinction in the 
rules, by defining a well-designed mandatory protocol for inter-computer 
games. You could build a GUI on top of that; the litmus test whether 
your protocol definition is OK would be to see if you can make a GUI 
which has essentially *no* knowledge of chess other than that it is 
played on an 8x8 board with 13 possible types of status per field.

> tournament organizers must be able to tell instantly which actions were
> the engine's and which the GUI's -- otherwise, it would be very hard to
> disprove a claim that a particular bad move was really chosen by the
> engine and not a bug in the GUI.

It's a mess. I truly believe that a good protocol would help this, 
together with an open source GUI that speaks this protocol (which could 
be mandated for cc-games). I'm going to think about this for a bit.

> So, while distinguishing between the GUI and the engine sounds like a
> good idea, it is not possible to police and
> would, therefore, make a bad rule.

The hurdle is a technical one. There should be no mix-ups between the 
GUI and the engine, and I almost feel like proving this point by taking 
this project on.

Best regards,

  Sidney