[Retros] Messigny 2013 results

[Kevin Begley]

When two conditions lead to ambiguous interpretations, it is necessary to
define a set of rules, and provide a unique name.

In fact, under such circumstances, it is best to avoid the popular
"condition1 + condition2" naming scheme, so as to provide clarity, with
respect to alternative interpretations.

It has been suggested that "condition1 + condition2" should be ordered, so
as to defer to the rules under the first condition listed.

It has also been suggested, quite naively, that the rules (and soundness)
depend only upon the implementation, by some version of solving software.
Obviously, that suggestion is completely bogus, and anyone claiming this is
suffering from considerable delusion (most likely, they either know very
little about software design, or have given the implications insufficient
consideration).




On Thu, Jul 4, 2013 at 2:34 PM, Joost de Heer <joost at sanguis.xs4all.nl>wrote:


> On 07/04/2013 10:56 PM, Thomas Maeder wrote:

>

>> Am 04.07.2013 21:10, schrieb Joost de Heer:

>>

>>>

>>> How does Madrasi+Maximummer work? Or Madrasi+Haan vs Maximummer+Haan?

>>>

>>

>> I'm sure that nobody has ever defined it.

>>

>> Popeye implements Madrasi+Maximummer with regular (non-Maximummer)

>> paralysis. I.e. in the position white ra1 black ra5, the rooks are

>> paralysed even if the distance to h1/h5 is longer than to the other

>> rook. Other programs may well do this differently.

>>

>> In Madrasi+Haan, observation doesn't go through holes. This seems very

>> natural, so I doubt that any program would do this differently.

>>

>> I don't get the question about Maximummer+Haan.

>>

>> All this seems unrelated to my previous post, though.

>>

>

> Case 1: With Madrasi, castling is just a kingmove. With maximummer it's

> both a king- and a rookmove. So what is it with Madrasi+Maximummer? In your

> position, queenside castling is possible under the madrasi rule, but if

> it's also a rookmove according to maximummer, then it's not legal because

> the rook is paralysed.

>

> In combination with Haan: If castling is a king-move, then only e1 should

> be a hole after castling. If castling is both a rook- and a kingmove, both

> a1 and e1 (in case of queenside castling) should be holes after castling.

> So haan would work differently in combination with Madrasi than in

> combination with maximummer.

>

> Joost

>

> ______________________________**_________________

> Retros mailing list

> Retros at janko.at

> http://www.pairlist.net/**mailman/listinfo/retros<http://www.pairlist.net/mailman/listinfo/retros>

>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.pairlist.net/pipermail/retros/attachments/20130704/28e6246a/attachment.htm>