[Retros] Fairy Retros - Codex question

[Joost de Heer]

> Dear Joost,

>

> A comment to your first reponse. You write:

> - - - - - -

> A problem stipulation has 2 classes:

> - Type

> - Constraints



> E.g. an orthodox #2 has as type 'mate' and as constraint 'orthodox'.

> A series-stalemate platzwechselcirce has as type 'series-stalemate'

> and as constraint 'platzwechselcirce'.



> For a fairy proofgame the type is proofgame, and the constraint is

> the fairy condition. I.m.o. all problems with type any of the retro

> types (proofgame, release, retractors, etc). belong to the retro

> section, and the constraint is irrelevant.



> - - - - - - - -



> If I understand 'constraint' correctly, then for a direct twomover

> with circe the constraint says that the problem is fairy chess. When

> we have a retro with circe-condition, then the constraint is

> irrelevant. Is this a correct interpretation?


Yes. In fairy proofgames, the defining characteristic for a problem is 
proofgame, not the fairy condition. For e.g. #2, this is different. All 
IMO of course.

About classification: You will always have hybrid compositions. E.g.
'Position after the 23rd move by black. #2' (Caillaud, Loustau,
Rotenberg; Phénix 10.1998). Where should this problem be printed? In the
retro column or in the direct twomovers section? Another example is
helpselfmates: neither help-play, nor selfmate. No matter what 
classification system you use, you'll always be able to find a 
composition that will fall outside the classification. A chess 
composition parallel to Gödel?

Joost