Talk:Forget / Keep Mechanism

Ouch, this seems to produce many problems. Specially the introduction of all these parameters for types. Also, I am not sure that this approach is sound and modular. It also looks difficult to teach. --Schoelle 15:45, 21 February 2007 (CET)

The currently proposed way of the ECMA standard also introduces a parameter for types (the question mark for detached types) as well as introducing new problems without solving the issue for generics (see here New_CAT_call). It would probably be easier to implement, understand and teach one conecpt to solve covariant feature redefintion, restriction of export status and safe generics than to introduce different ways for handling each case. --Juliant 20:28, 21 February 2007 (CET)

Try ECMA instead - it eliminates catcalls. --Colin-adams 14:48, 21 February 2007 (CET)

The ECMA meeting last week discussed the problems of the ECMA standard (see here New_CAT_call) and decided to have a look at another way of preventing cat-calls. --Juliant 20:28, 21 February 2007 (CET)