LSP (was Re: Function.prototype.bind) [offtopic]

David-Sarah Hopwood david.hopwood at industrial-designers.co.uk
Wed Sep 10 18:43:27 PDT 2008


Graydon Hoare wrote:
> Mark S. Miller wrote:
> 
>> [LSP] still a good design rule ...
> 
> Yup,

As stated ("... the behavior of P is unchanged ..."), it's too strong.
It should have been stated in terms of a subtype refining the
specification of a supertype.

>> in languages (like JavaScript) with nominal subtyping (via instanceof).
> 
> Oh goodness, don't say that while the type theorists are in the room!

Indeed. ES3 has exactly one type, and it has no name.

-- 
David-Sarah Hopwood


More information about the Es-discuss mailing list