Google Summer of Code
m.dentico at virgilio.it
Fri Jun 10 10:30:54 PDT 2005
"PB" <schizophonic at tiscali.it> wrote:
> Massimo Dentico wrote:
>> Note that the Liskov Substitution Principle (LSP), that I called harshly
>> "sacred" on another mailing-list,
> What is so strange with the substitution principle,
> be it in the Liskov variant or in my granny's variant?
> It just says that, if you want to replace A with A'
> in contexts C1, C2, C3, then A' must have the same
> interface *and semantics* of A, but just in contexts
> C1, C2, C3. In C4 it can do whatever it wants - granted
> that everybody is aware of the dogma "ye shalt NEVER put
> replace A with A' in context C4". It is a "blame you,
> blame me" crystal-clear principle of sheer commonsense.
> The fact it is seldom respected - better, that is
> seldom *understood*, is sheer nonsense.
It is not so simple with pointers and OIDs mixed in.
As in my replay to Robin Green, see the series of papers
"WHAT DOES SUBSTITUTABILITY REALLY MEAN?" and the paper
about Kaleidoscope, "Constraints and Object Identity"
(I think this paper clarifies the subject, it has
nice pictures and clear explainations of the problem).
More information about the TUNES