Re: LSP Was: Mixing OO and DB
Date: Sat, 23 Feb 2008 19:04:57 +0100
Message-ID: <1v39xnpsb9pa0.b22wp2jffy94$.dlg_at_40tude.net>
On 23 Feb 2008 16:00:21 GMT, Stefan Ram wrote:
> "Dmitry A. Kazakov" <mailbox_at_dmitry-kazakov.de> writes:
>>"Storage" is implied by out-substitutability.
>
> »Value« and »storage« (in the sense of a memory-cell for
> a value) are well-known terms.
Storage is indeed a well-known term. But it is used (like von Neumann machine or Turing tape) in the sense very different to a higher level concept of variable. Storage as a computer component is capable to hold representations of any values regardless their types. On the other hand, consider a language implementation (and there exist such), where the compiler for efficiency reasons allocates objects of different types in unique storage pools. With this implementation neither storage will ever be able to keep anything of another. Thus, again, your example is wrong.
> According to Google, if I did not miss something, you
> are the only one using the term »out-substitutability«.
>
> It seems even more strange, when you use it - under
> these circumstances - as if the reader is supposed
> to know, what this means to you.
The meaning is quite obvious. The term was chosen in order to stress that mutability (storage even so) plays no any role here. The problem can be stated in an immutable way.
-- Regards, Dmitry A. Kazakov http://www.dmitry-kazakov.deReceived on Sat Feb 23 2008 - 19:04:57 CET