Re: a new language, designed for safety !

[Georg Bauhaus <rm-host.bauhaus@maps.futureapps.de>]

On 18/06/14 18:39, Dmitry A. Kazakov wrote:
> On Wed, 18 Jun 2014 16:43:55 +0200, G.B. wrote:
>
>> On 18.06.14 14:30, Dmitry A. Kazakov wrote:
>>> On Wed, 18 Jun 2014 11:34:01 +0200, G.B. wrote:
>>>
>>>> (Pardon? A contract is a manifest piece from the equivalence
>>>> class of paper, signed by two parties.
>>>
>>> And that does not defines the parties it only constrains them, which was
>>> the point.
>>
>> In fact, you declared "contract" to be a
>>
>> "framework of constraints imposed on the implementations"
>>
>> which a contract between two legal parties is not.
>
> Legal parties? What are you talking about?

Contracts (the real ones) and liability.

Not about the term that has not even been hijacked by the language
designers---the LRM has "generic contract model" only---but about
those legal terms to which extended assertions (LRM) are a subordinate
means to an end: software business, and programers employed.

The use of "contract" is not just metaphorical, not just an analogy,
not just an abbreviation.
   I only wanted to point out that the reason of existence of
those extended assertions in Ada can be seen as related to liability
(via real contracts), not, however, as a definition of "contract".

In particular, whenever buyers of software want liability then they
are buyers, not logicians. A bargain means application of contract law.

If you think that the term "contract" should be defined by your
notion of impossibly exhaustive formal semantics, then it will
be difficult to see a point beyond artificial and tautologous
linguistic confusion. You'd substitute a notion (contract) with
some of its remote consequences in Ada (extended assertions).

I think it helps make things clearly defined if "contract" continues
to stand for what is has been standing for generations, and that
words form the LRM in the context of assertions be used for what they
cover, not for replacing a notion, "contract" rather than serve it.