On Thu, 7 Oct 2010, Lukas Bulwahn wrote:

The code_pred command for the invocation of the predicate compiler could also fit under this umbrella.

So the syntax would change from "code_pred" to "prove code_pred" which is actually better, because then users are not surprised that the command sets up some goal state (in most cases the empty goal) for the user to prove.

This scheme would be outside of Isar traditions for several reasons, such as the naming of toplevel commands (nouns, not verbs).

BTW, the very idea of an empty goal state shown to the user belongs to the PG interaction model. It is easy to imagine an Isar prover IDE that fills such holes seemlessly according to the formal structure of the text. The traditional mismatch of PG interaction and Isar structure may soon be overcome in the new document model.


        Makarius
_______________________________________________
Isabelle-dev mailing list
[email protected]
https://mailmanbroy.informatik.tu-muenchen.de/mailman/listinfo/isabelle-dev

Reply via email to