The Coq mailing list

["Andrew McCreight" <continuation AT gmail.com>]

Hi,

I have many hypothesis of the form
  X := A

At some point in my proof new instances of A will occur.  Currently, what I do is "replace A with X in * |- *", which works, but is annoying when I have 5 or 6 definitions.  A silly percentage of some of my proofs are taken up by this.

What I would like to do is to have a tactic that will automatically do this for all definitions, but the obvious match goal to do this fails with a syntax error:

  Ltac unzeta :=
    match goal with
      | X := ?Y : _ => replace Y with X in *
    end.

Toplevel input, characters 66-68
Syntax error: ':' expected after [Prim.name] (in [match_hyps])

Is there some other way to do this?

More generally, is using definitions like this a bad idea?  Should I use equality instead?  I like having a syntactic distinction between a definition of a variable introduced purely to improve readability and more profound statements of equality, but maybe it isn't worth it if such definitions have a second-class status in the tactic language.

Thanks,

Andrew