[noise] Pattern transformations

[Alex]

On Mon, 20 Mar 2017 05:38:45 +1000
Rhys Weatherley <rhys.weatherley at gmail.com> wrote:

> On Mon, Mar 20, 2017 at 5:27 AM, Alex <alex at centromere.net> wrote:
> 
> > I'm a little unclear on the idea of pattern transformations. Up
> > until this point the transformations I've seen have only been
> > applied to one or two patterns in an ad-hoc manner (e.g. fallback).
> > It's not clear that any single formal algorithmic transformation
> > can be applied universally to all patterns.
> >
> > In theory, could such a transformation exist? If it can, that would
> > be incredibly beautiful and elegant, because it would allow us to
> > have a notion of composition (among other things).
> >  
> 
> The "hfs" transformation for Hybrid Forward Secrecy [1] is
> algorithmic: replace e with e,f and ee with ee,ff.
> 

In this case you're proposing a protocol extension, which is something
far greater than a simple transformation -- you're adding primitives
like GENERATE_KEYPAIR_F, etc.

In section two, you describe how protocol naming is affected by the
extension. Without some rigorous definitions for protocol naming, we're
going to find that transformation composition is impossible, since the
way protocol names are modified is ad-hoc.

-- 
Alex