[darcs-devel] [patch1379] removed Darcs.Patch.Prim.V3 from unit tests

[Ben Franksen]

Ganesh Sittampalam <ganesh at earth.li> added the comment:
> I think if the V3 code itself needs any maintenance we should also
> consider removing that from darcs for the time being, as it's not being
> actively worked on.

Who wrote this code anyway and for what reason? Is it based on the Camp 
patch theory?

Unless it becomes a significant maintenance burden I think the code should 
be kept in Darcs. It's not as if we had anything to replace it and if one 
day someone starts to work on this again they can use the existing code as a 
starting point.

BTW I have been thinking a lot about patch theory during the last few 
months. I have even written up some math, using semigroup theory to simplify 
and streamline your permutivity proof and perhaps weaken the (IMO) rather 
strong assumption you named "consistency of failure". However, I got stuck 
when I could not prove a certain property(*) which I appears to be closely 
related to "consistency of failure". So there is really no big win here, 
besides possibly gaining a bit more insight into the structure of the 
problem.

Cheers
Ben

(*) In a few words: If P is the set of all patches and n>=2, we look at the 
inverse semigroup of partial functions on P^n generated by commuting 
adjacent patches in an n-element patch sequence. The property I need is that 
if y and x.z are idempotents (=partial identities on P^n), then x.y.z is 
idempotent, too. (Note that the "." operator is composition of _partial_ 
functions, here.) I haven't proved it formally (yet) but I am fairly certain 
this property is equivalent to saying that the inverse semigroup in question 
is E-unitary.
-- 
"Make it so they have to reboot after every typo." ― Scott Adams