regex/fsa theory

[Bob Ham]

Hi all,

I'm still chugging away on my regex rewrite uni project (which has been
stripped down to a much smaller grammar, so I doubt it'll ever be useful
for glibc,) but I've come accross a bit of a problem, which none of my
textbooks seem to want to answer.  Seeing as there are other people
doing regex stuff here (and likely other people that would know anyway,)
I though I'd ask here:

With the transition function for an nfa, does anybody know if:

  delta(r,s) = x + epsilon_closure(x)

or should it be

  delta(r,s) = x

assuming that delta returns a set of states?

Cheers,

Bob

-- 
Bob Ham: address@hidden  http://pkl.net/~node/

My music: http://mp3.com/obelisk_uk
GNU Hurd: http://hurd.gnu.org/