home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!haven.umd.edu!darwin.sura.net!zaphod.mps.ohio-state.edu!pacific.mps.ohio-state.edu!linac!att!ucbvax!CS.YALE.EDU!beigel-richard
- From: beigel-richard@CS.YALE.EDU (Richard Beigel)
- Newsgroups: comp.theory
- Subject: automata theory question
- Message-ID: <199209010003.AA02868@BEECH.THEORY.CS.YALE.EDU>
- Date: 1 Sep 92 19:53:52 GMT
- Sender: daemon@ucbvax.BERKELEY.EDU
- Reply-To: Richard Beigel <beigel-richard@CS.YALE.EDU>
- Lines: 11
-
- We know that DPDAs can solve every decision problem that NFAs can
- solve, because NFAs are equivalent to DFAs. The same does not appear
- to be true for function computation. Allow NFAs and DPDAs to read a
- character, write a character, or perform a stack operation during each
- step.
-
- Then an NFA can compute the function that maps xc to cx, for all x in
- {a,b}^* and all c in {a,b}. Clearly a deterministic Turing machine
- can compute this function as well. My question: Can a DPDA compute
- this function? Any other candidates for a function computable by NFAs
- but not by DPDAs?
-