Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!swrinde!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!ira.uka.de!fauern!faui43.informatik.uni-erlangen.de!richter From: richter@immd4.informatik.uni-erlangen.de (Joachim Richter) Newsgroups: comp.unix.wizards Subject: Re: Hacking and "Amateurism" Message-ID: <1991Apr16.121218.6603@informatik.uni-erlangen.de> Date: 16 Apr 91 12:12:18 GMT References: <1991Mar26.015635.23103@mintaka.lcs.mit.edu> <1991Mar26.163720.28379@en.ecn.purdue.edu> <1991Mar27.041126.9886@news.miami.edu> <1916@hpwala.wal.hp.com> Reply-To: richter@medusa.UUCP () Organization: CSD., University of Erlangen, Germany Lines: 23 In article jason@cs.odu.edu (Jason "dedos" Austin) writes: > > It's not too hard to show that it is possible to decode a >password. Every time the same salt and the same password is run >through the crypt function, the same code comes out. >(It would have to >or the thing wouldn't work at all) At the worst case, an exhaustive >table from coded to decoded passwords woul; give right answers. Even >if the relation is not 1-1 and each code has more than one possible >decoding, any of the valid decodings would let you log in. Of course, >this would be quite a large table to calculate considering all the >permutations. >-- >Jason C. Austin >jason@cs.odu.edu Right. Of course you can do that. You can also, for a given object file, find the source code, that, when compiled, gives that object code that way. No problem. Maybe it will take a while - some millions of years or so. But, since the number of source codes, that give the same object code is infinite, your chance is not so bad :-)