home *** CD-ROM | disk | FTP | other *** search
- Xref: sparky comp.lang.c:16059 comp.lang.c++:15852 comp.lang.pascal:6368 comp.lang.misc:3514
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!decwrl!sdd.hp.com!zaphod.mps.ohio-state.edu!uwm.edu!spool.mu.edu!agate!stanford.edu!rutgers!modus!gear!cadlab!martelli
- From: martelli@cadlab.sublink.org (Alex Martelli)
- Newsgroups: comp.lang.c,comp.lang.c++,comp.lang.pascal,comp.lang.misc
- Subject: Re: Software design =/= Programming
- Message-ID: <1992Nov04.073218.11970@cadlab.sublink.org>
- Date: 4 Nov 92 07:32:18 GMT
- References: <BwDour.J7C@cs.bham.ac.uk> <19OCT199217035699@trentu.ca> <1992Oct21.142003.7017@ucc.su.OZ.AU> <21OCT199219040925@trentu.ca> <mazz.719989080@pluto> <josef.720690811@uranium>
- Organization: CAD.LAB S.p.A., Bologna, Italia
- Lines: 38
-
- mollers.pad@sni.de (Josef Moellers) writes:
- ...
- :Although I have had some contact with good ole Goedel (I've done a
- :course in theoretical CS, I've read the book (Hofstadter), I've seen the
- :film (sorry, got carried away)), I always wondered what this
- :"unprovable" was all about.
- :When I was at school, I was taught that there existed things like
- :- axioms
- :- theorems
- :- lemmata
- :where "axioms" were the foundation of the whole, things that have to be
- :taken for granted, that cannot be proven, that "just are".
- :If You change one or more of Your axioms, You can arrive at a completely
- :different set of statements, sometimes interesting ones, sometimes wierd
- :ones. Nevertheless, "axioms" are nothing special.
- :Now comes Mr. Goedel and sais that there exist, in mathematical theory,
- :things that cannot be proven. Then I say "OK, these are axioms, so
- :what?" "Axioms", to me, are the very atoms (from greek "unseperable")
- :that make up "maths". Every now and then someone comeas along and shows
- :that one of these "atoms" can indeed be broken up, i.e. one of the
- :"axioms" is a consequence of other axioms, therefore can be proven.
-
- The point of Goedel's Theorem is, that for *ANY* formal system "powerful
- enough" (arithmetic on numbers, and anything able to include it, ARE
- powerful enough in this sense), *ANY* set of axioms which do not lead to
- contradiction (=>being able to prove both A and NOT A for some A, thus,
- by formal logic, being able to prove ANYthing whatsoever!), IS incomplete:
- there will be *true* statements about the formal systems that cannot be
- proven from that set of axioms!!! It does little good to say, "oh well
- then, I shall add those true statements as new axioms", because the new
- axiom set that you thus obtain will STILL be incomplete - there will be
- OTHER true statements which can't be proven in the system, and so on, AD
- INFINITUM. Adding an infinity of axioms at once is no better - unless you
- introduce contradiction, there WILL still be true statements that cannot
- be proven!!!
- --
- Email: martelli@cadlab.sublink.org Phone: ++39 (51) 6130360
- CAD.LAB s.p.a., v. Ronzani 7/29, Casalecchio, Italia Fax: ++39 (51) 6130294
-