A discussion of the term "truly random."
Subject: random: Glossary Term Date: Fri, 06 Nov 1998 22:06:25 -1000 From: Horst Ossifrage <WWII@wwww.com> Message-ID: <3643FF81.593E@wwww.com> Newsgroups: sci.crypt Lines: 92 random: Glossary Term 3 of 5 ____________________________ The word "random" is easily understood by most readers of the sci.crypt newsgroup. I will not define it here, but a few comments need to be published, anyway. Consider the difference between "random" and "truely random". There is no difference, and yet people write scholarly papers, and are not satisfied to call numbers random, they need to emphasize by calling a sequence of numbers "truely random". That practice should be stopped: by editors, or by style convention sheets that are commonly distributed to authors. There are three main classes of numbers considered here: random numbers, pseudo-random numbers, and non-random numbers. (See Terry Ritter's Glossary for more discussions). One more point: some sequences "seem random to people" and some "seem non-random". But that perception has almost no value in the realm of keys and ciphers. Flipping a coin has a human influence and seems to make random numbers but other forms of human influence seem to be non-random. The "Osprey Protocol" _____________________ The Osprey Protocol is a way to distribute keys and ciphertexts on insecure channels, like e-mail and newsgroups. Many varia- tions are possible, but the one described here uses 2-level super-encryption. A plaintext is encrypted with scott19u using the keyraw.key file that is created by cutting and pasting an entire message, like "mudge: Glossary Term 2 of 5" posted 11/5/1998, on sci.crypt, with the "space counts" added. That ciphertext is then encrypted using PGP 2.6.2 (IDEA) using a key that is composed of the "space counts". Space counts are defined by Rule 3, below. The PGP outputs ASCII letters as ciphertext by using the -a option. A chaining strategy is recommended, where the keys are in one message, without the corresponding ciphertext, and then a second message has the next key, and the ciphertext for the first key. Rule 1: The paragraph shape can be used to indicate to the initiates that the Osprey Protocol is being used. For example, a monotonically decreasing paragraph shape is the rule this week. Next week's rule will be indicated in later ciphertexts. This way, no explicit statements need to be made to alert initiates to the existence of the protocol. Rule 2: Use only the first paragraph for space counts. Rule 3: Determine the "space counts" as follows: add one space after the longest line and put a 1 after the space. The numbers are right adjusted as the spaces are counted for each line, and the count is typed at the end of each line. Try it for the first paragraph in this essay. Rule 4: String together all of the space counts to make the key for IDEA. If there are less than 16 digits, pad with zeros. If there are more than 16 digits, use the first 16. Discussion __________ The "mudge: Glossary Term" posting has over 1200 characters as the key file for scott19u. The key for the IDEA step has 128 bits or 16 ASCII characters. Use the 16 digits of the "space counts" as the key for IDEA. Here is the ciphertext that you may decrypt using the Osprey Protocol: -----BEGIN PGP MESSAGE----- Version: 2.6.2 pgAAAHjOfUx2V7TUJlGWodjtSIrN0nmrxo3Xhe+LPHZTYeAedCk27mKonpljOIRO zJszElBiZpzdf3eAOATHRC2wEpC09CstduEV0fkoP5bLvetDRofbKzpRbzwF9uTM ohrPYZ5ZnFR00IdOXJfe7/1BqXBF82TEuyObzns= =BSE6 -----END PGP MESSAGE----- Use this current posting for the keys for tomorrows part 4 of 5 Glossary Term postings in sci.crypt. In scott19u, use the menu choice with no padding, and when asked for a key, he means give a password. Press ENTER without giving a password. November 6, 1998 9:55PM Horst Ossifrage San Jose USA !
Subject: Re: random: Glossary Term Date: Sat, 07 Nov 1998 10:57:44 -0600 From: jgfunj@EnqvbSerrGrknf.pbz (W T Shaw) Message-ID: <jgfunj-0711981057450001@dialup172.itexas.net> References: <3643FF81.593E@wwww.com> Newsgroups: sci.crypt Lines: 7 I find your definitions "slanted." -- --- The public is harder to steamroller than some might think. --- Decrypt with ROT13 to get correct email address.
Subject: Re: random: Glossary Term Date: Tue, 10 Nov 1998 04:53:15 GMT From: scott@helsbreth.org (Scott Nelson) Message-ID: <3648bf9a.92225357@news.inreach.com> References: <3643FF81.593E@wwww.com> Newsgroups: sci.crypt Lines: 42 On Fri, 06 Nov 1998, Horst Ossifrage <WWII@wwww.com> wrote: > >The word "random" is easily understood by most readers of the sci.crypt >newsgroup. > Unfortunately, it's also easily misunderstood by many readers. >I will not define it here, but a few comments need to be >published, anyway. Consider the difference between "random" and >"truely random". Ok. The main difference I see is that random is a general term, used to denote a large class of things including some things (such as pseudo random number generators) which don't have all of the properties needed for "true" randomness. >There is no difference, and yet people write >scholarly papers, and are not satisfied to call numbers >random, they need to emphasize by calling a sequence >of numbers "truely random". I disagree, there is a difference. Just as there's a difference between "cat" and "house cat" or "number" and "integer." >That practice should be >stopped: by editors, or by style convention sheets >that are commonly distributed to authors. Again I disagree. It's very useful to be able to use short words for discussion, with the exact form taken from context. If by convention "random" always meant "truly random" then talking about pseudo random would be very cumbersome indeed. And since I'm posting anyway, here a link to Ritter's glossary, which is, IMO, excellent. Thanks Terry for providing it. http://www.io.com/~ritter/GLOSSARY.HTM ----------------------------------- Shameless plug for random web site: http://www.helsbreth.org/random Scott Nelson <scott@helsbreth.org>
Subject: Re: random: Glossary Term Date: Tue, 10 Nov 1998 22:32:25 -1000 From: newWebsite <polite@yahoo.com> Message-ID: <36494B99.27C@yahoo.com> References: <3648bf9a.92225357@news.inreach.com> Newsgroups: sci.crypt Lines: 43 Scott Nelson wrote: > Ok. The main difference I see is that random is a general > term, used to denote a large class of things including > some things (such as pseudo random number generators) > which don't have all of the properties needed for "true" > randomness. > > >There is no difference, and yet people write > >scholarly papers, and are not satisfied to call numbers > >random, they need to emphasize by calling a sequence > >of numbers "truely random". > I disagree, there is a difference. Just as there's > a difference between "cat" and "house cat" or > "number" and "integer." > > >That practice should be > >stopped: by editors, or by style convention sheets > >that are commonly distributed to authors. > Again I disagree. It's very useful to be able > to use short words for discussion, with the > exact form taken from context. If by convention > "random" always meant "truly random" then talking > about pseudo random would be very cumbersome indeed. Please answer 2 questions: 1) Are you certain of the definitions? 2) Are you truely certain of the definitions? If you can answer yes to one and no to the other, then you have a good argument to convince me that there is a difference between random and truely random. I am waiting... -- See the newWebsite on A Large-Key Encryption algorithm's inner workings! http://members.xoom.com/ecil/ for the finest file scrambling in the known universe! Expect nothing, get everything.
Subject: Re: random: Glossary Term Date: Wed, 11 Nov 1998 16:59:08 GMT From: scott@helsbreth.org (Scott Nelson) Message-ID: <364ab715.879846@news.inreach.com> References: <36494B99.27C@yahoo.com> Newsgroups: sci.crypt Lines: 96 On Tue, 10 Nov 1998 newWebsite <polite@yahoo.com> wrote: >Scott Nelson wrote: > >> Ok. The main difference I see is that random is a general >> term, used to denote a large class of things including >> some things (such as pseudo random number generators) >> which don't have all of the properties needed for "true" >> randomness. >> >> >There is no difference, and yet people write >> >scholarly papers, and are not satisfied to call numbers >> >random, they need to emphasize by calling a sequence >> >of numbers "truely random". > >> I disagree, there is a difference. Just as there's >> a difference between "cat" and "house cat" or >> "number" and "integer." >> >> >That practice should be >> >stopped: by editors, or by style convention sheets >> >that are commonly distributed to authors. > >> Again I disagree. It's very useful to be able >> to use short words for discussion, with the >> exact form taken from context. If by convention >> "random" always meant "truly random" then talking >> about pseudo random would be very cumbersome indeed. > >Please answer 2 questions: > >1) Are you certain of the definitions? > >2) Are you truely certain of the definitions? > >If you can answer yes to one and no to the other, then you have a good >argument to convince me that there is a difference between random and >truely random. I am waiting... Gee, I'd hate to keep you waiting . . . 1.) Yes. 2.) No. (If you want, I'll even answer yes and no to the same question. :-) Normally, I just gloss over people who want to argue semantics, but in this case, I think semantics is the whole point. I'll expand on my original statements. In my definitions of random, ( http://www.helsbreth.org/random/ ) I make a distinction between "true random" "cryptographically strong random" and "pseudo random" These terms come from a scanning of the literature. Now if I were choosing, I wouldn't choose "true" to describe the unreproducible property, but it's not my choice to make. The damage has already been done. There's some confusion between the common usage of "true" with the usage normally associated with random. In other words, "true random" has a specific idiomatic meaning, just as "true colors" means something different from "colors." Because english is extensible, truly random might have the more limited meaning of "a random generator having the property of being able to produce different outputs given the same starting conditions." Random might refer to a pseudo random generator, truly random can not. Now one could argue that random shouldn't include pseudo random, or cryptographically strong random. Random should be used only for things which measure up to the ideal of randomness; unbiased, unpredictable, and unreproducible. While I agree that it's an arguable position, it happens to be one I disagree with. I prefer to keep discussions brief and to the point. Short words and phrases should be assigned to those concepts we use most often. We don't talk about the ideal randomness very often, while pseudo random generators are a dime a dozen. If we were to affix a single meaning to random, I would think ideal randomness would be a poor choice. Better still is to allow random to a be dynamically assigned within the context of the discussion at hand. This is typically done by using the precise term at the beginning of the paper, then using the short form in the remainder. I like it that way, which is why I took the time to argue against the original poster in the first place. ----------------------------------- Shameless plug for random web site: http://www.helsbreth.org/random Scott Nelson <scott@helsbreth.org>
Subject: Re: random: Glossary Term Date: 12 Nov 1998 16:10:23 GMT From: jmccarty@sun1307.spd.dsccc.com (Mike McCarty) Message-ID: <72f19f$72s$1@relay1.dsccc.com> References: <3643FF81.593E@wwww.com> Newsgroups: sci.crypt Lines: 46 In article <3643FF81.593E@wwww.com>, Horst Ossifrage <WWII@wwww.com> wrote: )random: Glossary Term 3 of 5 )____________________________ ) )The word "random" is easily understood by most readers of the sci.crypt )newsgroup. I will not define it here, but a few comments need to be )published, anyway. Consider the difference between "random" and )"truely random". There is no difference, and yet people write )scholarly papers, and are not satisfied to call numbers )random, they need to emphasize by calling a sequence )of numbers "truely random". That practice should be )stopped: by editors, or by style convention sheets [snip] Unfortunately, the word "random" by itself has no particular meaning. It is not usefully used as an independent adjective. It has meaning only when used in combination with other words. For example these terms all have well-defined meanings: random variable random sample random sequence random (or stochastic) process None of these terms have any well-defined meaning random number random person random house (sorry, couldn't resist) In short, "random" is not really a word in itself. (I'm talking in pedagogical terms, I don't mean that it can't/shouldn't be used in casual speach.) Also unfortunately, as typified by random variable, the special meanings the special terms have are not particularly associated with being random. For example, a random variable is neither random, nor is it a variable; it is a function. Mike -- ---- char *p="char *p=%c%s%c;main(){printf(p,34,p,34);}";main(){printf(p,34,p,34);} This message made from 100% recycled bits. I don't speak for Alcatel <- They make me say that.
Terry Ritter, his current address, and his top page.
Last updated: 1999-01-19