RND.THD --- Copyright 1987 by Phil Wheeler An original compilation of Compuserve Model 100 Forum messages for use by Forum members only. The random numbers generated by the Kyocera laptops are generally uniformly distributed but not truly random. This set of messages deals with comparative results from different laptops, and the implications on cypher/decypher programs which us the RND() function in Basic. ** Updated on 10/10/87 ** Message range: 156960 to 157436 Dates: 9/21/87 to 9/29/87 Sb: #RND() on Laptops Fm: James A. McGreevy 70206,1115 To: ALL I was just wondering whether anyone has compared the RND() series on the M100 with the other laptops. Presumably the T102 will have the same sequence as the M100, but do the T200, NEC8201, Oly M10 or Kyocera 85 have the same generator? I've written a file ciphering program for the 100 and I was just wondering which other of these machines would be capable of deciphering M100-enciphered messages. If no-one's done this work yet (Tony or Phil or Denny ought to know), I can put a quick program up here to create a .DO file of the first 50 random numbers generated on the various machines, which should be enough of a comparison. Fm: Tony Anderson 76703,4062 To: James A. McGreevy 70206,1115 See the files PGMTIP.016 and RANDOM.TIP, both in DL8, for some additional information which applies to your question. Fm: James A. McGreevy 70206,1115 To: Tony Anderson 76703,4062 Thanks, Tony. I've seen those files, but they don't really answer the question. Guess I'll have to ask owners of the machines I listed to email me the first several RND numbers, then if you like I could leave up a file on it - or do you think that'd be too esoteric? CYPHER.BA1 uses PEEKs to examine the FAT in RAM, so it's M100/102-specific, but if the RND series are the same on the other computers, then it should be simple to convert them to read each other's files. If not... Btw, think I should put up an encrypted file in the DLs to see whether anyone wants to try decoding it? It's not a simple algorithm - I wrote it to challenge a statement that Peter Wright made in "Spycatcher" - and it may require a mainframe to decode. But some may enjoy the challenge. I'd like to find out how good it is, anyway. Cheers, --jam Fm: Tony Anderson 76703,4062 To: James A. McGreevy 70206,1115 Sure, you can put up a test file if you like. Marvin Miller did that once with a program he had available, CIPHER.MMM as I recall. Only he offered a "reward" to anyone who could decipher it. (grin) As for the random number generator, we know the 100 and 200 generate the same series of numbers, and can assume with some certainty (without looking) that it's the very same algorithm (hate that spelling! Should be rhythm!). Esoteric is right... probably no one else in the known Universe is curious as to whether all the machines develop the same series of numbers. Lemmee see.... there is a coding/decoding program already available for the 600 which is also runnable on the 100/200 and even PC's. Did you see that one? It's there in DL11 somewhere. Fm: James A. McGreevy 70206,1115 To: Tony Anderson 76703,4062 Nope, haven't seen that one. I'll pop over to DL11 as soon as the system comes back up. Been advised it's going down in two minutes. Cheers --jam Fm: James A. McGreevy 70206,1115 To: Tony Anderson 76703,4062 Btw, what kind of reward did Marvin offer? Just downloaded CODER.600 - 16 minutes for a thousand bytes does seem extremely slow, but then 600 BASIC's foibles are well-known here. No INSTR, no PEEK, no POKEs; I congradulate whomever wrote CODER for getting it up to that speed! The 100 is eminently more programmable. Oh, and how long did Mr. Miller leave his ciphered message up before he uploaded the program? I might wait a week for the hard-core codebreakers to take a crack at it, then upload the program. It's user-configurable (I have another program to aid in that), so I won't upload a version which will decode it. But knowing the algorithm I made up might help slightly. I've got to rewrite the docs anyway. Would you happen to know what 94! (94 factorial) is? The highest I can find is 69!, and after using computers for so long I've completely forgotten the Trachtenburg system (speed math). Can't find it in the library. Fm: Tony Anderson 76703,4062 To: James A. McGreevy 70206,1115 We have a program in the DL's for the 100/200 that calculates factorials, but with the portable's limitations, it only goes up to 48!. Phil Wheeler (71266,125) is a PHD Mathematician, perhaps he'd have the info, or perhaps could calculate it for you. As I recall (vaguely) Marvin offered $100 to the first person who could break the code of his sample file. It's been some time ago. Bill Templeton remembers, he mentioned it recently because there is a new code/decode program in the contest in DL14. I don't recall how long the file was up. Fm: Neil Smith 76257,3227 To: James A. McGreevy 70206,1115 James - All this talk piqued my curiosity, so I tried it on the two machines I have use of right now. The Olivetti M10 (North American)'s first random number is .40649651372358 and the NEC PC-8201A's is 951182 Talk to ya later, Neil Fm: James A. McGreevy 70206,1115 To: Neil Smith 76257,3227 Thanks, Neil, once again. Guess that CYPHER will only be a machine-specific program, though that's not such a bad thing considering it's purpose. The M100's first RND(1)=.59521943994623. Perhaps I could _build in_ a "random" number generator in a future version if anyone wants that. Fm: James A. McGreevy 70206,1115 To: Tony Anderson 76703,4062 Okay, I'll take the bait. I've uploaded CYPHER.COD into DL7 _with_ a $100 prize for the first person to decipher it. According to Neil Smith's message, it seems that the non-Tandy laptops don't generate the same random number series, so the program will only work between the Tandy lappers. I'll test the upload later today, though I've already decoded it successfully via EasyPlex. Fm: James A. McGreevy 70206,1115 To: Tony Anderson 76703,4062 Just checked CYPHER.COD, the checksum was correct and the decoding went fine (for me, that was easy!). Do you was me to announce this _other_ contest now or wait until the DL17 one's completed? Might generate some interests in contests in general if I announce it tommorrow. Btw, just picked up a T200 at the new sale price, so I can probably set the program to recognize it as well before I upload it next week. Fm: Tony Anderson 76703,4062 To: James A. McGreevy 70206,1115 CYPHER.COD is already in the DL's, and the prize is mentioned in your description as well as the "New Uploads" description in DL1 DEScription. That's probably enough. New Sale Price for the 200? How much? Fm: Phil Wheeler 71266,125 To: James A. McGreevy 70206,1115 Alas, I do not have the exact value. A rounded value is 1.0874 E+146. Most calculators only go up to E+99, which is why it is a bit difficult to come by. Sorry I couldn't remember the rest of the digits -- old brain cells . p.s. -- What are you going to do with such a BIG number? Not much of anything will accept it as input! Fm: James A. McGreevy 70206,1115 To: Phil Wheeler 71266,125 Thanks much. I just needed it to attempt to calculate the odds on deciphering messages created with the program I just wrote for the 100. Been a long time since I've taken statistics, but I thought I'd give it a whack. Suppose I'll have to find some tomes on coding and cryptanalysis as well, to figure in the shortcuts they've devised. Fm: James A. McGreevy 70206,1115 To: Neil Smith 76257,3227 Have you tried setting A#=RND(1):?A# to see what the result is? That might give a double-precision result. --jam Fm: Neil Smith 76257,3227 To: James A. McGreevy 70206,1115 Good thinking--That got .4919001758098602.