Dqcrq $I l JrnVU GBPD $GJI GJGBPDUVlIlGJGBPDlb ITIIJ V Ul sIIJ Ul $ WGBPi l1 iJWl l Ul W $GBPi l l $ f lBm IJUlDl1lD C1 D BI D i i i 1 SIS 1J0ED Jr------- I o TH1 SUPPORT TOOL U ooooo A HISTORIAN LOOKS AT SIG INT U Vera R Filby METEORBURST COMMUNICATIO S U ooooooooooooooooo J r oo o oooo Ii ' o o 6 J It 'i ' o a A B R IE F TR EA TI S E O N F I ' VE -'LAWS OF TELEPHONIC COMMUNICATION U ooooooooooooooo William M Nolte oooooo oo 10 BUT LIFE IS SUPPOSED TO BE HARD U oooooooooooo 12 I ooo CRYPTIC CROSSWORD U o ' o 'o oooo 30 RULES FOR THE CAMEL CORPS U o ooooooooo 31 TOWARDS BETTER SYSTEM DEVELOPMENT U o oooooooo 32 OLD PHONE BOOKS NEVER DIE U ooooooooooooooooooo William M Nolte ooo o oooooooo 38 CONSUMER VS COMPUTER A REVIEW U oooooooooooo 40 THE MAIL BOX U oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo 42 A TOY PROBLEM U oooooooooooooooooooooooooooooo I 43 KRYPTOS SOCIETY NEWS D ooooooooooooooooooooooo 44 1 r 1 Iooooooooo oo 'FillS BOef 1MBN'F eON'FllIN8 C09BWOK9 AiJ TJijKIJ J 6b SSIFIEB BY tl6MG66M 1 H REVIEW etl 18 Har 2812 2 eclassified and Approved for Release by NSA on '10-'1 2- 20'1 2 pursuant to E O '135 26 vl DR Case # 54778 DOCID 4009842 Published by PI Techniques and Standards for the Personnel of Operations VOL IX No 3 MARCH 1982 PUBLISHER BOARD OF EDITORS Editor-in-Chief I 1 7119 8322s production ooooooooooo 3369s _I 8555s Cryptanalysis ooooooooooo 1 1 311s Collection ooooooooooooooo Cryptolinguistics oooooo I 5981s Information Science o o 1 3034s Language oooooooooooooooooooo 1 $161s Machine support 1 084S '--- ' Mathematics oooooooo o I 8518s Puzzles ooooooooooooooDavid H Williams f03s Special Research ooooooooo Vera R Filby t 19s Traffic Analysis ooooooooooo Don Taurone 1 r s For individual or organizational subscriptions send name and organization to CRYPTOLOG PI or call 13369s Editorial One of the interesting aspects of editing a publication is the challenge of finding out what the readers think One of my predecessors has said that on a scale of one to ten a single non-angry letter is an eight whereas an angry letter is a two or maybe a one It is a fact that angry readers are more likely to write or throw rocks through the window with a message attached should that be counted as two messages than are the satisfied or edified or entertained or whatever readers How then is one to know if there any non-angry readers All of this makes me wonder if we should deliberately put in something to angrify the readers every so often We could declare a Letter Month and see what turns people off It probably isn't a good idea to do it deliberately there are too many pattern finder people around here and as soon as they tumbled to the pattern they would lose interest their adrenalin would decline and so would the volume of mail The system we have now might do us just as well It seems to be working pretty efficiently In a recent issue the wrong version of an article was sent to the printer one angry response came to us in technicolor It is being framed If the sender had included his her name we might have asked him her to help us as a proofreader We can probably make enough of these mistakes without trying to schedule them formally Which brings us to a related point the errors in the magazine as you readers see it are almost always mine rather than the author's I am the typist checker so if something offends you shout at me not the author P L 86-36 To submit articles or letters via PLATFORM mail send to cryptolg at bar1c 5 note no '0' in 'log' FeR eFFI8It'th 18K Qtl15JPY DOCIO 4009842 SEeltE'f Sf8til The PERSONAL Computer A retl l Wnt Cryptanalysis Support Tool u ' ' o ' o o by L - -----l Mar 82 CRYPTOLOG 853 SIiiSRE'f Sf8lE E Page 1 ' vtj HA 'Il i - I' l P L EO 1 4 c P L 86-36 86- 3 6 DOCID 4009842 SKSlHiT SpeRl Mar 82 CRYPTOLOG BlleRoI 'l' BP6H Page 2 I EO 1 4 c P L 86-36 rm '1 11 I gu _Lljl L j 93L 86 36 ea-m Mar 82 Page 3 DOCID 4009842 UNCLASSIFIED A Historian Looks At SIGINT U by Vera R Filby E41 ighteen senior NSA executives along with a few representatives from the National Cryptologic School and the Director's staff attended a seminar in the Director's Conference Room on 18 February to hear Ronald Lewin author of ULTRA Goes to War and the recently published February 1982 The American MAGIC Codes Ciphers and the Defeat of Japan speak on SIGINT and Decision Making in War The occasion was the second in a series of seminars for senior executives initiated and arranged by E6 the Cryptologic Management Department of the National Cryptologic School At the first seminar former Senator Birch Bayh spoke to a similarly small selected group E o so that a solid base of data was the greatest contribution of signals intelligence to decision making Of the commanders who understood and appreciated it he mentioned particularly Alexander who every day soaked himself in the production of his ULTRA room and Nimitz who understood the value of SIGINT from the beginning Mr Lewin stated and repeated that no praise is high enough for Nimitz as a commander After a coffee break questions variety of answers and issues Lewin's responses were these In his introductory talk before questions and discussion Mr Lewin presented conclusions to which his researches in World War II signals intelligence has led him Among these were o o o that the major decisions of governments were based not on intelligence but on political considerations o o that ULTRA's most valuable contribution was the day-in day-out accumulation of order-of-battle information and that constant familiarization of intelligence staffs and some commanders with the flow of signals intelligence led to a climate of understanding Mar 82 led to a ArDong Mr intelligence must work in action that is what it exists to do and because intelligence has failed because it did not conform to the pictures the leaders had in their minds and they were therefore unwilling to believe it He later made the point that the value of intelligence is not in having floods of it but in having it right A question on Anglo-American cooperation led to Mr Lewin's evaluation that after its tentative start in 1941 careful prowling around he called it the relationship by 1944 had become very very close Each country had something to give the other Both had systems of belief in common SIGINT CRYPTOLOG UNCLASSIFIED Page 4 DOCTD 4009842 UNCLASSIFIED cooperation varied in different theaters being least effective in the Asian and Pacific areas but in the Atlantic and Europe it was just about perfect -- a many-sided miracle an example of history There were questions on his chapter The Stab in the Back in The American MAGIC Mr Lewin's thoughts on security and protection of signals intelligence were expressed on several levels First of all he said he wrote the chapter out of a sense of tribute to the dedication of cryptanalytic people For Americans one of the central problems was walking a tightrope on the very highest wire On the one side was the vital importance of protecting the information on the other the danger of its annihilation by those who would ruthlessly expose it for unprincipled political reasons without regard for the national welfare He noted that the Rritish early centralized the control of intelligence that the Americans were less successful in resolving service rivalry -- that the Germans were riven with jealously protective rival services The lapses of American security were of intimate concern to the British it was their security too at stake Even now Mr Lewin remarked America has no viable way to manage the lem P 0b- In response to questions about his evaluation of signals intelligence in war Mr Lewin stressed its use to the commander This was his theme in ULTRA Goes to War the battle is the payoff In war intelligence is useful only to the extent that it contributes to the central action that it is present positively in the area of operations that it meets the needs of commanders And to this he added his conviction that commanders must understand and that they must have intelligence officers to help them understand People believe what they want to believe The problem is to get the truth to them Mr Lewin is a military historian and biographer Rommel Montgomery Churchill Slim Wavell and his SIGINT perspective derives from information released to the public record since the mid-1970's With his vast knowledge of World War II history he has this been able to discern and examine the previously invisible presence of SIGINT in the history of the war Listeners found his insights most interesting and perhaps a little disconcerting ATTENTION AUTHORS AND PUBLICITY CHAIRMEN CRYPTOLOG is always looking for readable material It is not a substitute for formal product or technical reports but it does reach most of the people who are interested in the technical underside of our business to borrow and modify a phrase We are now publishing once a month and hope to maintain that schedule How can you get an article into CRYPTOLOG Contact anyone of the editors listed on the inside front cover We don't need an abstract and only one copy is necessary as long as it is readable If it is typed using a Mag Card we can borrow your cards and use your original keystrokes If it is typed into a UNIX system send it to us using the UNIX 'sndmsg' command I send it to - -cryptolg@barlcOS no Oin 'log' and no spaces Mar 82 after you have typed in 'Control D' type 'f' for file and then name the file your article is in WARNING send only alphanumeric files this way If your file contains any special characters you will have to use ftp or cftp Give us a phone call and we'll let you know what to do If your article is on some other machine give I la call We may still be able to lise your keystrokes especially if your computer is connected to PLATFORM If it is not on any system don' t worry we will type it If your organization has a publicity chairman or program director have him her get in touch with us We can get vour news and announcements out to a large distribution list and we would be happy to work with you to get the word out about what your group is doing CRYPTOLOG Page UNCLASSIFIED 5 P L 86-36 DOCID 4009842 L UNCLASSIFIED 86-36 METEORBURST Communications UJ rn eteorburst communications MBC systems for tactical traffic are being offered for sale by two U S t manufacturers ROCKWELL and TELCOM There has been some Soviet interest in Soviet literature in use of meteorburst cornmunic2tion for rapid delivery of a short message The Department of Energy DOE is developing an emergency communication ECOM network for its atomic energy facilities using a sophisticated meteorburst system that can operate through a nuclear war Several large scale networks viz the SNOTEL net in the Rocky Mountains and the AMBCS net in Alaska have been operating for years under severe weather and propagation conditions demonstrating the cheapness an reliability of this mode of communication Meteorburst has also been considered for the strategic connectivity problem where low bandwidth links are use to assign and set up surviving wideband links The fact that several competing companies are in the marketplace and have been buying up competitors implies that a market potential is seen In addition to the obvious military market the third world civil networks are a likely place to apply this low cost technology bectuse of the shortage of shortwave frequencies Every day some ten billion tiny meteors consisting of dust particles orbiting through the solar system enter the earth's atmosphere and leave brief ionized trails which can reflect radio waves Radio signals in the range 20 MHz to 150 MHz can be reradiated from the meteor trails and if the trail geometry is correct for a given path some of the energy will be scattered down to the receiver The paths only last for a fraction of a second so typical systems operate at high bit rate bursts of 2000 to 8000 bps with a duty cycle a tive path to the receiver of 2 to 20 percent Mar 82 Because of the short path duration and random occurrences of useable paths voice communication is impractical but telegraph circuits work very well While circuit control was done by relays and conventional teleprinters the systems were not very useful even though research and experiments have been carried out since the early 1950's The introduction of microprocessors has changed the economics and operational attractiveness of meteorburst by shielding the communication operator from the vagaries of the propagation ARQ systems send little mini-packets of data at high bit rates so that an average of 1 to 4 teleprinter links is maintained Soviet tests showed quick delivery of short messages because very brief trails could be used The main features of the modern systems are path lengths up to 1200 miles Power between 10 watts and 1000 watts depending on the grade of service and throughput wanted Modulation may be FSK or PSK up to 8PSK Bit rates during bursts are typically from 2000 bps to 8000 bps at 8PSK The outstations in the SNOTEL net with over 475 stations cost from $7000 to $10 000 depending on who is asked The DOE ECOM net will be a full duplex grid with about 17 stations and the specification calls for ECOM to pass a 900 character message across the U S between any two stations in 90 seconds The software development to assure this plus reconnectivity is the problem ECOM will use 41 and 48 MHz channels but frequency ranges between 30 and 100 MHz provide practical service A small company in Texas did the basic engineering and hardware of the DOE SNOTEL net with Western Union acting as prime contractor ROCKWELL has bought up the SECODE company TELCOM wrote the RPQ for the DOE ECOM net for which Western Union is also prime contractor The SNOTEL net was used to telemeter snowfall data from inaccessible CRYPTOLOG UNCLASSIFIED Page 6 DOCID 4009842 UNCLASSIFIED mountain slopes in the Rocky Mountains at a rate of a few reports per day in response to polling from a central station The equipment was robust and operated from solar powered batteries at a few watts radiated during the brief transmissions The antennas are smaller than an outdoor TV antenna and this makes them suitable for tactical and remote area transmission Meteorburst might be useful for strategic connectivity within a country or a continental area with relay hops of up to 1200 miles to connect military and weapons bases and other vital facilities because it is narrowband and very flexible The unit costs can be reduced by competition and large scale manufacture Frequency reuse can be fairly high unlike HF because the forward scatter mode make mutual interference very unlikely It can also be used for theater area fallback dedicated and reconnect ion circuits without having to battle through the HF medium Unlike HF meteorburst links can operate on a fixed frequency 24 hours a day At dusk as the earth sweeps away from the meteor showers traffic capacity drops then reaches a maximum at 0600 local time The ability to use a single frequency on a link and to reuse frequencies facilitates the frequency management Because interference is negligible and the propagation and scattering is directional international coordination or registration of frequencies is unnecessary Therefore dedicated domestic and military nets can be set up and run without formal notifications to ITU For transoceanic paths the possibilities are surprising On the Atlantic route Scotland Iceland Greenland Canada is obvious Azores-Canary-Cape Verde-St Peter St Paul to Brazil is another route Or a link can run from the Azores to Miquelon to Canada It's also surprising how many routings there are via islands owned or governed by the U S U K or Commonwealth and France that allow hops across the South Atlantic Indian Ocean Antarctica and Pacific Access to the Mid East and Persian Gulf areas is possible by 1200 mile hops from places where U S or friendly bases or diplomatic installations exist The equipment is so small and cheap that it can be installed on small or large vessels and left unattended on uninhabitable islands deserts or mountains the attractions of wideband satellites overshadowed MBC Now because of the connectivity problem to say nothing of the gross vulnerability of almost all military and civil telecommunications to many threats MBC may look rather more attractive The low cost of the terminal equipment and the availability of working hardware and software from competing suppliers with the convenience of microprocessors for link and message handling -- might make this an attractive fallback option There has been some Amateur interest in MBC and the increased availability of Amateur digital radio equipment may stimulate further Amateur MBC This cheap over the horizon medium should be very suitable for special forces and missions and rapid deployments because they can take their own dedicated equipment and their own frequencies and be supported by a base hundreds of miles away which has satellite terminals to connect it into a global net Interference from TV broadcasting would discourage MBC in urban areas but in remote areas or in war operations such TV interference is very unlikely Summing up cheap flexible hard to intercept almost impossible to jam low power lightweight fixed or variable frequencies long range independent of all other power or communications systems with 50 to 300 bps average throughput nd high reuse of frequencies How have military communicators managed without this References 1 AFCEA 81 ROCKWELL exhibit Product information sheet ' eteorburst Communications TELCOM advertisement in SIGNAL 1981 E A Demin et aI Test Results GBPf Equipment for Data Transmission via Meteorburst Channels of Radio-Telegraph Communication Links Telecommunications and Radio Engineering 1977 No 1 Scripta Publications D C 1977 Western Union data sheet ' eteor Burst Communications AFCEA 1979 Rockwell International data sheet Meteor Burst Communications AFCEA 81 J A Meyer Digital Meteorburst Communications INTELCOM 80 Proceedings D W Brown The potential GBPf meteorburst communication Communications Systems and Random Process Theory Ed J Skwirznski Sijthoff Noordhoff Holland 1978 2 3 4 5 The attraction of meteorburst have been known for a long time and several times the U S and NATO military departments nearly went ahead with projects but in the early 1960's Mar 82 CRYPTOLOG UNCLASSIFIED - - - - - - - - - - - - - - - - - - - _ - - Page 7 DOCID EO 1 4 c P L 86-36 4009842 'l 8P S e8REl'l UUBM bylL -- P L lp16 Mar 82 CRYPTOLOG Page 8 i8P SBeIl Bi UUBIb EO 1 4 c P L 86-36 86-36 _ l2 h_ 8 E - _3E _ _ _o _ _ __ _ _ EO 1 4 c DOCID 4009842 Mar 82 CRYPTOLOG Page 9 'rep SI iSRI i'r lJUBa 4o DOCID 4009842 UNCLASSIFIED A Brief Treatise on Five LaYis of Telephonic Communication u by William M Nolte T54 LAW #1 YOUR FIRST CALL OF THE DAY DETERMINES THE OF DAY YOU'RE GOING TO HAVE KIND o Subcondition B She's not at her desk The truly significant aspect of this subcondition will be revealed to the caller with the next sentence There are two possibilities This law is a derivative of the street vendor's belief that the day's success is determined by whether he makes a sale to his first customer Cut the price give away the merchandise if you must but make that first sale She's in a meeting with so-and-so and I expect her to return at 1000 This person is clearly organized enough to leave a trail and is therefore a good be to return your call Action leave a message to return your call The same principle applies to using the telephone if the first call goes through to the intended recipient you will have a favorable success rate on the rest of the day's calls Unlike the vendor who works in a yes-no situation he either makes the call or he doesn't the telephone caller works in a more complex situation The call can be successfully completed of course or it might result in a busy signal the rough equivalents of the sale - no sale conditions encountered by the vendor The telephone involves however a series of other conditions all indicating a lack of success and each of them worse than the no sale or busy signal response o Subcondition A I'm sorry he's on the other line May I have him return your call Sure but nobody stays on the phone for more than a few minutes around here on official business at least and the mean time for a return based on such an assurance probably drags into days leading to Corollary A ANY CALL RESULTING IN THE RESPONSE DESCRIBED IN SUBCONDITION A WILL EITHER BE RETURNED WITHIN TEN MINUTES OR WILL NOT BE RETURNED AT ALL Mar 82 o CRYPTOLOG Gee she was here just a minute ago but I don't see her and I have no idea when she'll be back Now you're in trouble You have clearly called a floater who could be in the bathroom another office the drug store or fifty other places It is unlikely that he or she is working Action say you will call back and take your chances If you feel especially daring leave a message to have your call returned Do not however postpone your own trips to the bathroom cafeteria or any of those other places to await the call In fact the worst way to get this person to return your call is to sit by the phone for three hours ignoring persistent signals from your stomach or bladder The best way on the other hand is to leave your desk for thirty seconds to deliver a memo to someone in the next office You guessed it your call will be returned while you are gone o Subcondition C He's no longer at this extension Let me see if we have a current number This leads to UNCLASSIFIED Page 10 OOCIO 4-o09BltZ UNCLASSIFIED LAW #2 NEVER PROCEED PAST A SECOND REFERRAL IN TRYING TO TRACK SOMEONE DOWN The art of referrals is sui generis This is not the place for a full exposition on the nuances of the form but they are legion Simply stated assume that a referral has an 80 per cent chance of being outdated or simply inaccurate Computer simulations supporting this estimate are not available on request When in a gambling mood you may want to ask the person who answers the outdated or inaccurate new number if he or she has an even newer number Whether you then choose to take a chance on that number is a matter of personal taste o o Corollary A to Law #2 LAW #2 DOES NOT APPLY TO INTERNS interest of improving their morale if not their business But what do you do with a busy signal You are completely at the system's mercy -- and it has none Wouldn't it be nice if you could gain access to the switching mechanism and make a deal with it Okay I'll take the worst channel you've got with the lousiest bit rate or whatever and I won't complain even if it makes FANX sound 4000 miles away But I've got to complete this call This unfortunately is not a solution but there may be another one out there namely LAW #5 MOST OF WHAT GETS SAID ON THE PHONE DOES NOT HAVE TO BE SAID Interns are a different matter Unless it is a matter of life and death it is rarely worth attempting to track them down No one ever knows where they are When I was an intern I made it a point to surface only about the times plans were being made for barbecues and Christmas parties In cases of life and death do not use the phone Take a chance and of course you can improve your chances if you know the habits of the person involved and position yourself at the drug store or in front of the gatehouse the person is likely to use You could of course leave a message at the intern office but this is a real longshot Do you really think most career panels would have photos of their interns on the walls if the interns showed up there frequently Those are wanted posters When Oscar Wilde saw his first telephone the public relations type shilling the thing attempted to impress him by announcing ' r Wilde with this instrument we can talk to Texas But what do you have to say asked the skeptical visitor The former chief of history and publications insists that this story actually involved Thoreau and the telegraph not Oscar Wilde and the telephone Vince Wilson is a learned man Nevertheless the story is probably apocryphal in the first place second it would be a shame to waste a good line like that on a less than humorous character like Thoreau LAW #3 ANYONE ATTEMPTING TO CALL FANX BETWEEN 0900 AND 1400 DESERVES WHAT HE OR SHE GETS No commentary required LAW #4 THE PERSON RESPONSIBLE FOR MA BELL'S SLOGAN THE SYSTEM IS THE SOLUTION SHOULD BE DRAWN AND QUARTERED This is the most frustrating part of the process Street vendors at least have some control over their fate They can choose to take a loss on a particular sale in the Mar 82 CRYPTOLOG UNCLASSIFIED Page 11 DocrD 4009842 eoltP II' I'f A1 But Life Is JJufi wlied To Be Hard U p l by 1 lpI4 6 668 Maybe 90 but Pl4 the TA Division in the Office of Techniques and Standards is making life a little softer for the working traffic analyst In preparation are a series of How to working aids on the use of T33 Pl4 UNIX PINSETTER software for processing analyzing and reporting using terminal subsystems The first in the series How to Index covers the UNIX PINSETTER commands index permute and ptx Each command 1s covered in detail and examples are plentiful The working aids also contain a handy quick reference So once a command is mastered the basic details and variations of the commana are available without having to review the text S SeQ Though How to working aids are designed specifically for traffic analysts personnel in related SIGINT fields will also find some of the techniques useful If you are presently using terminal subsystems for processing analysis and reporting or will be in the not-to-distant future copies of How to Index can be obtained from Pl4 room BAl77 phone 3369s A second working aid How to Sort is being drafted and should be available 800n The following text with minor adjustments for CRYPTOLOG format has been taken from How to Index Mar 82 his working aid assumes only a rudimentary knowledge of UNIX PINSETTER software on the part of the analyst Consequently each technique and the U command lines necessary to accomplish that technique are covered in detail and examples provided The examples are basic and unrelated to any specific problem By studying and perhaps duplicating the examples shown the command and options can be adapted to individual target areas W U Points document to remember when using this 1Ir Examples shown are in the DELTA DATA 7000 type f keyboard format Keyboards for other equipment connected to UNIX PINSETTER software may differ but generally have keys providing similar functions Operational areas probably have equivalency charts if other equipment is being used If not Unit III - The RAND Editor - in the MPll9 course Introduction to UNIX may be of help UNIX systems to another on the UNIX work on most CRYPTOLOG sometimes vary from one host These examples were developed host BARDOLPHl and should other hosts Within the working aid the symbols and are used Data contained in e g i ignorefile denotes use is optional Capitalized data contained in e g BREAK is required and indicates a specific key on the DELTA DATA 7000 Page 12 P L 8 6- 3 6 DOCID UNCLASSIFIED keyboard In some cases a specific key may also be referred to as the key e g the key File names used in the examples and synopses are descriptive only e g filename means the name of a file newfile the name of a new file to be created etc and not a word or a specific file name that has to be used or included when entering a command For the sake of brevity some examples may be truncated after the first few lines Where this is done it will by 1IIIIetclllll be indicated The word CAUTION is occasionally used in the appendix It is used sparingly and only where necessary to alert the user to a situation that could result in the loss of data or where there are peculiarities in the system that need to be brought to the users attention Make sure files are untabbed before being run through any indexing program % untab filename newfile RETURN QUICK REFERENCE U COMMAND OPTIONS FUNCTION COMMENT index - only one _ symbol per command line lists words only unique words words filename line words to be omitted words to be indexed name of input file c u w i ignoreflle 0 onlyfl1e filename permute -ainputflle file name of input file -boutput file name of output file -iindicative data numbers of columns appended to each line -wpagewidth nr of columns 78 106 132 start end columns for indexing -sstart zone -eend zone zoning limits page width - indicative data - 2 divided by 2 and rounded down specific character for indexing -ddel1miterJ ptx inputfl1e outputfl1e name of input file name of output file Figure Mar 82 o CRYPTOLOG o Page 13 UNCLASSIFIED DOCID 4009842 6eUFrBElrl'FIAJo C eSQ of output c_ mands used tli INDEX works on11 special characters Cca output Consequently this c wmAUU best used when digits and special char - ters are either not a concern or need _ be suppressed INDEX works on the entire file and specific segments of 8 file can not be indexed It is possible to index only selected words or to omit selected words by using special files with the INDEX command PERMUTE handles letters digits and special characters and can be formatted to index only a selected part of a line For those reasons it will probably have the most applications for files containing intercept data There are no provisions to suppress or index selected words but a delimiter can be used to generate an index The PTX command generates an index of all special characters digits and words in a file Like the INDEX command it works only on the entire file There is no provision to suppress or index only selected words INDEXING U DESCRIPTION U U Several methods to index data are available to the analyst The commands INDEX PERMUTE and PTX are all excellent tools to index and arrange data to specific individual requirements By using these commands and related options the traffic analyst can employ UNIX PINSETTER software to quickly arrange and exhibit technical data in numerous ways APPLICATIONS U EO 1 4 c P L 86-36 Mar 82 CR PTOLOG S8I1FIBBII'l''bI tb Page 14 IWlNBJoB Y1A S8UHI'f SlloMRIEfJS eliJoY DOCID 4009842 UNCLASSIFIED % index -c filename RETURN THE INDEX COMMAND U with the output NAME U index a and and appear as as command data example file file filename For format naming of output sentence that the this using would SYNOPSIS U index - c u w i ignorefilel 0 onlyfile filename BASIC COMMAND DESCRIPTION AND EXAMPLE OF USE U U The index command generates an alphabetic index of significant words from a file When used without options the index command omits selected words such as articles and conjunctions e g a an and but the to treats capital letters as lower-ease disregards numbers or special characters and produces an output in this format word filename line-nr context For example using this sentence as a file and naming that file filename the command format and output of data would appear as Figure 3 % index filename RETURN The output of data from this command is in figure 2 shown % index -c filenameluniq -c RETURN with the output VARIATIONS IN THE COMMAND LINE AND EXAMPLES OF USE U 1 a 2 and 1 appear 2 as 1 command 1 data 1 example 2 file 1 filename 1 For 1 format 1 naming 1 of 1 output 1 sentence 1 that 1 the 1 this 1 using 1 would U As shown in the SYNOPSIS there are several options indicated by the data in that can be used with the index command As a general rule the c u and the w options are used independently but can be combined with the i and 0 options to tailor the output display desired The functions of these options and examples of their use are THE 'c' OPTION U U is option lists only the words A aount of each word can be made by piping i e an upper-case backward slash I is referred to as a pipe this causes the system to route your output through another program this option through the uniq command Again using filename the command lines and outputs would be Mar 82 CRYPTOLOG Figure 4 UNCLASSIFIED Page 15 DOCID 4009842 d t would appear aa appear fHen me olllllli d fil'm e 2 filename the data ftlenallle 3 output of data WQuld appear exalllple filenallle file filename filename 1 2 aentence aa a file and na1lling that file filename file filename the 2 filena e 2 OlIIIlJ4nd format naming fHenallle 2 naming that file output filenMe 3 and output of data sentence fllen81lle uaing this sentence as a fl1 filen me example uaing thia J J J 3 J 1 3 filenall1e 1 filenall1e c 1ll1ll8-nd 3 fi l enall1e dat a 1 filename e all111le 1 1 fil ena1l1e f1 lena1l1e file 1 filenall1e nall1ing outllut sentence using 1 filetU ll1e fO t co nd fom t using filenall1e 1 filename 3 f1 1ena1l1e 1 fil enall1e format For exalllple using filename filen me a llea o nd 1 fil enall1e 6 senten e DOCID 4009842 UNCLASSIFIED first make a file called ignorefile containing those words Now using the i option and the ignorefile the index command will ignore listed words while constructing an index Below are examples of how to construct an ignorefile the command line used with the i option ignorefile and an example of the output figure 7 To construct an nignorefile use the cat command and list one per line each word to be ignored % cat ignorefile RETURN a RETURN would RETURN appear RETURN data RETURN example RETURN for RETURN naming RETURN of RETURN output RETURN this RETURN CTRL-d THE 'u' OPTION U U This option will list unique words Words occurring more than once are not listed and capitalized words are considered unique from their lower-case counterpart Using the file filename the command line and output figure 5 would be % index -u filename RETURN Note The first column of output data is limited to seven letters if the word being indexed is longer than seven letters the next column the name of the file is offset Actually the index output contains a 'tab' character between the first and second columns of data If the first column contains a word whose length plus one for the follOWing space is longer the first tab setting the rest of that line will be offset one or more tab stop s usually eight characters per stop to the right Note CTRL-d means to depress the CTRL key while hitting the d key This closes the ignorefile and returns the cursor to the screen % index -i ignorefile filename RETURN The i option may also be used in conjunction with other options but unlike most other UNIX or PINSETTER applications the options are not listed separately e g -i -c ignorefile filename but must be combined i e -ic ignorefile filename Failure to do so will result in an error listing on the screen Also remember that whenever the i option is used with another option this combination of options must be followed by an ignorefile THE 'w' OPTION U U This option lists words without the context The command line and output figure 6 using filename are % index -w filename RETURN THE 'i' OPTION U U This option must be used in conjunction with another file containing words to be ignored by the index command For example if you do not want the words a would appear data example for naming of output nor this to be indexed Mar 82 CRYPTOLOG UNCLASSIFIED Page 17 DOClD 4009842 and filename filename 1 3 file and and output of data would appear 88 as filename filename 1 3 sentence as a file and output of data would appear a8 cOlIIID8nd filename 2 the command format file filename filename 1 2 4S filename filename format 2 filename sentence filename file filename the 2 1 a file and that file filename cOlllllland format using this sentence as that filename 2 naming the filename 2 filename the command using filename 1 For example using this sentence Figure 7 that file DocrD 4009842 UNCLASSIFIED COMMENTS OBSERVATIQNS AND SUGGESTIONS U U You will probably want to save the results of the data indexed To do so simply add a divert symbol key and a new file name to the command line e g % index -c filename newfilename RETURN will place the results of your index of filename into a new file called newfilename U Some index commands take a while to run By adding an ampersand -- uppercase 6 to the end of the command line the program will be run in the background thereby freeing the terminal for other work Three CAUTIONS are in order regarding the use of the o Do no t logout while a program is running in the background Use detach to get off the terminal Otherwise the system drops your job o Never run more than one job in ground at the same time o THE '0' U This option is the opposite of the i option It must also be used with a separate file or onlyfile that causes the index command to index only those words desired An onlyfile is created in the same way as the ignorefile For example % cat onlyfile RETURN and RETURN data RETURN sentence RETURN file RETURN C TRL-d will create an onlyfile containing the words and data sentence and file Now using the 0 option onlyfile and filename the command line and output GBP1gure 8 would be -0 on1yfile filename RETURN Like the i option the 0 option may also be used in conjunction with other options to achieve the desired screen output When used 1n conjunctions with another option the two must be combined e g -oc otherf1le filename and not listed separately Mar 82 CRYPTOLOG back- Always use a divert file when using the option If you do not the results return to the screen immediately after completion This depending on what you are then doing could be somewhat undesirable oooo OPTION U % index the UNCLASSIFIED Page 19 DOCID 4009842 EO 1 4 c P L 86-36 Mar 82 CRYPTOLOG Page 20 e8HHll 81h'dl11EbS 811bJPY- llF IBElll Il'b DOCID 4009842 8eUPIS I U'ftAb can be used with the permute command By using these options the output can be organized to the users specifications and displayed in several different ways The functions of these options and examples of their use are THE PERMUTE COMMAND U NAME U permute SYNOPSIS U permute -a input file -boutput file -iindicative data -wpagewidthl -sbegin zone -eend zonel -ddelimiterl BASIC COMMAND DESCRIPTION AND EXAMPLE OF USE U P L U Index order is special characters numbers and letters Upper- and lower-case letters are treated separately but within the same alphabetical order U We will use two files called filenamel figure 9 and filename2 figure 10 Filename 1 contains evenly formated call sign data while filename2 contains unevenly formated message text These two fil 8 are used as examples to demonstrate the basic permute command and options U Wi thout options the basic permute command functions like the ptx command except that all words are indexed For example usi l g the permute command and the file filenamel the cOllDland line and output figure 11 look like % permute -afilenamel RETURN 86-36 THE '-boutput file' OPTION U U Though an election it is likely that this option will be a part of most command lines when using permute The -boutput file works exactly like the divert symbol in that it directs the results of the permute index into a file within the working directory The -b part of the option is always followed no space by the name of the file either an existing or new file to be created in which results are to be placed CAUTION must be exercised so output is not directed into an existing file containing data since data already in that file will be overwritten U Below is an example of the use of the -boutput file option For this example the output file will be named text 1 and the input file filenamel will be used VARIATIONS IN THE COMMAND LINE AND EXAMPLES or US U % permute -afilenamel -btextl RETURN U As shown in the SYNOPSIS there are several options iudicated by data in l that Mar 82 After the command line is entered into system the following will appear on screen CRYPTOLOG s eMF IS EU'f I ttb the the Page 21 IMfiBbB YIlt 8eUHif SlWlIfBb8 SUbY DOCID 4009842 S8I1HBEII'fIhTo 1 4 c P L 86-36 Mar 82 CRYPTOLOG eOHFl fl15l1'fHrh Page 22 IDmflLI VIA ee IlI'f' ell1dm15bS SlibY ---------- - --'-- --- ---_ _---------------- DOCID 4009842 J SELECTIONS MADE NO INDICATOR LENGTH SELECTED PAGEWIDTH - 78 NO ZONING SELECTED DELIMITER SELECTED IS SPACE CHARACTER INPUT FILE - filenamel OUTPUT FILE - textl Return of the prompt symbol X to the screen tells you the job has been completed and the results placed in the text2 file Results would be exactly as shown in the previous example Note When the job is completed the prompt symbol % returns to the screen In the above example a permuted index of material contained in filenamel would then be in the working directory in a new file called textl Since no other options were included in the command line the output contained in text1 will be exactly like the example shown for the basic command U Use of the -boutput file option alone or with other options always results in a listing of SELECTIONS MADE appearing on the screen after the command line is entered into the system THE '-wpagewidth' OPTION U U This option sets the page width to either 78 106 or 132 characters Widths of 106 or 132 characters must be specified i e -wl06 or -132 If the -wpagewidth option fs not included in the command line the page width is automatically set by default to 78 characters -- the length of the screen If the output of the program is longer than the specified pagewidth the output will be truncated at the specified pagewidth but the system will try to include the data by wrapping it around or fitting it to the other end of the output line This is ind'1cated by the indicator ooo and should be apparent in the text In some cases with longer records data will be lost unless a longer page width is specified THE '-iindicative data' OPTION U U One of the reasons for and use of the -wpagewidth option becomes apparent when filename2 is permuted first with a page width of 78 and then again with a page width of 106 characters For example a pagewidth of 78 characters which is obtained by default when not specified would be obtained by the command line output in figure 13 % permute -afilename2 RETURN Even though no data was lost the input record was only 78 characters some lines of text are wrapped around as indicated by the ooo at the end of text Now using an option of 106 characters the command line output in figure 14 would appear % permute -afilename2 -w106 RETURN Note that some text is no longer I rapped around and that the symbol now appears at Mar 82 CRYPTOLOG Page 88HPlI'EHYIAL 23 thtdtBbE ViA Betlln'f Bl't2'dltlBlJS em I DOCID 4009842 COl'Jl I15 lff IKL EO 1 4 c P L 86-36 Mar 82 CRYPTOLOG EJ8 lF'IBEtITh' J Page 24 Ih dlBbE 'J IA EJ811I11'i' elh'tllllEbS 81lbY DOClD 8S1IFiBBfi'fbIrL the end of some lines This indicates additional text exists on the line but is not shown because of screen size If in the RAND editor hit the PORT-right arrow key and the screen will move over to the text To return the screen to its original location use the PORT-left arrow respectively If a zone greater than the maximum allowable is specified in the command line the program will abort and tell you via the screen why the program was not run For example the following command which exceeds the zone limitation 38 columns for a page width of 78 columns % permute -afilename1 -s10 -e60 RETURN U Another reason for the -wpagewidth option will become apparent with use of -sbegin zone and -eend zone options discussed below Like all other options the -wpagewidth can be used with the -boutput file option causes the error line below to return screen to the Can't fit zoned input indicator on page Execution terminated U If an -iindicative data option is used the size of the zone is further reduced Simple math can be used to determine the maximum allowable zone when an -iindicative data option is included The formula is page width - indicative data -2 divided by 2 fractions of a column are rounded down to the next whole number For example if the page width is 106 columns and the indicator is 17 columns then the maximum allowable zone is 43 columns 106-17-2 2 43 1 2 which is rounded down to 43 If the combined indicator and zone exceed the limit for the page width the program terminates and an error message returns to the screen THE '-ddelimiter' OPTION U THE '-sstart zone -eend zone' OPTIONS U e eeO The permute index can be restricted to a specific portion of the input record This is called zoning and requires that the start and end columns or zone be specified in the command line e g -s35 -e47 The -s oo and -e oo are interdependent and must be used together For example if you wanted to restrict a permute index to control callsigns columns 19 through 22 in filename1 the command line output in figure 15 would look like this % permute -afilename1 -s19 -e22 RETURN - U The size of the zone that can be indexed is limited If the standard 78 columns are used then the zone is restricted to a maximum of 38 columns If the 106 or 132 column page width is specified through use of the -wpage width option the zone maximums become 52 and 65 columns Mar 82 U This option is especially useful when working with data that does not have a set length or when you want to have only a certain item such as all text beginning with a 3 indexed The term delimiter used here refers to a specific character e g a 6 Z etc By default if the ddelimiter option is not included in the command line a space is considered to be the delimiter U When a specific character is used for a delimiter special rules apply It should be used either with zoning or with increased page width Failure to do so may result in the program being terminated an error message appearing on the screen and a core being placed in your working directory If this happens first remove the core % rm core RETURN and then either specify a zone e g -s30 -e40 or increase the page width e g -w132 or both in the command line CRYPTOLOG Page SQNFiBBN'nl 25 DOCID 4009842 Mar 82 CRYPTOLOG G9NFIBBn'fli EO 1 4 c P L 86-36 Page 26 IWIBbB ' 'H 88UIlI'f 81tz fiflBbS 81M L UNCLASSIFIED 86-36 U Fallowing is an example of the use of the -ddelimiter option being used with filename2 In the example the message text columns 32 to 63 is zoned and the I symbol is used as the delimiter The example is brought back to the screen but would normally be placed in another file by using the -boutput file option see figure 16 % permute -afilename2 -s28 -e63 -d RETURN example The program automatically placed I characters before the first column of data in the zone This allowed the first part of the message text which was not preceded by a I to be included in the index and o If an indicative data option h-a d-- b-e-e-n i ncluded in the command line and also because the indicative data and zone together exceed the limit for the page width of 78 columns the page width increased this data would have been retained U Now including the indicator and pagewidth options the entire command line and output figure 17 would look like U Users should be aware of an idiosyncrasy occurring when the and the -boutput file option are used together Basically when the command line is entered the system provides a process identification number frequently referred to as the process ID or PID and returns the prompt X to the screen Immediately thereafter the SELECTIONS MADE listing also returns to the screen Since the prompt symbol is already on the screen it does not reappear after the listing completes Consequently with the SELECTIONS MADE listing between the prompt and the cursor it is not apparent that the system is ready to accept another command However a new command can be entered immediately or if you prefer hitting the @ and RETURNS keys brings another prompt symbol to the screen U CAUTION Care must be exercised when the permute command is used on a large file In cases where a sizable zone of data in a large file is indexed the resulting output can exceed available temporary working space and cause your user group system to crash This problem is compounded by running the command in the background using the symbol If necessary large files can be split into several files using the split command permuted separately appended together and than piped through the csort command to achieve the desired output For example the command line % split -250 filename RETURN % permute -afilename2 -s28 -e63 -dl -i27 -w106 RETURN All the indicative data is now shown and the page has been expanded to 106 columns The file is best viewed in the RAND editor where the PORT keys can be used to see the entire text If your terminal does not have PORT right arrow and PORT left arrow keys you can get the same action by using Ctrl s for PORT right arrow and Ctrl a for PORT left arrow COMMENTS OBSERVATIONS AND SUGGESTIONS U U Like the ptx and index cOlDll14nd permute can be run in the background by adding an ampersand to the command line When running the program in the back ground an output file should also be specified to prevent the program returning to the screen when ever the job completes With the permute command this is done by including the -boutput file option not by using the divert symbol as is done with the ptx and index commands Mar 82 would split the data in file filename into separate files each with 250 lines of data The resulting files xaa xab xac etc can than be permuted separately but not in the background at the same time In turn these files can be combined and the output piped the I key through the csort command and its col and -0 options The col option specifies the column where the merge is to occur this will be the same as the start of the zone used in the permute command and the -0 option precedes the name of the file in which the results are to be placed For example the command line cat xa xb xc f1 Icsort 25 -0 f2 RETURN would append the files n xb and xc into f1 and pipe f1 through the csort command In turn the csort command sorts f1 on column 25 and places the results into a file called f2 Had the start of the zone previously used to permute the three files in this case 25 been omitted the merge would have been 'left-justified and not in the desired format CRYPTOLOG UNCLASSIFIED Page 27 UNCLASSIFIED as the special characters as the special characters I command oooo the special characters I contains the numbers contains the numbers 1 the numbers 1 2 and special characters I n sentence placed in a file as well as the special made to demonstrate the ptx The ptx filename $ and was made to This sentence placed in a ptx command line using the and 3 ooo command line using the file This sentence placed and the resulting output look The ptx command and the resulting output $ and was filename contains the and the resulting This sentence was made to demonstrate the The and the This as well as the filename ooo resulting output look like The ptx command line $ and and ooo as I II and ooo as well II and ooo as well $ and was made to demonstrate the ptx I and ooo as well as 1 2 and 3ooo filename 2 and 3 ooo filename 3 ooo filename contains and ooo as well as the called ooo This characters I and command oooo $ and was command line using the file filename contains the numbers 1 2 and 3 demonstrate the ptx command file called file filename ooo The filename contains the numbers 1 2 filename ooo The ptx in a file called like this line using the file filename look like this made to demonstrate the ptx command numbers 1 2 and 3 output look like this placed in a file called ptx command oooo $ and ptx command line using the file filename resulting output look like this sentence placed in a file called special characters I n and The ptx command line using the file This sentence placed in a file called this ooo and the using the file filename was made to demonstrate the ptx command well as the special characters II n Figure 18 NATrONAL CCPI PLJTGBPR lI m t 1 00 - ItCf j IlOoot oo MANUAL wRITER S ENTt Au GBP E 'l1 t 9ITU eEUR IAJ5T LrU7 OF _ a II1 ' 1 lOOt- t u rlft o 'TRII JC E or I I Mar 82 CRYPTOLOG UNCLASSIFIED Page 28 r oNl _1 oo E JvTR Ct l 'h_ J C or o I DOCID 4009842 UNCLASSIFIED THE PTX COMMAND U will ptx the data in the file filename and place the results into another file called newfile Because of the way the program was originally written this is an automatic divert and it is not necessary to place a divert symbol the key between filename and newfile Use of the divert symbol willj however produce the same results NAME U I'tx GBPermu ed inde SYNOPSIS U ptx inputfile outputfilel COMMENTS OBSERVATIONS AND SUGGESTIONS U BASIC COMMAND DESCRIPTION AND EXAMPLE OF USE U U The ptx command generates a permuted permute arrange objects in a series in all the possible ways in which they can be arranged Webster's 3rd New International Dictionary index of all words numbers and special characters in a file U When using the ptx command it is frequently advantageous to use the divert option and run the program in the background in order to free the terminal for other work To do so simply add an ampersand upper-case 6 to the end of the command For example % ptx filename newfilename U The index sort order is special characters first digits next and letters last Capital letters are treated as lower-case and punctuation ignored The key word appears in the middle of the page with preceding and following text also shown The words a an and 1I as is for lIof on or the to and and up are suppressed for key word use but included in preceding or following text This sentence placed in a file called filename contains the numbers I 2 and 3 as well as the special characters I I and $ and was made to demonstrate the ptx command The ptx command line using the file f1 lename and the resulting output look like this see figure 18 CAUTION If you leave the system while a job is running in the background use detach vice logout Otherwise the job is dropped Also make sure there is a divert file specified in the command line Failure to do so creates instant garbage when the job completes returns to the screen and combines with whatever you are doing at the time U Before running a ptx it is best to do a little selective remove useless lines % ptx filename RETURN VARIATIONS IN THE COMMAND LINE AND EXAMPLES OF USE U THE 'outputfile' OPTION U U The output file option will be used frequently It places the results of the ptx index into a new file within the working directory This is done simply by adding an additional file name e g newfile to the basic command line For example the command line % ptx filename newfile RETURN Mar 82 CRYPTOLOG UNCLASSIFIED 1 Page 29 frequently editing to DocrD UNCLASSIFIED CRYPTIC CROSSWORD Byl I _ 'ooo - ' 1 oo o o o o oo o o o ooo o oo o o oo o oo o o ooo o oo o o o o o oo o o o o o o oo o o o o I ItJ o o o o o o o ' ' ' t - ' o - ' f ' - o TIl h 1 13 16 IS 17 - 2 o TI 22 2 ' - ' 2 29 2 Relative expression of sUlTender 5 1 Do siaUar erowers produce buales 7 5 3 Affin recipe that puts apple juice back into chopped liver 9 9 CloseIIOUthed as a result of severe tic Awful I 9 10 The clove let out 4 Double or nothing out of race objective 4 It' 5 tOo thinl 5 S A wet rice concoction for a sUlaer beverage 3 51 11 Believe me Tick fever SymptOIlS include cause t o '1011 i't 6 6 oild alee about B fine lolf score I 5 12 Lunatic pets loat and holds up exit 8 13 Mever consult a nautical chart wles5 holdin a rulerl 6 7 _Sovereilll troubl e ir i t1a11y after car crashes into car I 8 15 The congress n lave public utterance and - Ie aMnds 8 8 London art gallery mved into South Dakota it's said 6 14 Horrible lB Settles one charled particle in aaona equals 8 19 Ahat Include a JUn for Christie 6 21 COIle back a ain to harvest fruit 8 Drop sale of cats 8 16 I4a1lo enters a rain oddly in South Allerica 9 17 Attellpt by the sound of it to fish for a poly on 8 23 Innate order where one finds fish 2 1 3 lB Insane report carries the load 6 26 The Spanish and the Gerun berry bush 5 27 15 the lout suina over the sticky stuff 9 20 Svears to student's whereabouts at end of tel'll perhaps 7 28 A nice insider 22 An anonYJlOus tipper once exposed a '0 5 berserk over bolllbsl 12 foraer dictator 5 24 She co s froa in dispersion around the first of October 5 1 For sallplers take tea 7 50K blosso s after 25 Being impolite we hear is to be re re ted 4 Mar 82 CRYPTOLOG UNCLASSIFIED Page 30 P L 86-36 DOCID UNCLASSIFIED Rules for the P L by t is frightening to contemplate the amount of time NSA employees spend in meetings There are staff meetings at all levels working groups meetings to solve a particular problem club meetings and meetings to find reasons for more meetings He's at a meeting is all too frequently heard on the other end of a phone line 4 Prepare an agenda or discussion outline and distribute it in advance Everyone should be aware of the contribution he is expected to make If possible include background information and ensure that pertinent documents are available to all 5 Keep the discussion on the subject and under control Digressions and war stories should be nipped in the bud This is an important way to help accomplish Rule 2 Above all don't let someone else assume command of your group I find that most people respond very favorably when a meeting is effectively kept on the subject at hand 6 Give everyone a chance to express his opinion The person sulking in the corner may have a more important comment than the most outspoken attendee 7 If you finish early dismiss the meeting I remember one leading Camel Corps member who accomplished his stated purpose in the first five minutes of a scheduled one-hour session This disconcerted him so he kept up a pointless discussion just to fill time I walked out 8 Briefly summarize the session before ending it This reduces the chance of later misunderstandings Each participant should be aware of any follow-up actions for which he is responsible 9 Prepare a written record for participants and for anyone else who should be brought up to speed Right more CC slang on the meeting Keep the record yourself until all danger of repercussion is past fJ As a veteran staffer I feel qualified to say that not all of this is time well spent I have fidgeted and chafed through many meetings that were unnecessary poorly run overlong and ended in chaos The people who run these I have affectionately labeled the Camel Corps because their product so often resembles the design of the proverbial horse Conducting a proper meeting should not be terribly difficult as long as the chairman follows a few simple rules 1 Make certain a meeting is really necessary A few strategic phone calls or memoranda may provide a more efficient solution to the problem 2 Schedule the meeting for a fixed period of time Except under very special circumstances no meeting should last longer than one or one and a half hours By that time you've lost the narcoleptics and doodlers and are in danger of losing everyone else It's better to schedule two sessions than to have one become unproductive 3 Invite all the right people and only the right people It s all in vain if you're missing what the Camel Corps would call a key player Also people who have no reason to be at a particular meeting tend to become hostile or to lead the discussion into an area where they can have input Work with the smallest possible group Mar 82 J Sound too simple and obvious Think about them whenever you take part in a meeting I don't think it will take you long to learn to spot the Camel Corps at work CRYPTOLOG UNCLASSIFIED Page 31 86-36 DOClD 4009842 UNCLASSIFIED TOWARDS BETTER SYSTEM DEVELOPMENT U J by I _ -_ _ p 1 mR53mmu This paper was originaUy written for the first annual conference of the Communications Analysis Association on communications anaZysis The AnaZyst in the 80's It appears here with minor revisions T he Analyst in the 80's will continue to be seriously affected by personnel constraints changes in the workforce and the evolution of the communications and technological environments Among the various approaches that will be used to deal with these problems will be an increasing use of tools to support the analysts in their various activities Central to these tools will be interactive computer systems Personnel constraints that limit the size of the workforce are a source of pressure for increasing productivity Changes in the workforce caused by the retirement of experts and the difficulty of recruiting and developing their successors are a source of pressure for finding ways to capture existing expertise for transfer to future generations Evolution of target environments to more sophisticated techniques and greater capacity is a source of pressure on diagnosis the development of attacks and their transfer into computing systems Responses to such evolution range from use of a known attack on a larger scale refinement of a known attack development of a new attacK and various combinations Continued development of more useable available and powerful computing resources for Analysts is the fundamental means of automating the more mechanical parts so that Analyst resources can be focused on the more intellectual For example some known old hand may be just what is needed for some new problem However effective exploitation and utilization of personnel may require the development of new computer systems Although the old hand attack might already seem to be sophisticated when it is carried out by a human processor additional sophisticated techniques may be reqUired to move it to a computer processor It may not be possible to move all of the human processing to a computer because some aspects of the human processing may not be well enough understood to be mechanizable This requires a close cou pIing between intelligent human and mechanical computer processors to produce a complete system solution As the target problem becomes more difficult in techniques or scale the system development activity becomes correspondingly more difficult Approaches to deal with these pressures share the idea of developing facilities to support the Analysts The facilities can be viewed as tools that can be applied to various parts of a problem A tool proVides a way to improve the interface between the human problem solver and the problem Those parts of the problem which are well understood can be automated so that valuable Analyst time can be devoted to the more interesting parts This is a continually-evolving process which raises the level at which the Analyst operates Suppressing the details that are no longer of interest exposes new and more interesting problems that were previously hidden attacks Mar 82 CRYPTOLOG UNCLASSIFIED Page 32 o L 8 6- 3 6 DOCID 4009842 UNCLASSIFIED Responses to the pressures are in the form of computer systems supporting increasing amounts of interaction In time these systems will evolve to become environments which will support almost all of an Analyst's activities This kind of response transfers pressures on the Analysts to pressures on the system Developers The general approaches used to respond to Analyst pressures can also be applied to system Developer pressures In the remainder of this paper the term User will refer to any person using a system to assist him in carrying out some activity Since we will be addressing system development for Users a brief sketch of system development from a User's view is useful to establish some context for what follows From the User's view system development starts with some notions of a need for a system to formulate system requirements After the requirements are formulated the system is developed and eventually delivered so that it is available to the Users To the User the system development activity looks like a delay between the formulation of requirements and the system's availability Although the Users may believe that the requirements embody their notions of what they need and the Developers may believe that the system developed satisfies these requirements the delivered system often contains some surprises tools and their importance include the following Tools are not used in isolation they may be used in combination with other tools J It is important that they can be combined into larger ones Developing tools which can be combined to form new tools is fundamental to the ability to build upon previous work rather than repeating it Tools interact with the User they may be completely automatic J It is important that excellent human factors be a major design objective in their development Tools interact with a variety of information and knowledge sources J It is important that these sources an easily accessible part of environment Before trying to answer this some discussion of the kinds of systems we are evolving to will be useful Tools may be applicable to a class of problems which differ in only some specific parameters Tools need to evolve as approaches to it evolve Just as tools have been and will continue to be developed to help the User solve problems the development of the tools themselves is a problem requiring tools The development of requirements for a system to solve some problem for a User is clearly a serious problem n its own right This is evidenced by the difficulty in deciding what problem is in need of a tool what kind of tool needs to be developed how the tool is developed how it will be used who will use it where it will be used and the like In fact the very existence of a tool for a problem and experience using it often provide useful and critical insight into the problem and expose the need for new tools Some general properties of Mar 82 _ the problem I systems have no intl'insic vaZue except to the extent that they satisfy their usel'S 1 -- are the and It is important that the tools can be evolved rapidly to maintain their applicability Tools and their Environments data J It is important that given the identification of the parameters it is possible to develop a single parameterized tool to insure the consistency of the different members of the class WHY I not CRYPTOLOG UNCLASSIFIED Page 33 DOCID 4009842 UNCLASSIFIED the various aspects of the development activity The structure imposed on the activity tends to be proportional to the perceived difficulty of development For some problems that are well understood requirements can be written by the User community and given to the Developer community If the resulting product satisfies the requirements then the Users should be satisfied with it This is a fictional approach for most problems of interest user expectations are rising as a result of previous development successes and the decreasing cost of system components The development of tools that satisfy these properties is itself a problem that can benefit from tools of its own Central to tools for the Users and the tools to produce them will be the use of interactive computing systems Time-sharing systems have been used as the basis of most interactive systems and are reaching limitations in the ability to support adequate human factors and desired functionality The decrease in hardware costs has reached the point where personal computing with network facilities is an attractive alternative The importance of a network cannot be overemphasized Personal computers alone are not a step beyond time sharing if they do not provide other kinds of resource sharing in a transparent fashion Traditional programming languages and techniques could be used but more advanced ones are emerging which also provide attractive alternatives The Requirements Problem We have seen the trend from pure requirement-generation by User organizations passed on to development organizations evolve to joint User development requirementgeneration Although this is a step in the right direction experience has shown that joint requirement-generation does not go far enough Systems built from joint requirements may not meet the expectations of the participating parties When the systems are completed and the problems discovered it is often too late to take necessary corrective actions The way in which systems are developed is heavily influenced by the perceived cost of Mar 82 This model takes a simplistic view of the need for communication between Users and Developers It does not provide for any feedback from the Developers on their understanding of the requirements until the product is delivered In an attempt to narrow the gap between the Users' and Developers' understanding of requirements joint User Developer requirement generation emerged The rationale was that if both parties are involved in writing the requirement then the chance of getting the desired product should improve Although joint requirement-generation does yield some improvement in the communications area it still takes a simplistic view of the role of requirements The generation of requirements assumes that enough is known about a problem to say what is needed Even if enough is known about a problem how can requirements be written so that any two people involved with them agree on what they mean And if not enough is known about a problem then requirements are even harder to write In reality the problems are sufficiently difficult that not enough is known to write complete requirements and it is not possible to write them in a way that guarantees uniform interpretation There have been a number of significant efforts to address the requirements problem and still it persists After some period of little progress one begins to look in other directions Could it be that there is something wrong with the idea of developing requirements or the way in which they are developed This will be explored next and alternatives considered The Prototyping Approach Requirements are hard to write and even when well-written and comprehended by all involved can lead to unexpected results When CRYPTOLOG Page 34 UNCLASSIFIED ---'--- _ __1 DOCID __L- _ _ t I_ _ 4009842 UNCLASSIFIED the system is completed and the problems are discovered it is often too late to take the necessary corrective actions This is because it may not be possible to change the system without substantial delays The form of a developed system may be very sensitive to some of its design decisions Changes to such critical design decisions can have pervasive effects on the developed system so that major parts of the system may have to be modified or developed again The translation of a system design into a developed system has traditionally been a manual one In addition the target of the translation even if it is one of the common higher level languages often requires detailed design decisions to be made prematurely just to get the system working These languages do not provide a convenient means of adjusting the level of the target to one that is high enough so that lower level design decisions can be deferred As a consequence of these factors the cost of the software development activity becomes high enough that emphasis is placed on trying to do as much as possible before any software development begins Although it is always considered a good approach to put effort into preparation the preparation and performance activities need to be carefully balanced It is becoming increasingly clear that the usual way in which requirements for a system are developed is not adequate because of the increasing frequency of surprising undesirable results of the translation of requirements into a system This is an example of too much emphasis placed on a form of preparation which does not adequately identify the important issues early enough Among these issues is how the system is really expected to behave which the requirement was formulated A product of the requirement activity can be a prototype which is used to start the development activity In the prototyping approach development becomes a prototype refinement activity Although traditional programming languages and techniques could be used they have some significant limitations These limitations are the same as those which led to the traditional system development approach An understanding of these limitations and a more advanced approach to programming languages and techniques that deals with them can break this cycle The limitations and their consequences include o The low level to which a design must be carried out to produce a working system because the language level is fixed and far below that of the problem domain J As a consequence more detailed and otherwise relatively irrelevant decisions may have to be made just to get a prototype developed o There is no explicit form of the design and no way to transform it into a working system except manually o Experience with system development has repeatedly shown that more insight is obtained by the use of a system than by simply trying to read about what is expected of it Given a system and its requirements the behavior is more readily evaluated from its use than from thinking about how it might be used This is not to say that all aspects of system evaluation can best be comprehended through examples In fact there are many important properties of a system that cannot be effectively determined by its use For example bugs can be found by using a system but it is not in general possible to know when all the bugs have been found The point is that a prototype of a system can be a very useful tool during the usual requirement and development activity To be most effective prototype development needs to be rapid so that it can be quickly made available to the User for Mar 82 CRYPTOLOG As a consequence changes in design decisions have to be carried out manually making changes very expensive and limiting opportunities to learn from experience A valuable product of the development activity is lost for future redevelopment The only real product is the crystalization of the design Although this is generally too fragile to be changed it is the only thing available short of developing the system again manually development aapaaity is a limiting faator in satisfying user requirements UNCLASSIFIED Page 35 DOCID 4009842 UNCLASSIFIED Unfortunately this does not remove the tension between the User and the Developer it just changes its character The decrease in cost of system components raises the expectations of the User The problem then becomes one of combining components to meet these new expectations This results in the problem solving time of Users and Developers in their respective domains becoming the critical resource The User wants new tools to help solve problems and the Developer needs to produce those tools But the development of tools is itself a problem domain which can benefit from tools The right tools have the potential of reducing development costs so that the Users get their tools sooner Tools to produce tools could become tools of the User if we had perfect insight then systems could be developed from their requirements but we don't Some Familiar Consequences of this Tension A more advanced programming language and development technique should address these two classes of issues directly Programming languages are needed which can be easily extended to problem domains characterized by the appropriate abstractions including notation objects and operations of the domain System development environments are needed which can support a multi-level design which captures the design decisions The combination of these two facilities can support the rapid development of prototypes the development of designs their transformations to production systems and their redevelopment The Tension between Users and Developers When new computer systems are being developed for Users care must be taken to insure that the resulting system is reasonably close to what is best for the User rather then what is best for the system Developer Each usage of best is relative to a different point of view based on the subject to which it is applied For example if the notion of best is that of most cost-effective in solving the problem the independent interpretation will still lead to conflicting results This is because the User view considers problem-solving effort of human resources as the most valuable while the Developer view considers the system itself as a valuable resource The proper point of view lies somewhere in between these two extremes Given the pressures on the User and the advancing state of computing technology it is clear that the point can be put very close to the User in the future As a result the situation dramatically changes because system resources have and will continue to dramatically reduce in cost Mar 82 Some consequences of this tension being resolved in a way that is not favorable to the User demonstrate the need to be more sensitive to the User view This goes beyond the system Developers because in some cases new technology is needed to fill important gaps Inittally the benefits of such systems outweigh their limitations After some experience with them the User expectations are raised and new issues are exposed There is a tendency to try to use a given system for broader applications Unfortunately the existing systems do not scale up well There are some systems which are much closer to the ideal in some areas than others but these have not all converged into a complete system yet Applications to New Technologies A popular topic in future computing technology is personal computing Although it seems clear that this technology offers many advantages over even existing advanced timeshared systems there is much to be learned about it At this time it may be hard even to write requirements for future systems based on new technology because of our lack of experience with it A useful way to gain experience in a new technology is to set up a laboratory for research in the area Some early explorations have been done in various places including R53 the Office of Computer Science Research which already has a strong foundation in the underlying technologies An example of such an approach can be found in the paper A Laboratory for Developing Personal Machines by 1 '--- CRYPTOLOG _ Page 36 P L UNCLASSIFIED 86-36 DOCID 4009842 UNCLASSIFIED Advanced System Development Environments No matter how advanced the hardware environments may be they cannot be used effectively without advanced system development environments An important component of an advanced environment is an advanced programming language But an advanced language alone is not a sufficient goal An advanced environment is needed to support the use of the advanced language throughout the lifecycle of the system Some limitations of most existing environments will be discussed and some indication of what more advanced ones are likely to be like will be given The limitation of the traditional programming languages including the higher level ones is the size of the gap between the conceptual level of the problem domain and the language Solutions to problems are expressed as programs If the programs can be written in a notation natural to the problem domain then the gap is very small As the problem domain moves further from that supported by a language more effort is required to reformulate the natural problem-domain expressions to the lower level supported by the language If the problem domain is one for which the language is designed then solutions will be easy to express in the language However if the domain involves concepts that are not in the language then problem solutions will also have to solve a second problem caused by the language itself The additional language problem may be sufficiently difficult to divert attention from the problem which was being solved in the first place Programming languages which are advanced in this way and which can be used in an interactive environment are not just tools for the Developer to produce systems for the Userj they can themselves become tools for the User The importance of the environment must not be underestimated Even with an advanced language additional facilities are needed to support its use These facilities need to have excellent human factors if they are to be effectively used by people throughout their activities Personal computing will provide the quality of computing resources needed to support the human factors Just as languages are not a sufficient goal environments to support their use are not sufficient either Systems to support system development or for that matter the NSA Analyst are only as good as the people that use them After all the systems are only mechanical embodiments of some things which the Users understand It is still up to the Users to have the intelligence to take the step beyond Conclusion The development of systems has been explored with respect to the pressures on Agency Analysts rapid advance of available computing technology rising expectations of the User community limited capacity to produce new systems and increasing difficulty of formulating meaningful requirements Among the various approaches that will be used to deal with future analytic problems will be an increasing use of interactive computing systems to prOVide advanced environments Human factors requirements are among the most difficult and most critical aspects of these future systems Prototyping has been suggested as a valuable experimental component in the requirement and design process Of course special dialects of some languages have been developed to support more natural notations for this problem domain with these data types and operations But suppose that some problem is best expressed in terms of another data type An unnatural notation will again be needed This problem exists within the very limited domain of numerical expressions The non-numerical world has a very rich collection of data types In many cases they are created for each problem From a human factors point of view the natural notation is the best This is an example of the tension between Users and Developers in the domain of programming languages Mar 82 CRYPTOLOG the use of a prototype system can reveal important insights that are not apparent from a requirements document UNCLASSIFIED Page 37 DOCID 4009842 FeR 9FFISIA WSK QWbJPY Ql lb '4our ilooks Nrurr itr uJ by William Slt Nnltr ID54 n addition to sundry other forms of aging paper the History and Publications Division T542 maintains a file of telephone directories prepared by NSA and predecessor organizations This file when used in conjunction with old organization charts and other documents has been a valuable resource in preparing books and articles produced as part of the Cryptologic History Program The directories are also helpful in answering the numerous reference requests handled by this office o They supplement our understanding of changes in Agency organization and frequently provide the most readily available source for the correct spelling of an individual's name or the organization to which an employee was assigned at a given time J Directories currently on file include Date Jan Sep July Oct Apr June Dec Station or Organization 1946 1948 1949 1950 1951 1951 1951 Vint Hill Farms Station Arlington Hall Station AHS Naval Communications Station Naval Security Station Naval Security Station AHS Naval Security Station Mar 82 CRYPTOLOG Feb Oct Oct Apr Dec Jan Oct Oct May June 1952 1952 1952 1953 1953 1954 1954 1954 1955 1955 1956 June 1956 July 1956 Fall 1956 Nov 1956 Mar 1957 Spring 1957 Fall 1957 Oct 1957 Mar 1958 May 1958 Oct 1958 Jan 1959 Sep 1959 Jan 1960 Jan 1961 Oct 1961 Mar 1962 Mar 1963 Mar 1964 June 1964 July 1965 Sep 1965 AHS AHS Naval Security Station NSA NSA AHS NSA AHS AHS NSA NSA Organizational Supplement NSA NSA PROD interim listing AHS NSA Organizational NSA Organizational AHS AHS NSA New Site NSA Organizational NSA NSA NSA Organizational NSA NSA Organizational NSA AHS NSA NSA NSA NSA NSA AHS Page 38 FeR 8FFlelkb BSE BUb 1 DOCID 4009842 FBll BFHSIM SSB Bllb'l July Apr May July Jan Jan Jan Jan Jan Jan Jan Sep May Apr Nov 1966 1967 1969 1971 1972 1973 1974 1975 1976 1977 1978 1978 1979 1980 1981 NSA NSA NSA NSA NSA NSA NSA NSA NSA NSA NSA NSA NSA NSA NSA W fDl '6ot to haVe 01 4 eR tPT Dt O 1 4 b$cl i i l h In DVecI il 1Ie f P L 86-36 tot Unless described as organizational in format and organization all the directories in the file are arranged alphabetically by employee name Some of the organizational directories list all or most of the individuals assigned to a particular element but most list only the chiefs or other principals Though useful for many reference purposes the directories are not infallible The move to Fort Meade and one or two major reorganizations must have been confusing and stressful for everyone involved but the persons responsible for the telephone system must have been especially pleased to see the dust settle after such reshufflings One practice of interest long since abandoned was the use of directories to reinforce security consciousness Dividing each letter of some directories of the late 1950s are security slogans of the sort seen on posters and other devices Dignity of man can be shattered by a careless tongue cautions one such reminder in the May 1958 listing The same edition also warns that A secret's a secret only as long as its kept a truism that appears only a few pages after the -listings for William H Martin and Bernon F Mitchell both of REMP 13 on 5323 secure 7147 outside Agency personnel holding outdated directories not listed above may wish to consider forwarding them to T542 for inclusion in the file Such action entitles the donor to membership in the Pack Rat Society and the gratitude of the staff By the same token this office also would like to acquire Baltimore and Washington area white page directories from the 1940s through the late 1970s We do not need every year within that span nor do we need 247 copies of the 1979 editions Persons with directories they may be willing to part with can call 2355s before putting them in the mail Mar 82 CRYPTOLOG Page 39 F9ll BFFISllm SSB BNb'l US DOCID 4009842 UNCLASSIFIED i HUMAN FACTORS CORNER Consumer vs Computer i I BY 1 --- P L lp13 REVIEW Consumer Difficulties with Computerized Transactions An Empirical Investigation T D Sterling Communications of the ACM Vol 22 No 5 May 1979 pp 283-289 If you are among the overwhelming majority of citizens who have wondered again and again why computers make so many mistakes this article has much to offer in the way of enlightenment if not in hope for improvement If you share my suspicion that many of the goofups in bills and mailings are unnecessary consequences of bad design and poor management Dr Sterling's paper will provide confirmation But the real surprise to me at least is the strong indication that a considerable nortion of the snafus are a consequence of deliberate and questionably honest business practices aimed at putting and keeping the customer at a disadvantage The answer to why computers make mistakes is only sometimes because programmers or data-entry clerks or system designers make mistakes In some proportion of cases the answer may well be because omputers were deliberately programmed to make those 'mistakes' Mar 82 Reprinted from Human Factors Letter 1-80 published by CISI Human Factors SIG Noting that errors in computer produced bills And various communications sent by government and other agencies or businesses to citizens clients customers and consumers are both commonplace and embarrassing to the computing industry oo Sterling asks the following questions How many and what types of errors occur What is the reaction of consumers after they encounter errors Few if any studies have been made to answer questions like these a fact which in itself seems strange to me In 1977 a study was made by an organization or individual that Sterling calls the Computer Ombudsman of Vancouver in cooperation with the Consumer's Association of Canada B C This paper discusses the study's findings From the point of view of data gathering methods in human factors work this type of study comes under the general heading of critical incident studies focussing on incidents where a human-machine CRYPTOLOG UNCLASSIFIED Page 40 1 86-36 DOCID 4009842 UNCLASSIFIED system broke down malfunctioned or occasioned some class of events or outcomes of interest in evaluating its strengths and weaknesses A questionnaire was mailed to a random sample of members of the Consumer's Association asking for experiences of household members with computerized transactions In-depth telephone interviews were then conducted with people who reported specific problems in the auestionnaire The Dopulation was according to Sterling representative of a broad range of middle-income groups He frankly admits that it was probably rather atypical in one respect namely the consumer orientation and the consequent aggressiveness and willingness to do battle in getting errors corrected It was found that 59 5% of the households reported no problems with computerized transactions but 25 1% had one 10 4% had two 3 6% had three and an eSDecially unlucky 1 4% had four such problems during the preceding twelve months By far the largest proportion of the errors were related to department stores credit cards services utility companies and mail order business Billing errors accounted for 81 2% while all dealing with governments accounted for only a tiny 2 4% of the goofups When we see what the study showed about the different types of billing errors a rather suspicious pattern seems to become even clearer Almost all the errors result in a net financial gain for the billing entity they include such things as charges for nonexistent expenditures inappropriate charges of interest and overcharges Sterling points out that errors in the consumer's favor may well have been under-reported but the balance is overwhelmingly in the other direction In 73% of cases respondents succeeded in getting the error corrected Whether success attended their efforts or not attempts to get errors corrected required one contact in 36 6% of cases up to three contacts in 30 4% and four or more contacts in an Rppalling 33% of -the cases It seems evident that both victims and offenders were admirably persistent and determined Sterling points out that each contact between a client and an organization requires additional 'overhead' time The pumber of contacts is a measure of effort The average time from discovery of an error to its settlement was 8 weeks but 15% dragged on for 20 weeks Consumers spent An average of 2 6 hours of their time in the attempt but 20 hours or more were expended in 20% of the Mar 82 cases In 10% of cases it took extra effort to remove incorrect interest charged on top of incorrect charges and in 2% of cases it was never removed Trying to correct an error was often made unpleasant for the consumer 16% of respondents reported having been coerced in some way to pay a disputed bill often with an implied threat of damage to their credit rating 8% were treated as troublemakers Some customers while continuing to deal with certain businesses that frequently offended elected to pay cash in order to avoid the hassles or to buy a magazine at a news stand rather than maintain a subscription While Sterling does not explicitly apportion blame for the error-prone practices he presents important evidence that points toward deliberate management philosophies as well as bad design of computing packages Another study made in 1976 showed a 7-day delay on the average between the billing date and the time the bill was mailed with a range of from 3 to 20 days delay In the 1977 study most of the victims of incorrect interest charges were also victims of this late-mailing practice In addition it is clear that in almost 60% of the cases where interest had been charged on some disputed amount it was not removed at the time when the charge turned out to be incorrect This must be a consequence either of bad program design or intentionally poor service Sterling offers the following in summary We might ooo ask if wi th the use of all the power computers add to the task of management it is really necessary to have systems that fail to correct errors fail to inform customers where they may address a complaint fail to remove interest in cases of billing errors fail to adjust the printing of bills to the capacity of an organization to mail them out and fail to do all the other things which careful examination of existing packages seems to uncover He expresses a valid concern about the trend in the business community toward electronic fund transfer and a cashless economy What will happen when the visible audit trail such as checks bills and receipts is replaced by electronic signals Next time you get an erroneous bill in the mail don't blame the computer the programmers or the data entry clerks Instead consider placing the blame on business and management practices that encourage the design of bad and dishonest commercial software packages CRYPTOLOG UNCLASSIFIED Page 41 DOCID 4009842 'fep SfleRfl'f UlIBRA Dear CRYPTOLOG Editor U I am shocked that you would print a personal note from me to you in CRYPTOLOG see Letters to the Editor CRYPTOLOG January 1982 without my knowledge or consent As a result of your action a person of U I sex ran into my office bared its chest and demanded to know if it won the Strangest Bust of the Month contest and what was the prize U A bust is not only an arrest when you are caught with the goods a marble statue of George Washington etc but it is also a cryptographic abnormality e g an error in encryption a fault in equipment a violation of rules for use that may lead to some analytic understanding or exploitation of a cipher system Mar 82' CRYPTOLOG EN g c P L 86-36 DOCID 4009842 F8R 8FFISVd BSE 811F publication P L 86-36 A Toy Problem U U Sorry no prizes by David J Tiren 9 P L B63 x5311 U Determine used I I think it wasl who said on more than one occasion Whatever I get I publish Or maybe it was Art Salemme Ed Stations 1 the 1 callsign 1 system 2 being 8 Day 1 DRQ YNM KPH ZLH TIH HWZ QIM OWl Day 2 FGT QBP ESF ZOE UAE ETI TAP RTN Day 3 LJI SFD BHV RCV XWV VUA IWD GUB Day 4 CON VKJ AMY UGY WZY YXW NZJ JXF Dear Wayne U A gentleman in one of the operational groups has correctly pointed out that Mark Twain died in 1910 and not 1909 as I had indicated in my article The Literary Bends January's CRYPTOLOG I respect his wish to remain anonymous But to set the record straight and maybe to do a little more drumming and trumpeting could you publish my answer to his letter U Answer next month the original answer from Dave came to us in a plastic bag if that's any help to you Some Thoughts on Lexicography S tuart H Buck CRYPTOLOG September 1974 Dear Sir You're right Sam died in 1910 Unfortunately I implicitly trusted John O'Hayre's statement on page 106 of his Gobbledygook Has Gotta Go I referenced his book in the footnote to the article He says it was 1909 What to do I guess not much can be done other than to express my sentiments like so Glancing through the latest bulletin of the Mongolia Society my eye came upon the following remark by John Krueger professor of Altaic Studies at Indiana University 'The very worst possible way to make a dictionary and ooo the way that nearly all appear to be made is to make a grand compilation of all existing dictionaries ooo ' For those who are interested here is remainder of the quotation Shall I point at O'Hayre and say he is the one Who caused the mistake I should never have done No I'll just throw up my hands in spite of the shame And cite that cruel law that uses my name I very much appreciated your comments Sincerely Al Murphy E4l Thanks Wayne SOLUTION TO NSA-CROSTIC No 38 ooo possibly abridging slightly and adding a few examples The obvious and ideal way seldom followed is to begin with a set of texts draw from them only the words used in those texts and create a dictionary out of the actual recorded usage of the literature or in the case of a spoken dialect from the noted speech of the speakers The latter and better course is self-evidently vastly more difficult and time-consuming AM Mar 82 CRYPTOLOG the Page 43 FeR eFFrerkh BSE SlIbY 86-36 DOCID 4009842 ' -' ' l I KRYPTOS Society t4ews U 1 4 d P L 86-36 ce eee I twill be the guest speaker at the spr ing meeting of the KRYPTOS Society He we tell us some Tales of the Unexpected This talk is based upon var ious manual cryptosystems with unusual twists that he has solved in his long career as a cryptanalyst I o IU His talk will be in the Friedman AUditorium o 12 May KRYPTOS Society _embers will be seated up to ten minutes before the talk non-memfers eftetHerd ' In addition a cocktail reception for IWill be held in the late afternoon Details will be announced IU The KRYPTOS Society is still accepting members Our last meeting featuring Mr Frank Raven was an overwhelming success For membership information contact Larry South E7 8153s 4718s Council P L 86-36 Mar 82 CRYPTOLOG Page 44 BBeRf '1 1 DOCIO 4009842 UNCLASSIFIED FOR RENT -r Reasonable Mar 82 P1-Apr 82-53-15139 CRYPTOLOG UNCLASSIFIED Page 45 This document is from the holdings of The National Security Archive Suite 701 Gelman Library The George Washington University 2130 H Street NW Washington D C 20037 Phone 202 994-7000 Fax 202 994-7005 nsarchiv@gwu edu