WUJVVUlWUJl1 5lBrnU1VV D $UJlB L10 V 13lD fil13 B UJ13tlJ 3 UJ W GBPijW f 3rd Issue 1992 t U 5t 1 EO P L 86-36 George Jelen THE THIRD OPSEC CONFERENCE o o o o o A QUANTITATIVE APPROACH FOR EVALUATING A F ACI LI Y1STECHNICAL SECURI TY o o ooooo o o o GOLDEN OLDIE 1 1 3 o '---- -J1 o 9 ---fL 10 16 '-- I o o WRITING IN A JOURNALISTIC STYLE o o L MINIMIZING THE BURDEN OF MAINTAINI1'lG SOFTWAlU Chris tine McCormick m 2l 1i eel ssi-ri d a n d A 'fillS DOCUMENT roved lor Release b eON'fMN N A on 10 o o o L D callimahos o P L I L l 86-36 86_3 6 23 Jb6 o Richard Sylvester 44 45 o o o 46 o o o 47 o o 48 o o o oo o 49 L I 10 201 2 UUIU JWUIlY MllTl jKlt b Lum u I CETA o o o o o o oo A HISTORY OF CRYPTOLOGY HARDWARE-SOFTWARE lU VIEW MAT NSA AND COMPUTER VIRUSES FROM THE PAST o LETTER o o BULLETIN BOARD o EDITORIAL o o o o o TO CONTRIBUTE o o o o 1 ur u nt to C 13526 'v1L lR C ase # 5 4 7 7 CLASSIFIE9 BY SN SSM 1J J E EC J SSIFY ON Origil l8tiflg I geAey's De t eFFFlifl8 t iol l Required NOT RELEASABLE TO CONTRACTORS DOCID 4036135 CRYPTOLOG Published by P05 Operations Directorate Intelligence Staff VOL XIX No 3 3rd Issue 1992 I PlT BLISFlER BOARD OF EDITORS 1 EDITOR t963-7595 Collection Marian Brown 963-1197 Computer Systems 1 963-5877 Cryptanalysis 1 96 1461 Cryptolinguistics 96 382 Information Resources 963-3258 Information Science 963 3456 968cg013 Information Security I Intelligence Community 1 1 963 5800 Language I 1GBP903 30 57 Linguistics I 9Ei34814 Mathematics I 963 3709 Puzzles 9 3 1461 Research and Engineering I 1 96 1-83 6 2 o Science and Technology 1 96 3544 Special Research Vera R Filby 968-6558 Traffic Analysis I i 963-33 j Classification Officer 1 Unix Support 1 Macintosh Support 1 Xerox Support Illustrators I I v P L 8 6- 3 6 rl963 5463 1 963 3369 96t 8362 963-6867 963-3360 963-4382 To submit articles and letters please see last page For New Subscription or Change of Address or Name MAIL name and old and new organizations and building to Distribution CRYPTOLOG P0541 OPS-1 f P L I I Please DO NOT PHONE about your subscription or matters pertaining to distribution Contents ofCRYPTOLOG may not be reproduced or disseminated outside the National Security Agency without the permission of the Publisher Inquiries regarding reproduction and dissemination should be directed to the Editor All opinions expressed in CRYPTOLOG aTe those of the authors They do not represent the official views of the National Security Agency Central Security Service FQK QFFIClIhL YSE ON18 86-36 DOCID 4036135 The Third OPSEC Conference George Jelen D2 Good morning It is my pleasure to welcome all of Initially applied to military operations during the Vietnam War OPSEC is now seeing application in you to the third annual NSAlCSS OPSEC day a rapidly widening set of circumstances and activiWe have all heard ofthe Age ofEnlightenment and ties The Secret Service is applying it to personnel many ofus have lived through the Age ofAquarius protection the FBI to law enforcement the deWell I'm here to proclaim this as the Age ofOpera- fense community to weapon system acquisition tions Security After being around since the days of the Coast Guard and the Customs Service to drug the Vietnam War Operations Security or OPSEC interdiction and the Intelligence Community to has finally come ofage I attribute this to two causes clandestine and covert operations At a national first partly as a result of the 1988 presidential level we are now examining how we might employ directive more and more people have become ex- OPSEC to help protect critical economic informaposed to OPSEC and have learned of its potential tion as well This latter application is clearly the This has led to the application of OPSEC to many growth area of the future different kinds of operations and activities And second people are beginning to discover that as they OPSEC has also found varied application here at are forced to shrink theirbudgetsfor security OPSEC NSA We have applied it to site clOS'ure plans to and its methodology can be very useful in choosing SIGINT support to military operations to where to focus their security protection and in apply- sensitive acquisitions and procurements to personnel movements to financial transactions to ing that protection uniformly and consistently logistics shipments and to counternarcotics There are many indications that the interest in support The purpose in every case has been to operations security is picking up OPSEC tracks improve the effectiveness of our operations We are now being offered as part of several major have come to understand that whenever there is security conferences including the International some advantage in concealing our intentions Security Systems Symposium and the conference OPSEC proves quite useful of the American Society for Industrial Security There is a second reason why operations security The National OPSEC Conference itself has been growing in attendance and the membership in the has been receiving more attention oflate and why this is likely to continue With the collapse of the OPSEC Professionals Society is continuing to former Soviet Union the U S has entered a period expand steadily The last couple of years have during which our expenditures for security are also seen a developing body of OPSEC literature being challenged The changed world situation Here at NSA a COSC has been established for the and the altered security threat that it has brought OPSEC field and NSAlCSS OPSEC Association with it have caused many to question the continhas been instituted and is aggressively planning activities and there is a panel actively working to ued need for security protection The question I hear all the time is Where's the threat It's a create criteria for professionalization 3rd Issue 1992 CRYPrOLOG page 1 6ft 6FFICtNL Ufffi fflftJJPY DOCID 4036135 reasonable question Most would answer that there is still a threat but that it is reduced and it is directed differently - focused more on economic and technology secrets than on military secrets Motivated by a need to reduce our expenditures and encouraged by this generally-accepted reduction in threat resources for security are being cut As this happens OPSEC and its methodology become ever more useful As resources applied to protection are reduced if we are not to sacrifice too mush security in the process this reduction needs to be done sensibly Two basic premises of the OPSEC discipline are that not all information justifies protection and that not all vulnerabilities are worth correcting We need to have some way to distinguish between what really needs protecting and what does not Presently we are spending far too much money trying to protect information that is either not worth protecting is already known or is fundamentally unprotectable This makes no sense and we can not afford to continue to do it This is where operations security can help The more important it is to be selective in the application of our security resources the more relevant the OPSEC methodology becomes As resources for security grow more scarce it becomes that much more important to apply them where they are most needed and where they can do the most good The proper application of the operations security discipline and its methodology can be extremely useful in sorting out what most needs protection and in making sensible decisions about where and where not we can best afford to cut resources Our decision process needs to weigh the importance of the information the motivation and the capability of our adversary the ease with which that adversary could obtain that information and the risk ofleaving the secret unprotected versus the cost ofprotecting it All of this is precisely what the familiar five-step OPSEC process does and all OPSEC professionals know how to do it The OPSEC process imposes a rigor that can be profitably employed in many security resource decisions sometimes with dramatic results For example when the U S was preparing for the arrival of Soviet inspectors as a result of the Stra- tegic Arms Treaty teams went around to a number of contractor facilities and Air Force bases looking at what special security arrangements would be required Applying the OPSEC methodology the teams were able to reduce the projected expenditures for security by more than seventy million dollars Once we have identified the information that is in most need of protection it is equally important that we apply security resources consistently and completely - that we do not spend money on a robust lock for the front door and leave the back door unbolted Here again OPSEC can be helpful The various traditional security disciplines do a pretty good job of protecting against direct disclosure But our secrets can be revealed indirectly as well as directly and OPSEC complements these other disciplines by seeking also to protect those same secrets against indirect disclosure as well Failure to consider ways in which an adversary might piece together the same secret from bits and pieces of information could mean that we spend a considerable amount of money in security protection and give away the secret anyway Without operations security the envelope of protection is incomplete We in D2 understand that our organization can not perform OPSEC for NSA or th CSS OPSEC by its nature cannot be centralized To implement OPSEC effectively it has to become part of the normal way everyone of us conducts our daily business In other words the NSA and the CSS will only have effective operations security when every single person understands it and practices it within his or her organization Everyone has to know what it is how it works what are its goals methods and mechanisms That is what this day is all about It is aimed at deepening your understanding of OPSEC so that you might be more able to apply it to your own activities Or employing the words ofthis year's theme the day is intended to help you keep your mission on target through the sound application of OPSEC principles I am pleased to see you all here and I hope that many of you can remain for the rest of the day 3rd Issue 1992 CRYPl'OLOG page 2 FOR OFFIGJA I USB ONLY DOCID 4036135 A QUANTITATIVE APPROACH FOR EVALUATING A FACILITY'S TECHNICAL SECURI1Y IR551 This Article is classified CRE'I' in its entirety Center for Security Evaluation CSE has been working on various initiatives to support the decision process in evaluating technical security at facilities One recent effort referred to as the Aggregate Countermeasure Effectiveness ACE model helps to provide a quantitative measurement that integrates the multitude of factors that impact technical security The model takes into account the value of the targets in the facility the capability of the threat and the overall effectiveness ofthe associated countermeasures A prototype of the model has been developed and is currently being evaluated by CSE with promising preliminary results P L 86-36 factors that impact technical security at a facility Results from the model would provide a quantitative barometer that could be used to compare the degree of technical threat at one facility with tha1 0 1 4 c of another The model had to be understandable p L 86-36 represent the overall state of technical security mSTORY In early 1991 the Standard Division of CSE decided that a new method was needed for evaluating the technical threat to a facility At that time a means was needed to help decide on the best combination of Technica1 Surveillance Countermeasures TSCM for the conditions particular to each facility To this end CSE set the wheels in motion to develop an analytical methodology that could be used to help make more informed decisions in two specific areas namely o in establishing TSCM standards and policies and o in evaluating overall TSCM investment strategies The intent was to support decisions in three general areas First to help evaluate alternative set o 1 4 c of TSCM either by looking at strategies specific t6' L 8 6 - 3 6 one facility or policies applicable to U S facilities in IZeneral r L - Jf Finally the model would be used to examine proposed resourcing concepts either at a particular facility or across a selected set of factilities The model was to address such THE INTENT OF THE EFFORT Simply stated the goal was to create a mathematical model that would tie together all of the 3rd Issue 1992 CRYPTOLOG page 3 Si CUT EO 1 4 c P L 86-36 DOCID EO 1 4 c F L 86-36 4036135 CRYPTOLOG page 4 SKCRI 1T 3rd Issue 1992 ii EO 1 4 c DOCID EO 1 4 c P L 86-36 4036135 using basic engineering principles 1 I THE SOLUTION WHAT THE MODEL DOES It was decided that this new methodology re- ferred to as the Aggregate Countermeasure Effectiveness model or just ACE would be created in two stages to expedite the development process and to provide a quick turnaround product that could be readily evaluated for its usefulness In the first stage a prototype microcomputer-based software package would be built for a reduced portion of the problem If the prototype is determined to be of value then a production version with full capability and user-friendly features would be created To develop the model technical security at a facility was analyzed as a general flow problem o DESCRffiING THE PHYSICAL PROBLEM For this flow model concept to work o a series of mathematical expressions were developed that 3rd Issue 1992 CRYPl'OLOG page 5 SI lORB't' EO 1 4 c P L 86-36 DOCID 4036135 1 4 c P 86-36 represented the relationships between the amount of sensitive material potentially compromised at a facility and the multitude offactors that determine its state of technical security the expression at the bottom of exhibit 1 illustrates a few of the factors The first step in translating the real world problem exhibit 2 to a mathematically oriented model was to define the physical relationship between the targets at the facility conversations between people workstations copiers computers etc the TSCM intended to protect the sensitive material processed by these targets I and the various techniQues used by the threat HOW ACE WORKS There are three basic parts to ACE exhibit 4 First the value of the sensitive material processed at the facility is estimated for each target based on the volume of activity and the worth of the material Second the effectiveness of the TSCM is computed for each path by summing the attenua3Td Issue 1992 CRYPI'OLOG page 6 SECRE'f EO 1 4 c P L 86-36 Eo DOCID 4036135 PHL 86 36 3rd Issue 1992 page 7 DOCID 4036135 tion provided by each individual countermeasure This in tum is degraded by the state of operational security at the facility both in terms of the adequacy of the preventive maintenance program and the level of p3 personnel procedural and physical security The natural attenuation provided by the controlled access distance is then added resulting in an estimate of the combined WHERE ARE WE NOW EO 1 4 c P L 8 6- 3 6 In October 1991 the prototype version of ACE was completed The microcomputer-based software package is being evaluated by CSE and run through a variety of scenarios The purpose is to perform a-preliminary sanity check does ACE effectiveness of the TSCM for each path Finally the capability and intent of the threat is brought into play provide reasonable results Can we explain and -- understand what is happening To help evaluate the model sample cases have been run using rough estimate data readily available on three facilities The preliminary findings are very encouraging The sample runs have provided realistic results and have highlighted interestingopflervations Eo 1 4 c F L 86-36 At present ACE is going through an extensive shakedown to validate its consistency and to examine its sensitivity to the precision of the input data number of targets attenuation of the TSCM risk of threat discovery etc Although a HOW THE RESULTS ARE USED CSE is now using the prototype version to evaluate technical surveillance countermeasures Actual data from an OCONUS site was input into ACE prototype The results were verified by CSE and the R55 team A brief of the results will be presented in September Actual data from OCONUS sites will continue to be collected and input into ACE I 3rd Issue 1992 CRYPI'OLOG page 8 EO 1 4 c P L 86-36 S SGHB tp L 86- 3 6 The Dying Programmer's Lament The programmers spoke in a fading voice 'That diamond shows it's a multi-choice And a loop is seen where the line returns And a block is cleared but my fever bums 0 pin me not to a completion date When the machine is down and the assembly late O think of the errors I might have made And the debug sessions so long delayed 'The symbolic deck with the cards transposed Subroutines opened that were not closed The card Operations dropped on the floor The Sponsor's shadow beyond the door Yet I fought the fight It will surely run At the next debug or the next but one So we buried him on his completion date When the machine was down and the assembly late And we Sighed for the errors he might have made And the debug sessions so long delayed L --_ reprinted from bits bytes C4 MACHINE INFORMATION PROCESSING BUUETIN Vol I No 2 June 1965 3rd Issue 1992 CRYPI'OLOG page 9 POR OPPICtA L 6fm eNLY P L 86-36 DOCID 4036135 oEO 1 4 c EO 1 4 d P L 86-36 3rd Issue 1992 CRYPrOWG page 10 BEeBE' SPaRE DOCID 4036135 SBOMr FSPcm E EO 1 4 c EO 1 4 d P L 86-36 3rd Issue 1992 CRYPTOLOG page 11 SECRE'f SP6ft DOCID 4036135 3rd Issue 1992 CRYPTOLOG page 12 SECRE EO 1 4 c EO 1 4 d P L 86-36 SPOIffi DOCID 4036135 3rd Issue 1992 CRYPTOLOG page 13 EO 1 4 c EO 1 4 d P L 86-36 ii T SP81B3 DOCID 4036135 eRE't' SPOIfE 3rd Issue 1992 CRYPTOLOG page 14 EO 1 4 c EO 1 4 d P L 86-36 i R T PQIQ DOCID 4036135 3rd Issue 1992 CRYPTOLOG page 15 EO 1 4 c EO 1 4 d P L 86-36 iCIti T PQD DocrD 4036135 Writing in ajournalistic styfe of greater interest Each reader and viewer has individual preferences So in essence every SIGINT customer constantly tunes in and tunes out topics of the day Additionally each customer also has individual preferences for receiving information Some prefer to read while others prefer to hear or see information presented to them That's why newspapers magazines radio and television_have all flourished FOUO Who determines story topics what will be published for print and produced for television and what priority each story will take Well in the media a team led by a managing editor or executive producer works together to answer these tough questions The team oversees day-to-day production while at the same time planning for 54 the next issue or program Quite often media senior managers get involved in these collabora U During the past year we have witnessed a tive editorial group discussions to provide addimyriad ofloca1 national and international events tional input into pre-publication planning special The Birds Move Into Camden Yards Hurricane event coverage and post-publication reviews The Andrew Sweeps Devastation Across Southern objective is to serve the customer's thirst for inforFlorida and American Troops Victorious in mation in an accurate timely and highly competiPersian Gulf' were but a few of the headlines that tive fashion The journalistic process of identifycaptured the attention of diverse audiences What ing stories prioritizing topics and overseeing do each of these stories have in common Well no production never ends matter where you read heard or saw these stories the information presented to you was POUO The journalistic topic experts are the planned published and distributed to you in a analysts reporters correspondents writers and journalistic style and format producers who have intimate fi1' t hand knowledge tro CD The journalistic approach to reporting tells a story It seeks to provide the reader or viewer with a direct presentation of the facts with minimum interpretation You remember the kind of reporting delivered by Walter Cronkite when he signed off at the end of each Evening News show with And that's the way it is of developing and breaking stories The journalistic process of reporting empowers these individuals to produce and meet deadlines The journalistic process also supports follow-up reporting and timely correction updates CU The journalistic process of reporting is much like a total quality management TQM process Journalism products are geared toward satisfying U The headline gives the reader or viewer a diverse customer demands for information the quick synopsis of the main element in a story It attracts the reader or viewer and serves as kind of editorial boards are collaborative process action teams and the staffs and support mechanisms of a marketing hook to get people to buy newspapers the journalistic process are empowered to put out watch television or read NSA SIGINT reports the best quality products possible while striving Never thought we were in the marketing business for the competitive edge The journalistic process with our SIGINT products did you But indeed of reporting and TQM go hand in hand in satisfywe are ing both the customers' demand for excellence and the desire ofjournalists to contribute and be an POUO The headline also allows each reader or viewer to pass over a story and move onto another integral and essential part of the process 3rd Issue 1992 CRYPTOLOG page 16 FOR OFFIOlMJ USE ONLY ---- ----- --- DOCID 4036135 Reducing the Burden of Maintaining Software program will possess most of the seven quality factors Le reliability understandability testability modifiability usability efficiency and portability but it is seldom possible to cost-justify all seven for anyone program Another block to all factors receiving equal importance is the fact that some of these factors are in conflict one must be sacrificed to improve another A common example is the efficiency understandability conflict Again the project manager must rank the seven quality factors and emphasize to the programmer which factors are to be delivered for the specific system in question QUALITY-ENHANCING TOOLS The use of quality-enhancing techniques and tools will improve the maintainability of a system Seven suggestions will be proposed to help minimize the burden of software maintenance If these The quality-enhancing techniques and tools that will be discussed here are structured techniques methods are implemented the resulting system restructuring reformatting and prototyping will be less of a bU den on the data processing department that supports it One method used to Structured techniques improve the maintaInability of software is for the project manager to set explicit maintenance objec- Structured techniques should be utilized in all tives and priorities Another is to use qualityphases of a system structured techniques imenhancing techniques and tools that will improve prove understandability i e quality and theremaintainability and will improve the system's fore reduce the overall software costs Structured documentation Establishing activities that techniques standardize the style of the software assure quality choosing a maintainable program- system this standardization helps programmers ming language and establishing file systems that become familiar with the system mpre quickly are independent from the programs are three their understanding of the system is more comdesign concepts that will greatly improve mainplete the quality of the system is upheld Structainability Finally I will give suggestions on how tured code is the introduction of standardization to contract for a maintainable system when that into the program's form Modularization is the system is supplied by a vendor traditional approach for enhancing quality the SETTING MAINTENANCE OBJECTIVES The best way to build anything into a system is to ask for it therefore setting explicit maintenance objectives and priorities will improve the maintainability of a software system By setting maintenance objectives and priorities we make it possible to influence quality and therefore maintenance According to the Weinberg studies each project team will achieve the highest goal set for it whatever system qualities are stressed those are the qualities that will be delivered The manager of the project team must tell the maintenance programmers what quality factors their system is expected to contain A maintainable theory here is that independent pieces will simplify the program's understandability and the maintenance task When modularization is taken another step further structured programming results Structured programming is a modularized system that represents a logical and hierarchical relationship Coupling is low the execution flow among modules is simple and easy to understand Restructuring The objective of restructuring is to improve the understandability of the existing software system and therefore improve its useful life Complex error-prone and frequently changed modules are 3rd Issue 1992 CRYPTOLOG page 17 FQR QFRC b -I s l JSE ONUI DOCID 4036135 prime candidates for restructuring A caution we must be careful not to take poor unstructured code and develop poor structured code The goal of any structured technique is to improve the quality understandability and maintainability of the system Reformatting documentation promotes system usability When user documentation is poor misinformed people report errors These hypothetical errors are really differences in the interpretation of the system's functions Most managers agree that proper documentation is a good idea but they seldom require it Program documentation is used to help the maintenance programmer understand the internal structure of the program It is also used to demonout the introduction of restructuring Reformatstrate the software's coupling how the modules ting the introduction of indented code standard interact within the system the systems interaclabel conventions one instruction per line and tions with the operating system and within other standardization of keywords are much less risky software systems Program documentation inthan restructuring and have also been shown to cludes external program specifications program improve understandibility flowcharts source code commentary and system flowcharts The most useful documentation is Prototyping high-level This documentation explains the Prototyping is a quality-enhancing tool that can be overall purpose of the program and describes the just as useful in maintenance as it is in the devel- relationships among the various program compoopment cycle It provides usability since it allows nents External separate from the source code the maintainer to understand the needs and redocumentation is necessary HIPO diagrams and Warnier diagrams are two examples of external quirements of the end user Prototyping is most useful during life cycle support since it minimizes documentation Low-level documentation Oine-bypossible user misinterpretations It is a valuable line descriptions is not necessary The best way though underutilized quality-enhancing tool to provide low-level documentation is through the use of self-documenting programming languages IMPROVING DOCUMENTATION Program documentation is produced in the design phase problems occur since it is''rarely updated to Documentation along with quality is one of the reflect maintenance changes factors that contribute to the difficulty ofmaintenance If a system's documentation is improved Data documentation is needed in addition to so will its maintainability Documentation can be program documentation There are two ways to classified into four types user operations prodocument data data modeling and a data dictiogram and data documentation Since each pernary forms a specific function maintainability is improved if all forms are present This is primarily o The data modeling provides a graphic model because the maintenance programming team will identifying the structure of the data and its funcbe able to find exactly the information it needs tional dependence without extensive searching o A data dictionary lists all the forms of data used User documentation provides instruction on the their definition how they are used where used use of the system's programs Instructions are and who is reponsible for them provided for the entering of data interpreting of Data documentation needs to be included but output information and reacting to error mesoften is not sages Usually this consists of a user manual but a more usable approach is an online documentation system This online transaction would be available on the end users CRT High quality user If a manager considers restructuring too risky there is still a technique that can be utilized with- 3rd Issue 1992 CRYPI'OLOG page 18 VOK OfRCkY YSK OM JPY' DOCID 4036135 QUALITY ASSURANCE Another method for improving the maintainability of software is the establishing of explicit quality assurance activities Commonly called quality assurance audits these activities are important in maintenance as well as in development phases of a system Briefly there are four types of quality assurance audits covered the more easily recoverable it is and the less costly the error is to correct We have already discussed seven factors that contribute to quality i e reliability understandability testability modifiability usability efficiency and portability and by measuring these we can measure the cumulative quality of the system This should convince managers that they should practice quality audits CHOOSING A LANGUAGE o Checkpoint reviews are used in the development of new software They are used between the phases of development to check the development work as it progresses again the sooner the error is found the less expensive it will be to correct o An treceptance review is a special checkpoint review that occurs between the development and production stages An acceptance review sometimes called project turnover is the last chance to ensure maintainability before the software becomes operational and becomes the responsibility of the maintenance staff Choosing the proper language can affect the program's maintainability Low-Ievellanguages are difficult to learn and understand as are programs coded in a low-level language 'C' language is easier to understand than Assembler because Assembler is not structured well and does not support meaningful variable names Recognizing this the project manager should choose the highest-level language possible Fourth generation languages should be utilized when possible Fourth generation languages are easy to use understand and modify Therefore development o PerWdic maintenance audits are used on and maintenance in a fourth generation environoperational software to recognize changes in qual- ment is faster Since most fourth generation ity Since software systems are not static periodic languages are non-procedural defining what is audits are necessary and any changes in quality to be accomplished not how these systems can be should be investigated Because of the importance modified by the end users and ma not require the of overall system understanding in the maintehelp of an analyst Even if analysts are required nance function it is helpful to have the they can obtain the results faster by the use of a maintainers involved in the development of the fourth generation language than if they were to system Ideally they should be involved not only write out program specifications in the maintenance acceptance reviews but also at other checkpoint reviews It is generally thought that the use of fourth gen- o The benchmark audit is used on packaged software and it will be discussed in the section on improving maintainability in packaged software Audits are the most powerful techniques for introducing and preserving software quality Managers often feel that quality assurance activities are not necessary especially in a maintenance activity They usually state that audits cannot be costjustified and quality cannot be measured Although audits seem like a tiring time consuming activity they actually reduce the time allocated to maintenance the earlier an error is dis- eration languages will help the maintainability of a system The quality of the system is improved by the use of fourth generation languages Understandability is better since the language is simplier and less complex The code must be structured since unstructured code is not obtainable with the use of a fourth generation language These languages often are equipped with selfdocumentation capabilities thus reducing the maintenance difficulty FILE STRUCTURE INDEPENDENCE I think we all can agree that a file cabinet is a flexible way to store data data can be added 3rd Issue 1992 CRYPI'OLOG page 19 FeR ePFICtAL USE eNUI DOCID 4036135 removed or rearranged without major difficulty This is not always true of a computer file system computer data files often have flexibility problems They often have high levels of redundancy and inconsistency can occur when data are in different stages of update Since computer files are inflexible requests for a new data arrangement can take weeks or months users to ensure that the software package performs according to their expectations If the vendor refuses to allow a benchmark audit as part of the selection process another vendor should be considered The conduct of these benchmark audits are an important part ofthe acquisition of an application package Once a package has been selected an appropriate contract must be written The user must have a Seemingly trivial changes to a file system can set good contract to ensure that the vendor will keep off a chain reaction or two The goal is to avoid the software package in good working order A changing programs when a file's physical struccontract contains some important specifics A ture changes The data base environment was introduced to solve this problem The intent of the reliability clause is needed to guarantee adequate data base environment is to isolate the program maintenance support it should specify how quickly the vendor can be expected to respond to a from changes in the structure of the data files This environment allows the program's perspecrequest for service how quickly the software error tive to be different than the physical record The will be repaired after it is reported the methods programmer perceives a make-believe record for correcting software errors and the penalties and therefore does not worry about changes to the incurred by the vendor if the reliability guarantee data's physical structure the programmer can is not met The software lease agreement should represent all data structures and dynamically assure the customer knowledge of and access to new releases of the software package The vendor create new access paths One of the factors discussed earlier that contributes to the difficulty of should provide a renewal option in the contract maintenance is the dependence of the programs on This clause allows the purchaser continuing mainthe file structure the data base environment tenance even though the supplier has a short provides program and file structure independence contract term The final clause is the termination clause This provides the purchaser with the PACKAGEDSOF'lWARE source code in the event ofvendo bankruptcy Package software applications should be purCONCLUSION chased with maintenance in mind Again the principle applies you get nothing unless you ask Nice suggestions you may be saying but I have for it We should plan for the future maintenance heard all of them before It just is not that easy when a software application package is purchased to change Although these statements are true I Each vendor's reputation should be examined will feel that with a little effort the time and money the vendor provide adequate service The custhat is put into maintenance each year will be tomer is dependent on the vendor to perform reduced Although all seven of these methods will maintenance on the software the software must help reduce the burden of maintaining software be kept in good working order This is especially the system should be considered when determintrue when the customer does not have access to ing the appropriate method Generally the first source code and or is not allowed to modify the three methods will make the biggest impact on software maintenance Setting explicit maintenance objectives and priorities using quality-enhancing Current users of the packages under considertechniques and tools and improving program ation should be sought and potential users should documentation will improve the quality and undiscuss the package's performance It is best to derstandability of the computer system Using conduct a benchmark audit to ensure the quality these suggestions alone can reduce the burden of of the protential software package A benchmark maintenance significantly Sometimes underaudit is a program test conducted by the proposed standing the problem is half the solution J 3rd Issue 1992 CRYPTOLOG page 20 FOR QR lGIltIs BSE 6NIIl DOCID 4036135 8F lOIlB'I' Chinese-English translation aids Its purpose is to promote cooperative efforts among linguists lexicographers computer specialists and others in compiling and updating computer-stored machine-readable Chinese-English dictionaries and glossaries and to make available the products of those efforts to its members and insofar as feasible to other users ofChinese-English translation aid_so CETA D Membership of the CETA group consists of agencies and individuals interested in ChineseEnglish translation There are no formal requirements for membership in the CETA groups aside from an interest in the furtherance of its objectives HISTORY P L 86 36 A631 8-000 Sometime this year a data tape will be delivered to B6 containing a data set of some 250 000 Chinese-English dictionary entries Development of a local software package will make those entries retrievable by English term or by Chinese term in characters -both long and short forms-or by Standard Telegraphic Code from networked workstations in B Group The delivery of that tape will mark the high point in the Agency's long and at times difficult participation in an organization called CETA For a period of six years I was the Agency acting for DoD representative to the CETA group Following is a brief history ofits 20-year existence COMPOSITION D The Chinese-English Translation Assistance CETA Group is an organization of persons from the U S Government and the private sector who share a common interest in the development of U In mid-1964 the U S Government sent academia a list of government needs of gaps in China research The government in this case was the China Committee of the interagency Foreign Area Research Coordination Group FAR The contact point in the academic word was the Joint Committee on Contemporary China JCCC established in 1959 under the auspices of the Social Science Research Council and the American Council of Learned Societies and funded by the Ford Foundation The late John Lindbeck of Harvard then chairman of the JCCC in his reply to the government initiative made it very plain that the scholars he represented covering most of the private China studies programs in the United States at that time had sufficient vague indications of general interest in government-private cooperation to support such a project He rejected most of the topics on the list however such as Minority Groups in China and The Relationship between China's Foreign and Domestic Policies as non-starters in developing truly meaningful government-private cooperation U Lindbeck focused instead on a lowly project near the end of the list Development of a Comprehensive Dictionary of Modern Chinese Terms Here he said was an area where government and academia simply had to work closely together if the need were to be met and time was running out 3rd Issue 1992 CRYPTOLOG page 21 iiJU 'I' IWHlbB VIA eOMHff OHMiifBLS ONif DOCID 4036135 U Beginning in 1964 the former Foreign Documents Division FDD now under the Foreign Broadcast Information Service FBIS assembled materials for use in compiling a general-purpose listing of contemporary Chinese terms with English translations One of the source materials was a Chinese-Japanese dictionary containing many Communist Chinese terms others came from China or from U S Government institutions including the Joint Publications Research Service and the Foreign Service Institute The original FDD plan was to merge six dictionaries and glossaries with computer assistance and to have the resultant compendium published ultimately in a thoroughly-researched commercial dictionary form encouraged to attack not only definition of that problem but also its solution o Further development ofCETA's living dictionary with appropriate purpose and quantity should retain first priority among the Group's efforts U The first run ofthe CETA dictionary was published-and distributed in September 1971 but an oversight occurred that caused long and recurring argument within the group over the dictionary The first run was quite crude and was intended primarily for contributors to look at and evaluate with a view toward acquiring considerable guidance and input from them But a caveat to that effect was omitted and as a consequence many people looked upon it as merely a crude U A CETA Workshop in March 1972 was decompilation of reversed English-Chinese dictionarsigned for just that purpose It produced the ies of no great value they overlooked the fact it guidance and momentum that led to joint funding ofCETA by 9 of its 12 member agencies beginning was a printout from an online database that could be easily corrected supplemented and edited as with FY 1973 In two days of panel and open discussion in the informal workshop atmosphere opposed to a typeset book with its attendant difficulties the mixed group of 290 government representatives and 24 academic participants from 20 pri s-eeo Agency representatives however were vate institutions worked smoothly together most interested in this database especially in anticipation of the SEMESTER system Eventu U After a well-pitched keynote address by E ally after editing a large subset of the database Raymond Platig Director of the State was loaded into the SEMESTER system and Department's Office of External Research which became the central core of compu rizedlookup had funded the workshop including travel from capability for B Group Chinese transcribers By all over the United States and from three foreign 1986 many analysts were demanding more countries the CETA dictionary effort was explained along with the varied projects of 16 other readily accessible and manipulable computerresident dictionary database so greater effort was institutions that bore on the Chinese materials processing problem under discussion At the end expended to refine the CETA database and to make more readily and conveniently available to of the discussions which many described with Agency analysts some emotion as the most fruitful they had ever known in a conference situation four main conclusions were reached o Steady increase in the flow of materials from China on research on China was likely o Efficient processing of those materials in Chinese and English for both government and private use in research was a definable problem ofconsiderable importance CCQ Now with the acquisition of the Chinese character set on the SUN system and the opportunity to buy large relatively inexpensive storage capability all of the elements finally merged to bring the best of CETA onto the desks of Agency analysts o The CETA man-machine system should be 3rd Issue 1992 CRYPTOLOG page 22 8B0RE'I' ----- -- ----- -------- DOCID 4036135 In commemmoranon ofthe 50th anniversary ofthe World War II 1991-1995 The wartime contingent oflinguists codebreakers and cryptanalysts was an exotic melange ofmultitalented people many ofwhom had already made their mark in the world Some were foreign-born some had already had prestigious careers in academia There were missionaries and biblical scholars mathematicians classicists andlinguists there was a Russian Polish noblewoman who had been rescued by an American gunboat a Hungarian prima ballerina who had been awarded a medal by the Pope a professor ofphilosophy who was also a navy officer cryptanalyst and Chinese linguist and there were others who later were to achieve fame Eugene McCarthy McGeorge Bundy William Bundy Edwin M Reischauer also columnists Joseph Kraft and Charles Barnett journalist Al Friendly and the elder statesman of bridge Oswald Jacoby Among them were musicians counted by the dozen Scratch a cryppie find a musician or so it seemed Peter Nickels a conservatory-trained concert violinist doubled as the conductor of a dance band in the late 30'sl onducted concert versions of Gilbert and Sullivan at Arlington Hallwith a cast of equally gifted musicia1t 1 I who in 1962 received theJOth Annual V -_ I International Jazz Critics' PoWs Best Unknown Andthen there wasBandmasterL B Red Luchenbach USN Band #16 assigned to the CALIFORNIA On December 7 his ship was hit and sunk Meanwhile his bandsmen vanished The fleet personnel officer refused to tell him where they were but Red persisted and eventually joined them at Combat Intelligence Though it was to be a temporary assignment the musicians proved too valuable to be replaced and so they served as machine processing specialists and cryptanalysts By the end ot the war Red had a commission Later in civilian life he represented IBM at the Navy Security Station The star was the late Lambros D Callimahos dubbed the Paganini ofthe flute b Y music critics a flute virtuoso on the international scene in the 30's At the prestigious Mozarteum in Salzburg Austria he was the youngest professor ever Also in the early 30's he developed an abiding interest in the history of cryptology LDC was a collaborator with WilliamFriedmanonMILCRYFT I and II wrote MILCRYPT III and N and other papers and monographs He developed and taught the seminal course in cryptanalysis CA - 4 0 0 that GradusadParnassumencompassingthecryptography of the then known cipher systems I Trumpet Player in Jazz 10 jazz piano This lecture was delivered at an unknown time and player-cum editor an tlso very many others place and to an unknown audience This lesturewas entitled on your programs The History of Cryptology It's had several other titles one of which was 26 or Bust 26 x 25 x 24 x 1 To those of you who use the slide rule to 1 digit of accuracy it is four times P L 86-36 10 to the 26th Cryptology is an ancient profession in fact the second oldest profession one that abounds in drama and fascination and one that has had a 3rd Issue 1992 CRYPI'OLOG page 23 P6It 6PPICtkL ust 6NLY DOCID 4036135 profound impact on the tum of events in history priest Greek priests a married man may become a priest but priests cannot marry We start offwith communication The Greeks did have messengers In sending communications from one commander to another they had the usual runners but sometimes they wanted to be on the safe side and conceal the messages For one idea the Greeks would shave the head of the slave and inscribe the message on the bare skin of the head And then you wait awhil eferredmessage--and the slave was dispatched not executed but dispatched to the distant commander who would shave the head of the slave and read the message If the message were particularly sensitive it would be a onetime slave These slaves are also as you well know normally distributed The Greeks also used secret inks the juices of various berries milk etc brought out by heating They used sputum and other eflluvia On concealment Hieronymous mentions that messages would be concealed in the belly of a hare or inscribed on a wooden tablet and then covered with wax or even inscribed on the leaves covering the putrid ulcers of disguised beggars Also on concealment Sir Francis Bacon in 1623 wrote his renowned work The Advancement ofLearning wherein he showed how he could disguise some of his innermost thoughts thoughts which might be considered heretic in those days by means of concealment within a covering text It is the same system used today in our modem teleprinters On the subject of concealment there is also Boccacio those of you who may have read the magnificent work in the original Italian you know that Boccacio gave methods of information retrieval that is by means of the positioning of curtains or shades but that's neither here nor there Cryptography was practiced among other things by the ancient Egyptians only God knows why because their hieroglyphic writing was enough of a disguise as it was except for the learned class the priests And on the subject of priests let us not forget that it is the ruffians and priests and scoundrels who have made great advances in cryptography and cryptanalysis throughout the ages I myself am the son of a EKYTAAH 400 B C Lysander used a device known as the scytale in the dictionary it is pronounced sitale s-c-y-t-a-I-e but I don't like sitale it is a E K Y T A A H The scytale consisted of a baton that the marshall carried about yea long tapered with a notch at one end and the marshall would take off his belt and wind the belt around the scytale and then write the message in Greek of course across the bars then he would take off the belt and dispatch that to the distant commander who having a scytale of the same size would affix the one end of the belt on the notch wrap it around and 10 and behold the Greek plain text came out So that was our first transposition system Also in 400 B C we have Aeneas the Tactician who wrote a voluminous tome on the defense of fortifications one chapter of which was devoted to the subject of cryptography This was the first treatise on cryptography o 100 B C Julius Caesar in corresponding with Cornelius Baudus Opius-whoever the devil they were and others--used a simple scheme of replacing each letter in the Latin alphabet by one three removed from it in other words A plain would become D cipher B plain would be E cipher etc But this was too much for Augustus' brain he preferred a simpler scheme where A plain was replaced by B B by C etc Now we jump to 1200 A D and the Papal States By this time there were active cryptographic bureaus in the clergy and the Papal States were the first ones to engage in systematic crypto correspondence They often substituted vowel marks for the vowels leaving the consonants unchanged In other words Mississippi would be spelt M SS SS PP where one dot might be an I two dots an E etc In 3rd Issue 1992 CRYPTOLOG page 24 FOR OFFIOfAiJ use ONLY DOCl D 4 O 3 6 1 3 5 1378 Gabriel de Lavinde of Panna who worked for Clement VII-and if there are any historians here they know Clement the seventh did not flourish at that time it was the anti-pope Clement the Seventh Lavinde wrote an SOl a signal operating instruction which is on deposit in the Vatican right now and he gave many alphabets for which he had multiple equivalents for the single Latin letters Now the meaning of26 There are 26 ways of scrambling the letters in the sequence ABCDEFGHIJKLMNOPQRSTUVWXYZ You all know the number If you set up here a set of letters for the plain component for the cipher under A plain you may put anything you wantsuppose you put an X You had 26 choices here Having an X here there are only 25 choices here-make that an O To gain some idea of the great size of this number if you had 1 000 machines capable of testing 1 000 000 different alphabets per second it would still take you over one billion years to go through the gamut of all alphabets However since you have a 5 probability of hitting before you reach half-way through you can say roughly you have the expectation after 500 million years Leon Batlista Alberti's cipher disk In 1470 Leon Battista Alberti wrote his Trattati In Citra He was an architect painter musician writer on art and the most universal genius of the First Renaissance He invented among other things the cipher disc In 1404 on Monday July 4th-imagine what prescience the man had-Cicco Simonetta an Italian but nevertheless born on the 4th of July wrote a little tract on cryptanalysis the oldest tract extant and he observed what you can do with Latin secret writing by capitalizing on frequencies on patterns of words and on vowel identification His methods were so good that they hold even today He didn't get very far in 1480 he lost his head That was Monday July the 4th 1404 In 1531 David Trithemius the Gennan Abbot wrote volumes I and II of a projected 4-volume work He never finished the work but it was the first extensive treatise on cryptology He was also a magician-after all in my business every little bit helps-and he was accused of being in league with the devil and his books were burnt but fortunately he wasn't OM of Giovanni Btltzula Por 's dpMr dUks You don't though in solving a simple substitution cipher have to make all these trials We do work on frequencies the fact that the letters composing language are not equiprobable We do work on repetitions not only of letters but also of digraphs trigraphs et cetera and of long sequences that you hope are complete words And we do work on patterns like the word CEMETERY has an A B A pattern 3rd Issue 1992 CRYPTOLOG page 25 POR OPPletAi J USE ONIN DocrD 4036135 Three Victorian amateur cryptologists Sir Charles Wheatstone inventor of two important cipher systems Lyon Playfair First Baron Playfair who gave his name to one ofWheatstone's ciphers and Charles Babbage who solved many difficult ciphers o for the 'e me' The word BATTALION has a A B B A pattern for the repeated letters 'a t t a' et cetera So quite early in the game when people realized the weakness of simple substitution they thought they'd get aroWld it by having variants So where E is 13 per cent in English instead of having one cipher symbol that would stand out 13 per cent you'd now have 5 symbols that stand out about 2 or 3 per cent each nary rectangle is enciphered by the other diagonal of the imaginary rectangle and PC is NE This was a brilliant idea because it suppressed the frequencies of single letters However Mauborgne did show them one method of solution And he made good anyway because he rose to be Chief Signal Officer All these remarkable ideas in cryptography were offset by even more ingenious ideas in cryptanalysis We'll come back to codes in a minute The departure was in the latter part of the nineteenth century-I'll skip over that for a minute and go on to what happened In 1914 a First Lieutenant Mauborgne published a paper put out by the Army Service School's Press Ft Leavenworth Kansas and the title An Advanced Problem in Cryptography and Solution had to do with the Playfair cipher It was invented by Sir Charles Wheatstone who did invent the Playfair cipher but it was Lord Lyon Playfair who sponsored it in the Foreign Office and gave it its name But that's all right because Wheatstone didn't invent the Wheatstone bridge Wheatstone applied Cristi's dimension to the measurement of the bridges so you see it balances out in the end On top of that for those of you who might be interested Wheatstone also invented the concertina Instead of encrypting one letter at a time in the Wheatstone Playfair you encrypt two letters at a time So EN is CP this diagonal of the imagi- Leon Battista Alberti 1404-72 considered the father of western cryptology Alberti was also a talented musician writer artist and athletea universal man In 1902 a chap by the name ofF de la StelleF I suppose was Fran ois but it could be Felix nobody knows-wrote in 1902 a book called Traitt de Cryptographie He mentioned in this book a system by which some letters might be enciphered by single digits and other letters by pairs of digits This was an academic curiosity Wltil the '30's when this system took on major proportions not as is but with certain kinds of disguise I mentioned briefly some ideas of simple substitution and also the idea of transposition as exemplified by the scyta1e 3rd Issue 1992 CRYPl'OLOG page 26 FOR OFFIOfAiJ tJSe ONLY DOCl D 4 0 o o o 3 6 kl 3 5 That was in 1586 In 1765 there flourished in the true sense of the The earliest klWwn digraphic system Giovanni Battista de la Porta replaced each pair of letters with the sign at the intersection oftheir row and columns In the 16th and 17th centuries transpositions came to the fore In a transposition system you retain the same language elements of the original message except that you permute them about Their identities remain the same but their positions have been changed The substitution systems of course the positions remain the same but the identities change The first idea of polyalphabetic substitution was given really in 1470 by our friend Alberti in his cipher disc In 1563 we have Giovanni Batista de la Porta He was a physicist inventor of the camera obscura which was a predecessor of the Kodak and he was a healer of the sick just as we are healers of sick messages He is known as the father of modem cryptography because it was he who pushed the idea ofpolyalphabeticity In 1586 a French gentleman by the name of Blaise de Vigenere was travelling and one of the things he picked up while in Italy was the idea of a square table to which he gave the name 'the Vigenere tableau' He didn't invent it moreover he never said he did but he gave his name to this idea with which you could perform true polyalphabetic substitution with no limitations word a truly great man Giovanni J acopo de Seingalt otherwise known as Casanova he was a remarkable scholar a savant a person who well I can't say we should all emulate but at least study because I'm sure we could all learn from him One of his remarkable exploits was the fact that-thank you somebody who was asleep is now awake--was his solution of a polyalphabetic cipher 100 years before the method of attack was announced to the world by the German Major Kasiski Of course Casanova was a privileged person he managed to get cribs in the most remarkable places But the way this came about I'll give you very briefly the background of all this-oh my goshthe background will have to rest-if any of you wish to see me privately I'll give you the full story about what happened to the Marquis when he solved her cryptogram 1863 was the date ofKasiski Now we come to codes The ancients went from simple substitution to variant systems to dissimilar writing wherein certain groups of characters or for that matter certain plaintext words took on a new meaning like 'ALMONDS' means 'I won't be home until Friday ' So codes came out quite early in the game but it wasn't until 1640 the great French cryptanalyst Orsignon that the two-part code came into being In a one-part code the code groups are arranged alphabetically and the vocabulary elements are in alphabetic order also In a twopart code one section the encoding section has the vocabulary elements in alphabetic order with a scramble of the code elements and then another part with the code elements in alphabetic order for ease in decoding The idea of an enciphered code soon followed because of certain weaknesses of unenciphered code systems We might go back for a moment -to the earliest cryptanalysis 550 B C -Please note the dateDaniel read a cryptogram for Bathsheba It was a good stunt there were symbols on the wall He not only pronounced the symbols but also gave the meaning And since there was nobody 3rd Issue 1992 CRYPl'OLOG page 27 EOa OQICIAI -gs g 0NI8 DocrD 4036135 there to contradict him why he made hay while the sun shined In 1510 Yulan deSoto of Venice solved many ciphers including those of Charles the Fifth which had been intercepted by the Papal Court and not solved by it So he was an outsider like an NSA consultant that France was using sorcery so there was a miniature Pearl Harbor investigation on Viete who to avoid conviction of sorcery a capital offense told all In 1525 the British lion begins four centuries of successful cryptanalysis In 1556 Alberti solved a message for the Portuguese ambassador who had lost his own code This wouldn't happen today I mean if a diplomat loses his own code he doesn't come to NSA asking for help 1567 A prior in St Peters according to Vigenere deciphered in less than six hours a large page of cipher in the Turkish language of which he did not even know four words Having travelled I can imagine what the four words were 1589 Fran is Viete also know by his Latin name ofVieta who as Privy Counsellor in Antoine Rossignol the father ofFrench cryptology That's 1589 1595 in June Viete by this time was a good blabbermouth in conversation with the Vatican ambassador to France revealed that his ciphers were being read That's the worst of all the sins that one could commit John Wallis clergyman and mathematician England's first great cryptanalyst France solved a 50D-group code of Spain's Philip the Second Philip bitched to the Pope 1626 Rossignol who was remarkable for keeping his trap shut began a cryptanalytic career with Louis the Fifteenth When Louis was dying he told the queen that Rossignol was one of the men most essential to the State He was fifty-six years a civil servant Brigadier Tiltman is second only to Rossignol Wallis in 1645 the great English mathematician began a career of five decades as an active cryptanalyst The Brigadier is in the middle of his fifth decade He was under Cromwell and he solved the secret cipher of Charles the First In 1689 still going strong he solved the cipher of Louis the fourteenth a 6OO-group two- and three-digit code In 1821 Jean Francois Champollion a miserable little fellow with a pale complexion yellow 3rd Issue 1992 CRYPl'OLOG page 28 FeR eFPICtAi tfSE 6NLT DOCID 4036135 skin slaty eyes was very gifted as a child He told his brother who later became his keepernot in that sense I mean his brother sacrificed his life for the more adept junior-little Champollion became imbued with the idea of reading the Egyptian hieroglyphs He fell to work reading everything he could get his hands on Modem Egyptian et cetera-he was a gifted linguist-and in 1821 Champollion succeeded in breaking the secret of the hieroglyphs The hieroglyphs were broken bymeans ofthe Rosetta Stone a piece of black basalt three feet high two feet across with three inscriptions evidently parallelinscriptions in Greek Egyptian demotic and hieroglyphs He had what we call isologs and he was very very fortunate because to recover an unknown language takes some cryptanalytic 'in' some cribs and this was the height where he had parallel texts in three versions Incidentally although Sir Charles' idea was in 1867 in 1807 an American got there first Decius Wadsworth who was later chiefofOrdnance U S Army He invented the same principal as the Wheatstone cipher device and executed it even better mechanically This device rested in limbo and the British knew about it and in World War I the British wanted to use the idea Now the solution was known the solution where the plain component was the normal sequence and the cipher component an unknown mixed sequence But there was no known solution to the Wheatstone Cryptograph with two unknown sequences But we'll come back to that story after a few minutes Let me do some more on cipher devices In 1891 a French major on the General Staff a reservist by the name of Etienne Bazeries published in 1901 an essay showing the idea of a spindle with 20 disks permutable on the spindle The Rosetta Stone each of the disks had an alphabet a mixed sequence inscribed on the periphery you arrange the disks on the spindle according to the key set We come now to cipher devices again to our friend up the disks along a guide bar for your first twenty Alberti in 1470 the first one to dream up a cipher letters of plain text and send the cipher text disk Then by 1500 the idea of a cipher disk everyotherrow At the other end the deciphering occurred to many many people a disk let's say of clerk would have the same disks arranged in the 26 letters revolvinginside a frame ofanother circle properorder he would setup the ciphertext on the of 26 letters The idea was invented dozens of disks and look around One and only one row times In 1867 Sir Charles Wheatstone inventor of the Playfair thought up an ingenious little mechanical contrivance to which he modestly gave the name 'The Wheatstone Cryptograph ' It for its day was the most sophisticated idea in cryptography but it lay buried in the archives Bazeries Spindle 3rd Issue 1992 CRYPI'OLOG page 29 FeR eFFIOIM a WE Q DOCID 4036135 would have plain text all the way across He had a picture of this device in his book and across the guide barwas the sentence Je suisindechifferable I am indecipherable Let's see the book was published in 1901 it was 1891 when Bazeries thought of it but it was in 1893 when a buddy of his the Marquis d'Viares another one who didn't have to work at cryptology who was very skilled-I consider this piece of cryptanalysis the greatest piece of cryptanalysis considering the age in which it was done D'Viares showed Bazeries how he could arrive at a general solution of the device having only possession of the disks This you would assume in time of war that the device would be captured It was Bazeries who published it first then in 1915 an American army captain by the name of Parker Hitt invented the device again He was the third inventor because I didn't mention that the first inventor was Thomas Jefferson but his papers weren't discovered until 1926 and he showed in his papers the idea of 36 disks on such a spindle and that was a most remarkable cryptographic idea of its day Now we come to World War I Radio To coin a phrase radio is a two-edged sword and you can go on from there Every lecture you hear at NSA has that phrase in it so it's wise to remember it On the 26th ofAugust as you all remember in 1914 the battle ofTonnebre lasted three days 100 000 men were killed or wounded and missing in action Two Russian commande amsonovcommanded the Second Army Rennenkampf commanded the first Army They didn't like each other they had no contact with each other not even on the staff level Samsonov went out in the field with the old code but in the meantime Rennenkampfreceived the new code from Moscow so he promptly destroyed his old code He sent a message to Samsonov who couldn't read it He asked for a relay in the old code He couldn't get it because Rennenkampfhad destroyed his old code So then Rennenkampf proceeded to send his messages in plain language The Germans couldn't believe their ears They read the messages where the Russians were supposed to be they sent out reconnaissance patrols found out that the Russians were there then every dayit wasn't every day for long just a few days they waited for the day's take before they made up their battle plans and in three days everything was lost So that is a fine example of how things can go wrong The next item I wish to cite in World War I was 16 January 1917 The Zimmerman telegram Perhaps the most famous cryptogram in history Zimmerman the Foreign Minister sent a message to Bergstoff here in this country for transmittal to Eckhardt in Mexico This telegram offered Mexico parts of Texas Arizona and New Mexico if Mexico would enter the war on the side of Germany Of course this is a dreadful abuse of the hospitality ofa neutral country The British solved the message and conveyed it to us with some misgiving because they didn't want to reveal what they had been doing At the beginning the anti-British faction here thought that it was another trick to get us into the war but Zimmerman was queried in Berlin and he admitted to having sent the telegram Six days later we were in the war The U S Army went to war with three dreadful systems One was the War Department Telegraph Code which was safe because it was large enough so you couldn't hide it under your tunic you'd have a bulge and it had to have a certain amount of security to go with it because of the size and the U S Army Cipher Disk with reversed standard alphabets the solution of which you'd do in the first lesson of Military Cryptanalytics Part II and you could solve a single message you could solve a portion of a single message In any case that's what we had as the mainline system and for emergency we had the Playfair but on every SOl every two or three days when they changed the key word there was a warning please don't use it because it's weak insecure 3rd Issue 1992 CRYPTOLOG page 30 POR 6PPICb i o USE ONLY DOCID 40 3 6 1 3 5 So that was on the U S Army side Riverbank Laboratories headed by a megalomaniac named George Fabyan A private concern The Germans were more fortunate they had double transposition they had complex At this joint there was a young geneticist named polyalphabetic substitution systems and very William F Friedman What happened was that ingenious combined substitution transposition this chap Fabyan had a hobby ofBaconism He systems known as the ADFGVX cipher It was wanted to proved that Bacon wrote Shakespeare so remarkable that if I sent you a message and or vice versa and he got a young lady in his then I had a power failure-not me the sta employ by the name of Miss Elizebeth Smith to tion-when I sent only half the message you read what there was in English there wasn't still could reconmuch God knows struct the entire to help him with his message from the hobby So then she half that had been had small classes in sent You can't do OK l ft cryptology and one that today of the gents was i GBP d C5 i I ' __ - oo this William I The Germans were 1 Friedman who did nc lnlO methodical so-andntor ' ene very well and so he ott so's and instead of r e later took over the having check proce1 3 1 t 1 I ''' lt4'Tl section dures to make sure ute uu S1 r that the system was Anyway Somebody 1I e working right what remembered '7'1 s noo they didwas they sent Friedman and the a U' 'e Jtiro s- aphorisms or people at Riverbank parables early in the so they decided to morning first day of send six short meschange and they had sages out to various axioms one of Riverbank LaboraZimmerman Telegram the most frequent -tories to see what parables was Morgen could be done with Stunde hat Golden Munde the earlybird catches them Now they were only about twenty or the worm So whenever there was a key change thirty letters each and that's really not a very or a system change the Allies would searchthrough good test but Friedman by 'horsing around'early in the morning for a short message find it those were the words he used-managed to put against it one ofthese parables and in an hour or two they'd have that system GDUIl UJ - -lC nUtclCI' 'T 13042 SUZZ IC9Z 1 401 2t lC'Cl lUll '3 t 1' 4 11 2t4 l a'O I 1$ r ct t024 1 ' 2 MZI 11 t 4Ul ZH 10 1' 0Xl2 thOI 2 ' l't%H 6ftl 1300$ 2l2tO 5un 0311 0221 13'911 1 eM 4473 1t137 1U1' 4 'Oe _ 1 1 24 o 1 M - r W' 2t' of 2t 4 Ul C Me $ - le4 '1 I 21 I C lro2 t' - 'i 7311 lU 131 17'i1 1 ' 12 aI$1 aot ' ool ZV -t ' o l ttt 2 U34'r 1 K U049 f iI92 '14 3t _ lW' ' I 11 w'1 1801 lC 31 I'7J42 U M t l t%O ge 92 t t' BII9 15 'l' IOM7 1 l211$ 5oa z '1es't U 11502 93 1O C7'l' t lnZ 9DO let 7'l 'II I ' Now we'll come back to the Wheatstone Cryptograph episode The British wanted to introduce the Wheatstone Cryptograph into World War I but were reluctant to because if the Germans capturedit they too would have the indecipherable cipher-remember there was no known solution ifyou had two mixed components So it wasjudged unsolvable by the British the French and the Americans both in the AF and in Washington until someone remembered the group ofpeople out in Geneva lliinois industrial laboratories called 3rd Issue 1992 CRYPI'OLOG page 31 FOR OFFICfM l JSE ONl Y DOClD 4036135 scrounge out cipher component But he didn't know what to do with it because it wasn't until 1923 that he discovered a very strong principle the reduction to monoalphabetic terms So not knowing what to do he called in Miss Smith told her to sit down make herself at ease put on lipstick-I suppose she had some on before but---anyway and fm going to give you a word you tell me the first word that comes to your mind So he said 'machine' she said 'cipher' I forgot to mention that the cipher component was based on 'machine' --a transposition mixed-alphabet based on 'machine' So he didn't know what to do but he asked Miss Smith and she gave it to him The plain component was based on 'cipher' To keep it in the family he married Miss Smith The solution that went back to the British was quite embarrassing because the method he described for getting the cipher component was so strange that even today you can't fathom it It was part astrology part cryptanalysis that's hard Cryptograffe which is the firm headed by the Swede Boris Caesar Wilhelm Hagelin More of that later It was Damm who really invented the first rotor In 1924 a German by the name of Alexander von Kryha invented a gadget which had an astounding number ofpossibilities like the number I wrote on the board before the Kryha machine and he got a buddy mathematician to explain in precise mathematical language but even if you went through all the alphabets in time blah blah blah you could never go through them However the device is solvable and even on a single message This broke the man's pocketbook and also his heart in fact he committed suicide a few years back perhaps because he couldn't push his device This mathematician buddy of Kryha's came up with the statement that the number of possibilities with this machine was 1 4 times 10 to the 64th And since the number of atoms in the universe according to Sir Arthur Eddington is only 3 times 10 to the 74th you can see it's a very favorable comparison indeed Factorial 26 is only in the order of magnitude of 10 to the 26th and here we're talking about 10 to the 64th enough to understand but then about the plain component I asked Miss Smith and she gave it to me So that's the story about the Wheatstone Cryptograph in World War I B C W Hagelin-a brilliant engineer who came forth in the early 1920's with a whole series of devices The first one was ingenious contraption-fractionating principle What it amounts to is this you press a key on the typewriter keyboard it sets into action two rods mechanical rods which are the left-hand the row components and column of a fractionating square in other words A is 1 6 then K is 2 0 for instance this 1 6 would be enciphered separately by separate schemes recombined through this square to get a single letter output In other words one letter input is fractionated into two halves they go their separate ways in a complex fashion and join together in holy matrimony at the other end Then in the early twenties a number of cipher devices came forth the early Damm device not damn DAM M Aubry Damm operating with chains and gears and what not This Damm firm was predecessor to the Achting-Bolotek This was a wonderful idea except when examined by William Friedman and his people in the early S IS days Signal Intelligence Service of the U S Army when this was found wanting It's like a young girl who has had many offers of The Hagelin machine 3rd Issue 1992 CRYPI'OLOG page 32 peR 9F'F'lOw BSB 9NLJPY noelD 4 0 a60 a 6 l la3 -a marriage and never the knight on a white c 'larger or whatever it is They broke up every cryptographic idea that had been proposed to them or that they heard l I The Hagelin machine was a very fine advance over the M-94 but still no good compared to what we should have as a major power The better idea now this was a tactical machine something you could carry about with you they even had the paratroopers landing with the 131 pins their own two pins and the 131 pins of the M-209 and all 54 lugs zeroized on the assumption that after the poor devil hit the ground if he was still capable of so doing he'd get his key list out and do this while the tracer bullets were going over his head The Enigma of-they demolished but they didn't come up with any good ideas themselves They realized the enormity of their crime They know no matter how complex something looked the solutions may not have the same order of complexity That's why in 1923 we adopted for the lower level cipher device a tactical cipher device the M-94 which is this disk device that harkens back to Jefferson days This was in use until 1942 The solution is a childishly simple matter especially since it was published in 1893 I ' Significantly we're the only nation that adopted it the French refused to have anything to do with it We had nothing better until one of Hagelin's devices came along accompanied by Hagelin and this ingenious device about seven pounds this big gave rise to a very long key the alphabets were known components reversed standard alphabets but the key was 26 x 25 x 23 x 21 x 19 x 17 So that's a very long cycle It didn't mean that for a solution you had to wait for a message that long but it did have certain fundamental weakness that nevertheless in spite ofthe advice ofFriedman and the Army we did lease the North American rights and thousands of the devices were made by the L C Smith-Corona Company during the war rm sure there must have been more than one violation of security where a man dropped with his machine already set up The first patent for a wired wheel machine was in 1918 in Germany In 1923 the Enigma was patented These machine involved discs known as wired wheels there's switching combinators With every encryption or decryption one of these discs changes its position to give rise to a new alphabet So if you had five of these cipher Hebern Electric Super Code 3rd Issue 1992 o CRYPI'OLOG o page 33 FeR eFRCIA YSE 9NI JV DOCID 4036135 wheels you have a potential of 11 880 000 The inventors of the first German rotor machines we're not sure whether it was Scherbers or Kom In 1923 an American by the name of Edward Hebem out in California also invented a rotor machine to which he gave the name the Hebem Electric Super Code He interested the Navy in this device it was a three-rotor machine but he was asked to build a five-rotor machine At that time the Army and the Navy were not talking to one another and Friedman who was on the Army side of the business found out that the Navy was interested in buying some of the these The Navy wouldn't let them have one of the machines so the Army bought two Then Friedman said the machine was weak Actually he said so out of his hat because he had no idea how to tackle it It had a potential 90 billion cipher alphabets Any selfrespecting cryptanalyst should have thrown up his hands in horror Anyway he found out that the machine the Navy had was not the same as the one that he had because he asked for some letters to be encrypted with a certain setting across the wheels So to make a long story short the Navy sent ten messages encrypted on this device and Friedman who was terribly unmathematical-in fact he was just very very lucky He blundered his way into solution time and again In 1917 an American engineer by the name of Gilbert S Vemam thought of a way for encrypting teleprinter signals As you know the teleprinter code is a binary code-I used to say two things taken five at a time but mathematicians objected-it's five things each of which can take one of two states on or off whatever you want to call it So the symbol for an E let's say on a teletype tape is a hole and in the next four hole positions there's no hole That's an E This is a T et cetera So Vemam thought up an idea of having a key tape prepunched random tape used to key a plaintext message tape Then he thought he could do one better this key tape of course should be one-time because then the security is infinite if the key tape was produced at random But of course it's clumsy and then there's the difficulty of distributing the tapes et cetera So then he thought why not have two key loops let's say of a thousand characters and 999 characters so since they're relatively prime it takes 990 000 encryptions before you get back to the same arrangement of the two tapes He even had an idea for less securityhe proposed the idea of a single key loop In Me II I wrote that the security of this scheme however is either negligible or only two or three times that amount To us it makes sense It's interesting that Friedman in 1919 meantime had been a good boy and gone overseas he was a captain with the AEF in the code compilation section After the war he went back to the Riverbank Laboratories and there examined some traffic sent for test purposes by the State Department So the State Department was about to use this two-tape idea To be on the safe side they sent a series of messages to Riverbank Friedman and his staff worked two or three weeks on this traffic and got no place They were sore in mind and body This was the first time they worked for such a long period without solution Everything they had come across they solved and they couldn't understand what was wrong with them there must be something wrong with them they were all losing their buttons and they wanted to leave him It was a total of five weeks that they worked One by one they wanted to leave and Friedman said Give me a last chance Let's go over our steps we'll work one more week if we get no place we'll give it up So they went over their steps by the way they had no tape printers for getting hard copy 3rd Issue 1992 CRYPl'OLOG page 34 F9R 9FFIOfAiJ l JSI ONLY DOCID 4036135 they had the teletype tapes sent to them and they had to transcribe the holes et cetera with the Baudot code in front of them on sheets of paper And in the transcription one chap left off a character that happened to be at the crucial spot By the end of the week they solved the system WORLDWARll We can go in the last five minutes to World War II to give you some idea of the COMINT successes we've had By the way these items came out in a very dry article in Time magazine shortly after the war when they had the Pearl Harbor investigation when everything was revealed to our detriment That particular Time issue said that through MAGIC which was the then cover name for our COMINT product it enabled a relatively small U S force to intercept a Japanese invasion fleet and win a decisive victory in the battle of the Coral Sea thus saving Australia and New Zealand It gave the U S full information on the size of the Japanese forces advancing on Midway enabling our Navy to concentrate ships which otherwise might have been thiee thousand miles away and thus set up an ambush which proved to be the turning point of the Pacific War directed U S submarines to sea lanes where Japanese convoys would be passing and made possible the reading of messages from the Japanese Ambassador Oshima in Berlin often reporting interviews with Hitler giving our forces invaluable information on German war plans When the lid was blown which was a dirty shame because the world at large knew the cryptanalytic potential of the United States In the report of the Joint Committee on the Pearl Harbor attack there are two other quotes I would like to read you With the exercise of the greatest ingenuity and all obvious resourcefulness regarded by the committee as meriting the highest commendation the War and Navy Departments collaborated in breaking Japanese diplomatic codes Through the exploitation of intercepted and decoded messages between Japan and her diplomatic establishments the so-called Magic a wealth of intelligence concerning the purpose of the Japanese was available in Washington Another quote Important diplomatic messages were intecepted transmitted to Washington decoded and translated and disseminated with utmost speed Not infrequently they were in the hands of the authorized recipientsofMagic in our government as soon as they were in the hands of the Japanese overseas Many of the civilian and military personnel engaged in the handling of the Magic worked long hours far in excess of those prescribed without additional compensation or special recognition Now this is in italics The success achieved in reading the Japanese diplomatic codes merits the hightest commendation and all the witnesses familiar with Magic material throughout the war have testified that it contributed enormously to the defeat of the eneey greatly shortened the war and saved many thousands of lives In fact one estimate-they went through three or four sheets of foolscap to show this -and it was General Chamberlain who was then G2 out in the Pacific-that one dollar spent on COMINT during World War II was the same as a thousand dollars spent elsewhere Editor's Note It seems evident that we in many ways inherited the many talents ofhis father who was a renown theologian a gifted linguist and an authority on Byzantine music Coincidentally his father was my mother's professor ofmusic and theology in Athens So it seems only right that I in tum became LDC's student -not in theology or music but in that landmark course in cryptanalysis CA-400 3ni Issue 1992 CRYPTOLOG page 35 FeR OPPICtAf USE 6NLY DOCID 4036135 T Report Fujitsu's Atlas G-160 Machine Translation System A Preliminary Evaluation P L 86-36 P L 86-36 EOL4 P L 86-36 I I In 1986J points with regard to specificc JapPI i C a t l o ns are di 'ed and the overall system evaluated ___________________ began evaluating commercially available Japanese-to-English machine translation MT systems developed and manufactured in Japan After a three-year evaluation period which focused on appraising the quality of raw translations produced bYIleaJ lyeverysyste m on the marketc Jdecided in July 1989 to PU chase Fujitsu's Atlas G-160 systeIll aperson al computer PC -based systeIlljlJstplaced on the market that April c J nitiated purchasing arrangements in July 1989 and took delivery of the system in late December 1989 1 ---J1has been evaluating all aspects of the Atlas's role in producing translations from a range of original Japanese text This report is a product of that evaluation L - Dma e t current machine translation technology in general follow in Section IV While it s e fica1lyaddressest t t tlass Ystem l Ulditsuse 1it1liIltheE Jmvir IllIlent this rep0rtisbeing presented in tile hope that it will help readers gain some insight into the machine translation proce s enablethem to measure current sys m capabilities against their own particular translation requirements and ultimately place them in a better position to judge whether a machine translation system could be right for them L The Role of Each Component in the Document Processing-Translation Sequence The document processing-translation sequence using the Atlas system involves five basic steps In order to present a meaningful evaluation of the o Japanese-Language inputting and creation ofa Atlas G-160 system it is first necessary to briefly corresponding Japanese-language file on the introduce the system's components and each system inputting can be done in anyone of three component's role in the document processingways-via keyboard floppy disk or OCR translating sequence This information is o pre-editing of the Japanese-language document presented in Sections I and II Just as crucial to in the newly-created file optional the output quality as the machine itself is the type o translating the file using one of two programs and quality of documents inputted for processing batch or interactive translation o post-editing of the translated file optional Section III of this report outlines the kinds of o outputting of the translated document as a documents c Jis translating or hopes to printout in one of several formats original translate using the system The actual evaluation Japanese text with side-by-side English of the system as a whole begins in Section IV translation English translation only etc Based 011 the background offered in Section I through II the system's strong points and weak P L 86-36 3rd Issue 1992 CRYPrOLOG page 36 FeR eFFICtA-L USE 6Nh j DOCID 4036135 High r------------- Written Style Clarity Conciseness Worst o - - -- r----- - - --r-- ---1 Sciences Liberal Arts Low Subject Field High Figure 1 Document Types Atlas Most Effectively Translates Document Inputting and Creation of Japanese-Language File This step involves taking a Japanese language source document--either in printed form or on a floppy disk-and creating a new electronic version of that document in a Fujitsu EPOWORD-G file which subsequently can be processed by the Atlas program This process can be done manually via the keyboard basically re-typing the original document into the computer by transferring from floppy disk or by use of the OCR o Manual Input The Atlas incorporates a JIS key-board for the manual inputting typing of documents into the system Conventional Japanese text consisting of kana syllabary and kanji Chinese characters is input as either Roman letters or kana Additionally keystrokes convert the letters or kana to the required kanji The keyboard enables the input of 10 000 kanji in cluding the most common Chinese simplified kani the English Greek Russian French Italian Spanish Portuguese Danish Scandinavian and Norwegian alphabets and hundreds of symbols used in mathematics science and technology and graphics compilation o Floppy Disk Input The system incorporates an MS-DOS conversion function that enables input via floppy disk o OCR Input This device consisting of a scanning unit which looks much like a desktop copier a connection unit and related software reads in hardcopy documents placed on the scanner brings up a bit image of the document on-screen for the operator to selectively edit and then produces a standard EPOWORD-G file containing the newly-created electronic document The OCR can process typeset or word-processed Japanese text in a variety offonts pitch and type sizes although it cannot process text whose foreground or background contains color It has a character recognition speed of about 20 characters a second and can process a page containing about 1 400 characters in roughly three minutes The OCR was designed to process Japanese text only although Fujitsu is currently developing an English-text OCR The device is simple to operate the operator sets various processing parameters vertical or horizontal text format pitch document name etc pushes the START button and feeds the document through the device Each page's bit image is displayed on the monitor allowing the operator to electronically adjust skew erase spurious images and select exactly how much of a page is to be processed Each page is then processed by the OCR and compiled as a Japanese-language document The operator then merges the pages to form a single document and begins proof-reading prior to translation Pre-editing Pre-editing along with post-editing is what separates a raw translation from a polished 3rd Issue 1992 CRYPrOWG page 37 FOR OPPICtkL U 6NLY DOCID 4036135 translation It basically involves clarifying and simplifying complex grammatic structures into text more amenable to successful machine translation Since the Atlas system provides better Le more accurate and more understandable translations from text characterized by relatively short clear and concise sentences pre-editing even ofthe most basic kind result in significantly improved end translations On the Atlas system the process includes for example artificially pluralizing nouns and pronouns enclosing words and phrases in brackets to clarify governance and string boundaries and replacing complex syntactic structures with simpler ones Pre-editing can be carried out in two ways on the Atlas manually or through a software program called suiko brush-up o Manual Pre-Editing This method carried out either before or during the course of interactive translation see Translation Methods involves the operator simply going through the document manually and changing words and or grammatical constructions which the translator perceives based largely on previous experience will represent a stumbling block for the Atlas translation program Generally the operator is concerned with only the most egregious errors since minor changes can be made during postediting o Suiko Pre-Editing This method employs a distinct program selected from the Atlas menu which is run against a document file the program flags on-screen such items as incorrect kanji not quite the same as a spell check function the incorrect use of parentheses overly long sentences missing subjects inappropriate kanji and superfluous or ambiguous expressions The operator then has the option of correcting the items flagged or ignoring them Although extensive the process is extremely time-consuming especially if a multipage text is being processed In fac c Jperators do all pre-editing maIlually during interactive translation Translation Methods Th Atlas system carries out translation in two user selected modes batch and interactive P L 86-36 o Batch Mode In this mode the system attempts to translate an entire document file in a single operation No operator intervention is involved The only information displayed on the monitor screen during batch translation is the total number of sentences in the original text and the number of the particular sentence being translated o Interactive Mode The interactive mode involves extensive operator involvement the operator selects exactly which sentences in a text are to be translated pre-edits re-translates adds or changes words in the Atlas dictionary system or post-edits as necessary Atlas Dictionary System The Atlas system incorporates an 80 OOO-word standard dictionary and a user's dictionary as standard onboard programs In addition specialized dictionaries in over a dozen scientific fields such as physics and information processing see Section 2 B above representing the full set of specialized dictionaries offered by Fuijitsu as of late 1989 also are available options The dictionary system allows the operator to adjust English-Japanese word equivalents in order to improve translation accuracy-changing words adding new words deleting words or adjusting word priorities In the user's dictionary there are two ways to register word priorities The first method is the so-called kan'i simplified method used for registering nouns only When a Japanese noun encountered in a text has not been registered correctly in any of the three dictionaries standard user's specialized the operator extracts the noun inputs it onto the kanji dictionary screen types in the appropriate English translation selects the appropriate noun class object personal noun phenomenon organization etc and selects the appropriate English plural form The operator then goes back to the text and re-translates the sentence s containing the newly-registered noun Registration of other parts of speech is a far more complex process involving for example careful classification ofa verb's semantic and syntactic functions in both English and Japanese 3rd Issue 1992 CRYPrOLOG page 38 FOR OFFIGI L 'b USI l 9NbY P L DOCID 4036135 c Jlrattb afo ti nReports These Post-Editing Methods The Atlas system allows for post-editing ofthe translated English language text either as an integral part of the interactive translation process see above or as a separate step once the English translation has been isolated in a separate English language text file The software features a number of word-processing functions common to U S word processors such as move and copy commands but is still somewhat awkward to use Output Methods The Atlas system can output translations in two forms-printout and floppy disk o Printouts Through the use of Atlas's media conversion program processed documents can be printed out as original Japanese text only original pre-edited Japanese text only original pre-edited Japanese text with English translation with or without Japanese-English vocabulary for those sentences the system cannot translate or English translation only This flexibility allows the operator to maintain a hardcopy of each stage of document processing transla if desired o Floppy Disks Atlas output Le English translations can be output onto a floppy disk but when transferred to English-language word processors will contain only the English translation and no part of the original Japanese source document II Document Types For Translation On The Atlas Documents processed thus far onDAtlas system are of two broadly-defined types draft Japanese-language information reports produced by Japanese national and other Japanese-language documents ranging broadly from technical reports and newspaper articles to contracts and specification tables and charts The differences in the two types of documents are important to consider as they have a major impact on how the documents are processed on the Atlas system and what results are achieved see SYSTEM EVALUATION I 86-36 I Japanese-language documents produced in-house byc Jan ysts contain largely science and technology-related information extracted from original sources After translation into English these draft reports are edited and eventually published as final-form information reports Most of these draft reports contain English-language glosses fOl'the more complex specialized tenns and all reports submitted for machine translation are in the form of work-processor printouts Because the reports are written by thec J own analysts the style in which the reports are written can be controlled to a certain extent The draft reports are generally no longer than four or five pages Other Japanese-Language Documents These documents while in Japanese come from various sources and generally do not contain Englishlanguage glosses for specialized vocabulary Ihasno controLoverthe style Moreover in which these documents deal with P L 8 6- 3 6 communications and electronics and are at least ten pages long All are submitted for translation in the form of xerographic copies or hardbound books ill Sysytem Evaluations Capabilities vis-a-vis ASD Applications Having briefly described the components ofthe Atlas system the role each component plays in the MT process and the types of documents meant to be processed on the system a more meaningful evaluation of the Atlas's capabilities in processing o uments can not be offered The I central question fo f course is this Can the Atlas Il vide acceJ tabletranslations of th et v0 roadlY-definedtyptlsofl I documents P L 86-36 ro The answer to this question is essentially this since its operational introduction in February 1990 the Atlas system has shown itself to be generally NOT suited for producing POLISHED translations ofl Idocuments particularly given the wide variety of topics and writing styles these documents encompass rather the Atlas is most efficient and effective in producing RAW 3rd Issue 1992 CRYPI'OLOG page 39 POR OPPICtAi l USI l ONLY DOCID 4036135 1 ROSTERS OF PROPER NAMES personal names plaQ names company names etc POOR without devoting a lot of time to dictionary registration OUTSTA 'iDING after completing dictionary registration 2 NEWSPAPER ARTICLES FAIRLY GOOD for articles on new products or technologies TERRIBLE for articles on politics economics Atlas cannot handle journalistic Japanese w out extensive pre-editing 3 ABSTRACTS OF TECHNICAL PAPERS EXCELLE 'T concise style well defined terminology 4 TITLES OF S T-RELATED BOOKS AND PAPERS OUTSTANDI -'-G concise style technical vocabulary probably already in technical dictionaries 6 USER'S MANUALS FOREQUlPMENT EXCELLE 'T unless the tranSlation must contain both Japanese and English teXt e g translation of Atlas User's manuals 7 TABLES CHARTS MAPS GRE T rranslations but cannot reformat back into original graphics the latest Arias Ver 13 DOES have this capability however 8 SHUKA ''' MAGAZINE ARTICLES VERY POOR generally vety colloquial style requires extensive pre-editing 9 UDC DRAFT INFORMATION REPORTS POOR TO 'ERY GOOD depending upon style and subject of original document see fig 1 10 NUMERICAL DATA priee lists etc OVTST -''iDING 5 CONTRACTS WORTHLESS too many problems with long-winded legalese styles Figure 2 Results of Testing Atlas' Ability to Raw Translate Various Types of Documents F L 86-36 translations from documents written in a very con cisestyle and dealing with as narrow a technical field is possible see fig 1 In other words the Atlas systeQl cannot provide acceptable translations of ALL type ofl documents instead its strength lies in its ability to provide in a minimum amount of time a high volume of raw translations of certain kinds of documents see fig I 2 more expensive mainframe-based MT systems Its weak points on the other hand include technical limitations of the OCR in processing certain character styles and document types the need to manually check all OCR output the extreme difficulty of the pre-editing process the fact that the translation program provides no translation whatsoever in cases where it finds the text too difficult and the relative awkwardness of the system's English work-processing capabilities in the post-editing process Underpinning this evaluation are the strengths and weaknesses of the Atlas system itself There are many positive features to the system which streamline and simplify the processing-translation Evaluating the system's strengths and weaknesses in the various steps of the MT process sequence However it is unfortunately the system's inherent technical weakness which place reveals the following specifics restraints on the Atlas's overall translation Input Process More than 99 percent of the capabilities documents processed to date on the Atlas have been input via the OCR The device has proven Atlas Strengths and Weaknesses indispensable to the Atlas translation process as Their Impact on the NT Process it allows hardcopy documents to be read into the Among the system's strong points is its ability to computer without having to type them in input via both floppy disk and OCR in addition to manually word-by-word This of course saves manual keyboard input Also its interactive time on the part of the operator and represents a translation mode and extensive dictionary systems major system strength However in many cases add an important measure of flexibility and much of that time can again be lost when the expendability to the MT work process Overall operator goes to proof-read and correct the OCR the Atlas menu system and operating inputted file Depending upon document environment are well-designed and user friendly characteristics the OCR device can reproduce text and the fact that the system is PC-based makes with an accuracy rate as high as about 95 percent the system a very attractive alternative to larger for original hardcopy that has good contrast and contains only Japanese text and as low as 40 3rd Issue 1992 CRYPI'OLOG page 40 FQR WFICIAb Y8I 1 QNLJPY DOCID 4036135 percent for third or fourth-generation xerographic copy with spurious or colored images poor skew English-language text or mathematical formulas etc Any inaccuracies must be corrected through manual proof-reading Japanese work-processing systems do not have highly-developed spell-check capabilities which in the case of many OCR mistakes - can be extremely time-consuming and tedious l Specifically the OCR has difficulty with text containing any of the following o non-Japanese script including English o handwritten text o text with color foreground background o poor xerographic copies o complicated mathematical formulas Greek symbols etc o underlined characters o free-form designer characters o characters smaller than 7-point in size o Layered characters not clearly separated see fig 3 o unusual characters outside the finite set included in the OCR software I I P L 86-36 proficiencY to be done properly and requires a considerable amount of time to pre-edit any but the simplest of writing styles Moreover the suiko or brush-up program that comes with the Atlas and which is intended to help facilitate the difficult pre-editing task actually does little to reduce the amount of time required For these reasons Dperators have found that attempti n polished translations-and doing a complete job of pre-editing-is generally impractical opting instead for a raw translation A quite satisfactory raw translation can very often be produced relatively quickly for S T-type documents so long as the source text is clearly and concisely written and characterized by a clearly defined set of vocabulary which is often the case with such documents If subjected to some simple post-translation English-language post-editing moreover such raw translations can sometimes approach the quality of fully preedited polished translations P L 86-36 Translatwn Process and Dictionary Systems The batch translation mode because it does not allow for human interaetion for editing purposes and cannot be viewed by the operator has been found bYL operators to be almost useless The Although this seems long most normal wordinteractive translation mode on the other hand processed or typed-set Japanese-language text has proven to be the better method by far and conforms to standards within the range of the certainly represents one of the Atlas's strong OCR's capabilities One solution to the OCR points This mode offers the operator tremendous dilemma i e quick read-in time but possibly flexibility It gives himlher the freedom to decide time-consuming proof-reading and correction how much or how little to pre-edit the ability to process is input via floppy disk If the source document call be illiti Y crea d suchas aD make dictionary changes in the Inidst of the translation process and the option of redraft information report or downloaded such as translating any particular sentence as many times an on-line data base file etc as an electronic as desired An operator for example would document then it could be input into the Atlas system directly via floppy disk bypassing the OCR generally have the system translate the first few and accompanying proof-reading entirely Having sentences of an input text Based on how well these sentences were translated the operator both the OCR and floppy disk as input options would then register any new words in the user's provides much flexibility and is a definite strong dictionary pre-edit as necessary and re-translate point of the Atlas system the entire text interactively repeating this Pre-Editing Process If the Atlas system has an procedure This procedure generally leads to a Achilles heel then this step is it The reason translation that while not a truly polished one is being that it is the difficulty of this particular step substantially better than a raw one which makes completing a truly polished Post Editing The Atlas system while it does translation such a painful task This process is possess some rudimentary English-language the most complex of all system procedures It word-processing functions generally lacks the requires near-native Japanese-language 3rd Issue 1992 CRYPI'OLOG page 41 POR OPPIOw USB ONM T DOCID 4036135 A Lavered charaa ers not easily delineated ' 7 7 ' 'O B Characters not clearly separated Figure 3 Example of Character Types Frequently Misread by the Atlas OCR u s English-language word-processing programs ments fields It has thus been found that using a Wang or IBM System for post-editing is far more efficient than For example a 20-page typewritten document dealing with telecommunications and written by using the Atlas specialists in a very plain style can be input via Output Methods The ability to furnish to post- the OCR proofread translated in raw form and output as an Atlas printout in anywhere between editors and consumers hardcopy printouts 4 hours very few OCR mistakes and 10 hours containing the pre-edited Japanese text the English translation and English translations of many OCR errors This raw translation can be individual words in sentences the Atlas was given directly to a consumer for determining whether further translation is necessary If no unable to fully translate represents a strength of the Atlas System However since files for transfer further action is required by a translator hunto English-language word-processing systems via dreds of dollars have been saved by not having to floppy disk can contain only the English work for two to three weeks to produce a polished translation the use offloppy disk output for translation that is not needed by the consumer If a polished translation IS required substantial anything but polished translations is prohibited amounts of time and money are still saved c J IV Conclusions and Recommendations operators have found that on average and Atlas raw translation of the same type of dOCUIllcnt The Atlas 0-160 system represents state-of-thefollowed by off-system post-editing by a translator art machine translation technology yet at the results in producing 10 pages in the time it takes same time displays some of the fundamental technical limitations seemingly generic to all such to translate seven pages by conventional methods only systems today Simply stated while the machine has been well engineered and slickly packaged it is not ablenor does it claim to be able-to produce consistently coherent and accurate translations across broad or varied subject fields Current algorithms and associated natural language processing techniques simply are not that far advanced And while the use of extensive time-consuming preand post-editing procedures can somewhat remedy this basic shortcoming such procedures are almost always extremely inefficient For these reasons the Atlas MT system does not represent the ultimate answer to a1lDl'8J 1ll on needs Rather it is a tool which-with the proper investment to time and effort-can yield definite benefits in improved productivity in specific docu- On the other hand c J perators have found that a polished translation of ven five or six pages of a handwritten document not related to an S T field and written in a too-elevated too colloquial or too illogical style invariably takes anywhere between four and five times longer on the Atlas system than by conventional translation methods Use of the Atlas for translating this type of document is simply not worth the trouble The important thing therefore is to use the MT system where it is most effective where it can assist by speeding up the translation process instead of slowing down that process Forc J that means using the Atlas fpr documents which areconcse and well written from thestandpoint of style andvocapulary whichis most often the 3rd Issue 1992 CRYPrOWG page 42 fOR OFRCIM B6B ffiiLJPY P L 86-36 DOClD P L 86-36 4036135 case for documents in S T-related fields Moreover it means concentrating-for the most parton producing a high volume of raw translations for subsequent screening by consumers Encouraging the use of input vi fl PPYdisk-Ddraft informatiollreports for example-would also contribute to increased productivity by eliminating the need to manually check for OCR errors In sum if used in a way so as to exploit its strengths instead of its weaknesses the Atlas MT system can be a valuable tool in increasing translation productivity yet it is no dream machine and should never be purchased or used as an excuse for not hiring training and retaining the very best translator work force possible v System Components Hardware Fujitsu GIGO oc with a single 5 25 floppy disk drive Expansion memory to 8 megabytes I35-megabyte hard disk drive JIS keyboard Color display Mouse Kanji printer Optical character reader OCR connection unit Software Atlas-G set 30-dot character group MS-DOS data connector Image processing option OCR control option Dictionary biology medicine Dictio industrial chemistry Dictionary meteorology seismology astronomy Dictionary mechanical engineering Dictionary civil engineering construction Dictionary physics atomic energy Dictionary transportation Dictionary electricity electronics Dictionary mathematics information Dictionary plants factories Dictionary automobiles Dictionary biochemistry Dictionary information processing Cost Cost of Hardware Cost of Software Total Cost JPY5 974 700 JPY3 6I4 950 JPY9 589 650 o o o oo We welcome reviews and reports of hardware software training materials books technical literature and conferences ooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo 3rd Issue 1992 CRYPTOLOG page 43 FeR 6FFICfAiJ USI ffiftN DOCID 4036135 o o o _ o o o o o o o o oo o oo o o oo o o ooo o oo o oo o o oo o oo oo o o o o o oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo o o o o iPOY an admission ofvulnerability i e CU There has been considerable discussion in the Agency about computer viruses and the oo possibility that NSA's computers and computer o networks have been invaded In discussing the viruses writers must be aware of the classifica tions that apply to the discussions o o CU As most of us know a computer virus is a software program designed specifically to repro duce itself and to modify or destroy computer o software damage or destroy equipment or como promise sensitive data Some of us however are unaware that any personal computer network or office automation system is susceptible to inva sions by a virus infecting any host in which the oo program is used Owing to the insidious nature o of a virus any unwitting user can become an unwitting propagator oo o D We have established classification guidelines o to discuss viruses that may be summarized as follows oo U the fact that NSA is aware of computer viruses and that we take steps to minimize the risk of introducing viruses into our automated information systems AIS or networks is UN- o o CLASSIFIED oo CD the fact that we employ commercially oo produced software to scan for virus infections oo also is UNCLASSIFIED oo the acknowledgment that NSA has experienced the intrusion of a computer virus in any of its systems is classified at a minimum CONFIDENTIAL oo o o o o o o oo oo o POUO The disclosure of the extent of infec- o o o tion or the name of the specific virus that may oo have been discovered in an NSA AIS or network is classified at a minimum SECRET o VOTTO Specifics concerning an infection such as the severity of the infection 'the extent of damage done the complexity and expense of eradicating the disease or the impact of the virus on operations is classified TOP SECRET in some instances if certain details concerning the AlS the network or the database in which the virus was discovered are revealed any of the above revelations may require handling in COMINT channels CHVCCO or even in codeword channels POUD In summary be careful what you say and to whom you say about NSA and computer viruses For further information concerning the classification matters pertaining to computer viruses contact your classification advisory officer CAD your local Computer Security Officer or Computer Security Manager or J06 the TCOM office of Operational Computer Security o o oooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo 3rd Issue 1992 CRYPI'OLOG page 44 FOR OFFIGJA - UBJ ONIooY o DOCID 4036135 From the Past The Department ofDefense is commemorating the 50th Anniversary of World War II in 1991-1995 Unlike the immediate post-WW I period when cryptology went underground the organizations involved during World War II remained on the post-war scene Even in the darkest hours ofthe war it had become evident that one ofour one-time allies was already an adversary To a great extent this cryptologic agency owes its continuance to the Cold War that had its origins in the bitter war years 3rd Issue 1992 CRYPI'OLOG page 45 P SBGRF l'l' YMRRh EO 1 4 c P L 86-36 DOCID 4036135 they sometimes require lots of work to refine and forge a final product Mavericks are the prospectors who find them Letter But sometimes it's easier for a leader to decide that the sparkle of a new discovery is iron pyrite fool's gold -and tell the prospector to leave it in the ground Unfortunately when that happens too often the loss can extend far beyond one idea Like prospectors whose claims never pan out some mavericks just give up and stop having new ideas Still others keep having them but for new employers after they resign or are penalized for sins like unpredictable creativity -against which at least one military officer has been strongly counseled The truth is that mavericks make us feel uncomfortable they question authority they rock the boat and they don't understand how we do business To the Editor But these three phrases are often the best descriptions of really new ideas For example Kodak When General Matthew Ridgeway retired as them in 1948 when they turned down an applied Army Chief of Staff-this was the man who had entered Nazi Germany as commander of the First inventor named Charles F Carlson when he proposed a new copying process Carlson's idea was Airborne Division who relieved General Douglas to use high-voltage electrostatic charge to attach MacArthur as Supreme Allied Commander in Korea who stopped the Chinese advance while in fine black powder to plain paper then to heat the command of the Eighth Army-he was asked what paper and melt the powder into its fibers Kodak turned Carlson down flat the proce s was too he considered to be most important accomplishcomplex the machines were too expensive and ment of his career His response was simple 1 anyway the whole thing was unnecessary--everyprotected the mavericks General Ridgeway's one at Kodak knew that if you want a picture on answer may seem surprising but it reflects a paper the best way was to start with a picture on profound understanding of both the realities of Kodak film organizations and the requirements for their success Kodak was wrong Carlson took the ideas that Kodak turned down and sold them to an unknown In theory protecting the mavericks is easy company named Haloid Corporation Today mavericks are the people whose new ideas and Haloid's name is Xerox approaches make their organizations uncomfortable today but that will be invaluable in solving the problems of tomorrow We all know the TQM school solution -new ideas are good and people who have them must be encouraged rewarded and protected What could be plainer and why was General Ridgeway proud of such a simple thing There are two important lessons here but they're not simple ones like don't turn down another Xerox or don't make 'bad' decisions There won't be another Xerox-the next revolution will start with another idea-and at the time the decision to tum it down wasn't a bad decision it was a good decision made in the wrong context Of course the real problem is much harder Good ideas are like gold they are rarely unalloyed they Carlson's process was expensive it was complex are often found in unattractive surroundings and the copier repairman is still a standard figure in 3rd Issue 1992 CRYPrOLOG page 46 F'Ott 6PPICtA L USE ONLY DOCID 4036135 business cartoons and it would have been in context and suggest the new solutions for difficult direct competition to Kodak's traditional busiproblems--questions like Why are we draining ness-and there was a good chance it wouldn't the swamp and solutions like First thing let's work outside the lab In the everyday context of a tame all the alligators Their ideas may be anKodak operating manager investing-not declinnoying frustrating embarrassing and even ing-would have been the bad decision After threatening perhaps we picked the swamp ourall the project involved high-cost substantial risk selvesD but they may also lead us out of the muck and the best foreseeable outcome would be and up to drier ground Perhaps Robert Kennedy launched another competitor for Kodak's existing thought of himself as a maverick when he wrote positions in the markets for film and photographic Some people look at things as they are and ask supplies The real measure of the right deci Why sion-Xerox's subsequent commercial successI think of things as they could be and ask wasn't available until years later Why not So what are the two lessons I think they are Don't think of your mavericks as trouble-makers these or boat-rockers think of them as principal staff 1 Always look for the largest reasonable context for thinking of things 'as they ought to be' -and in which to evaluate a decision and if they do tame the alligators how easy it will be 2 Have the courage to risk some bad decisions to drain the swamp In the Kodak-Xerox case making the decision in Anonymous the largest context would have begun with recognizing that Kodak's real business was putting pictures on paper not selling film and photographic supplies Seen this way Carlson's invention of plain-paper copier that could be used by people with no special training right in their offices fits right in as a logical extension of and not a competitor to Kodak's existing business Of course neither of these lessons is as easy to apply without the benefit of hindsight as the saying goes When you're up to your ass in alligators it's difficult to see that your objective was to drain the swamp It's even more difficult to see something like economic enhancement through the provision of retail jobs in a suburban shopping facility in a soon-to-be-drained swamp as the largest reasonable context With alligators alongside it may seem like the only reasonable context for any decision is keeping all my body parts or just not becoming someone's lunch At times like these we need reminders like the slaves who accompanied victorious Roman generals in their triumphal parades reminding them of their mortality But our reminders should be to consider the largest context and have the courage to take risks And that's exactly why we need mavericks They are the people who ask questions in the largest ooooooooooooooooooooo oo Bulletin Board oo oo oo ooooooooooooooooooooo FOYO L143 has a degausser for large objects like disk drives It can handle objects 8 x 18 x 22 The degausser is located in SAB-3 It has been approved by T03 now J06 FOW In the past we discarded defective disk drives This will allow us recoup our investment For example SUN disk drives cost about $16 000 each FOYO To degauss your drive write a memo to J06 with specifics about the drive 3rd Issue 1992 CRYPTOLOG page 47 FOR OFFIGJA - l JS8 ONLJPY DOCID 4036135 Editorial ABOUT TIlE DlSTRlB JTI ON OF CRYPTOLOG We regret to inform subscribers that we lost our ability to distribute individual copies beginning with 3rd Issue 1991 We're looking for an alternative For the time being we ask you to be patient And as there are many ongoing changes in organizational designators and people are moving about it will take quite a while until everything is on even keel again So please be patient Meanwhile if you have missed copies you can ask for them using the form shown below or write a note to CRYPTOLOG Distribution P0541 Ops-1 As there are over 3 000 subscribers it stands to reason that you should not call about your subscription or distribution For your convenience the contents of the issues published beginning with 3rd Issue 1991 are shown below NO TELEPHONE REQUESTS FOR BACK COPIES WILL BE HONORED 2nd Issue 1992 3rd Issue 1991 EDITOR IAL o o ' DooPs -_ ll ii _ 1 _ 101 _ ACIlIEVDa IIT 'nItOOG 8 I'EClI'LZ TIlE AlQI YST MID AtJTCIMUIa SIGIJrr Df All OPD SOCIZn o o 13 17 u n VIEW OF L o o L Geo J l j'i Har l II Jlrow a o ' Stuart Bw lc o - - - - - 'fIlE u x ICOCi It PD a DO I AJID lUlU I OODIG IIDE o o o 1lRI Ar1'D XJGl III 03CBID20RLU _ IQJOR 31 iii J Enf01'E ADDRESS o noN 'rIlE AS'r EO 1 4 c P L 86-36 ClDE '%'0 SIGIK'2' mtUl CIc 33 0 IZftEIlS 011 A MODEJUI y sroaT GOLDEII OLOIZ o o TO COH1'1IJIUTE o o 2 o 2 _ J 3 ELECftCIrIc i'tJIILXDDG onZ 1lEPC8C' aADIO aIIOrADCaSft It I S J I QAa 1ft S%G U nc r L mI'lCllUL o 1'0 'UIIlIf o 86-36 1st Issue 1992 GBPR Be't'ty 1faDa't_ 3 o o o o Peter D JIol E J WL 1M MEMOIlIAK o o o o o JOO LE - ooo ' 10 13 'L - rI o ticl ard uz S l g TECHNICAL l I'tUA' ORE REPORT o 33 CQNrERENCE REPOR'T J4 ITCC' 91 F1 CM T1IE PAST o oo L I J OFSECASA '%'OOl EDI70RZAL GOWEN OLDIE TO CONTRIBDTE I o 3rd Issue 1992 W1IEJI E THE no N TJU aE MEVEJl WAS A IOGE'YMAN_ - 1 5 3 OM 2'D or 'HZ 0YS'fD o v au DICl'Oar o vun %DIE o -41 '- 1 trlDIDaAnc s s o ooo o o o o o o MISSlc s ru-crlc s f GJ Izuu s J DSOllRZl o 42 - - snu oo A 4' I % I I G t C_ err A USTOa o o o 17 1 1 or caD %OLOC't o o o o L D CIolliaabo 2 3 w on- Jylvootu IoST o o o _ o o _ o o _ ' L ud n a t TIlE 42 43 44 - 45 1 GBPT E1 45 o o _ o o o o o o o o _ oo _ o o o o o o o o o 46 D aa JtD_ o o o o o o o o o o o o o o o o o o o o o o o o o 47 EDI'fOJLI 1tJ o o o o o o o o o o o o o o o _ o o o o o o _ ' 'O conJt taoTE o o oooo _ o o o EO 1 4 c EO 1 4 d P L 86-36 PLEASE SEND ME TIiE ISSUES MARKED BELOW 3RDl991-- NAME ---------- 1ST 1992-- ORG---------- 2ND 1992-- BLDG -------- 3RDl992-MAIL TO CRYPTOLOG DISTRIBUTION P0541 Ops-1 3rd Issue 1992 CRYPTOLOG page 48 FOB QFFIGMdJ USF 6NLJPY o o o o o o ooo oo 49 86-36 DOCID 4036135 CRYPTOLOG Editorial Policy CRYPTOLOG is a forum for the informal exchange of information by the analytic workforce Criteria for publication are that in the opinion of the reviewers readers will find the article useful or interesting that the facts are accurate that the terminology is correct and appropriate to the discipline Articles may be classified up to and including TSC Technical articles are preferred over non-technical classified over unclassified shorter articles over longer Comments and letters are solicited Weinvite readers to contribute conference reports and reviews of books articles software and hardware that pertain to our mission or to any of our disciplines Humor is welcome too Please note that while submissions may be published anonymously the identity of the author must be made known to the Editor Unsigned letters and articles are discarded ooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo N B We regret that we must now do cost accounting For every original o submission-not written for another purpose- please indicate how much time o you spent writing your article or letter and what your hourly rate is ooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooooo If you are a new author please request Guidelines for CRYPTOLOG Authors How to Submit your Article Back in the days when CRYPTOLOG was prepared on the then state-of-the-art a Selectric typewriter an article might be dashed off on the back of a used lunch bag But now we're into automation We appreciate it when authors are too N B If the following instructions are a mystery to you please call upon your local ADP support for enlightenment As each organization has its own policies and as there's a myriad of terminals out there CRYPTOLOG regrets that it cannot advise you Send two legible hard copies accompanied by a floppy disk or cartridge as described below or use electronic mail In your electronic medium floppy disk cartridge or electronic mail please heed these strictures to avoid extra data prep that will delay publication o do not type your article in capital letters o do not right-justify o do not double space between lines o but do double space between paragraphs o do not indent for a new paragraph o but do paragraph classify o do not format an HD floppy as DD Or vice-versa-our equipment can't cope o label your floppy or cartridge identify hardware density of medium software o put your name organization building and phone number on the floppy or cartridge The electronic mail address 6 CRYPTOLOG publishes using Macintosh and Xerox Star It can read output from the equipment shown below If you have something else check with the Editor as new conversions are being added r------------------------- SUN 60 or 150 MB cartridge ascii only XEROX VP 2 0 2 1 5 1 4 floppy only WANG Macintosh 3 1 2 DD disk only IBM Compatibles 3 1 2 DD or HD 51 2 DD orHD Stand-alone or Alliance Please furnish a copy in 1EXT as well as in your software as we may not have all the software upgrades Please furnish a copy in ascii as well as in your software as we may not have all the software upgrades 3rd Issue 1992 CRYPTOLOG page 49 FOR OFFICJA L YBIS O a JPY P L 86-36 DOCID 4036135 TillS BOCI JMBNT CONT hIN C09BWQR9 AtAT RIJ b NOT RELEASABLE TO CONTRACTORS 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