• • IOOJVOOlila'ilk l E100f Ci ' v IBBl IWl IO l I wua1mo lll 1Wfi lillllD HIDDEN LOSSES IN COMINT PRODUCTION •• • • Robert Gould •• •• •• 13 arm • oar11•• 80DI911111 · · · - · · · PJIILZELII Pl•- - Declassified and Approved for Release by NSA on 04-09-2021 pursuant to E O 13526 MDR-109431 EO 3 3b 3 PL 86-36 50 Most of the Goiden OZdiss published in CRYP'l'OLOG so f I 1' have bssn light in tone but the Editor feels that this dBpartment should also Nprint csl'tain sel'iOWI wrks that continue to remain golden and that reade rs may have missed when they wre fiPst pubtishsd One such wPk is the follol ng article by PetiNd NSA-er Robert Gould which s o l'igina ity published in KEJWORD June 19 1 use 3605 sages •which he decides to read and file ' that • f •• is f le rather than process to his checker or confirmation of the decision Ass'Ullling the lin •• guis has not delayed his decision in favor of••• proiessing other messages first and that the checker clears his read-and-file box promptly • tffll time lost may be slight but if neither of tnese conditions exists the effectiveness of the •• it is jeopaTdized If the unit contains a • •• celatively large proportion of new personnel •inexperienced and uninstt'Ucted in requirements •• • the hazards are magnifi ed • •• le is that of the new emThe sim le t e ployee • In discussing the function of language in tae production of COMINT much emphasis has been • placed on the h l'111 that can be done by mislealing translations Such errors may be potenttally the most damaging but there are other ercors that can and do result in consistent and lortg term losses that may never be made up So of these which are discussed in the following paragraphs are obviously ticklish questions and no recommendations for solution are made here Only the people directly responsifle for the individual problems are in a positiQn to recognize and correct the causes • Failure to Recoqni ae Inteili gsnce In o tion e trans ator s m sta e was un erstan a le in l i ght of his newness It is •• a little harder to understand why he had the rer sponsibility of uking a decision on the message • nte 1gence tat 1s not rec inguist will probably be lost per• manently and nobody will know that anything is• going wrong If the traffic is a miscellany and processing• is limited to issuing translations of individuat This failure may mean not recognizing the in- messages the losses will probably not be great telligence content at all or not reoognizing since the intelligence content per jndividual • that it has a very limited life span Its remessage is low anyway If the material is the • sults may be an irretrievable loss Q f intellikind that is exploited in depth for the prepara• gence because of lack of timeliness or of an tion of reports or if it can only be exploited • outright failure to publish CoD1111enly this erin that fashion the loss may be 1110re signifi- ror originates in unawareness of J'equirements at cant unless the linguist-analysts know their • the level of selection either b ause the anarequirements and the various methods for satis- lysts have not been properly inslructed or befying them cause the customer's informatioll needs have outIn one of our areas a stripped his expressed require nts A review ceived for information on of processing methods may reve 1 still other Joyed hazards though In npst organizations the inguist who is ______________Jsubmits those mese April 76 CRYPTOLOG Page 13 9 P flDODDll rlANDLE VIA COMINT CHANNELS ONLY - _ 'iOP il81 BI • • a • EO 3 3b 3 l PL 86-36 50 USC 3605 • a II • • • a I fects the newcomer_ priniarily is simply inadequate exposure to the language No linguist has at his dispoSdl the entire vocabulary of a language and to •inexperienced if confronted with volW11eS oS gew and varied material frequently caMot-11eally deterru ne its value by reading it 4'td they play safe by translating The Checki ttZeneck Languag e•rors exact a regular toll One simple mi ranslation if caught may cost very little t he lhecker merely has to make a cha nge ---o11 and if tim • permits instruct the erri ng trans I The lator fallr But one checker supervising a _ H'-9 __ n ' _ a i -i P e r 'e l q 'u' ' i re m -e n t c o ul'T d be sa t i 's fied large 1tumbft of translators may find that an ac was discovered only by chance since the response cumulation of such corrections measurably lowers necessarily depends upon the opinion of those his l 'i'oduitivity and affects the timeliness of who read the traffic end product A good many errors are not simple misotranslations of single words but fundamental Failure to Recogniae Essential Tec1mi cat mi conce tions of the meaning of an entire esInfo-r ration sJ ge hese may require extensive or even comProbably the mst easily docU111ented losses are plete 1Jewriting and a consequently greater loss 11 those caused by a failur - e t o r e co '·san1 · z e t h e i m - '11 0f time• oortance of information r • there al'e large nt111bers of inexperienced Where persdnnel there sometimes arises a situation in whic h the entire production effort hangs upon the work of a very few checkers It is not simply that the quality would suffer were it not fof them but that there would be no reliable ptoduct at al 1 This is theoretically only a temporary situation but temporary in this ease may mean a couple of years and if such a l 'lllr ' ' ' ' ' ' ' T 'l ituation exists the appropriate managers should _ _ • ere 1s no way or esti mat1ng mat uiese •know about it _______ _ _over _ _a period _ _ _ot years _ __ _ Scraping By 1osses may have_been The wol'st error is attempting to operate with inadequate resources It has not been uncommon to find areas in which the entire lexical and graimnatical resources consisted of a couple of bilingual dictionaries and an elementary introduction to the gramnar No language is satisfactorily represented in such scanty refer ences and the result of this inadequacy has been that really crucial messages could not be - - - - - - - - - - • issued in time or with sufficient clarity to ---- - - - - -- - -- - • be useful Much of the blame belongs to the MispZaoed Effort COMINT professional who often does not recognize Circumstances may require that the decision how much he depends on expert knowledge acto translate a message be left to the line quired imperceptibly over a period of 10 years linguist In such cases people have sometimes or more He consequently is inclined to rely spent hours on a message that doesn't contain on the obvious native intelligence of his junia shred of intelligence that cannot be pubors to guard them against errors from which they lished as COMINT or that is too late by weeks can actually be protected only by knowledge or In the worst instances t he production of real lacking this by adequate working aids COMINT is delayed by the translation of trivia or the product in general is degraded The time Mistakes will always occur and some of them lost can also become serious if an analyst is will be serious The object of these couments directed to devote research to a term report for is not to er adicate errors but to signal cerwhich there is neither need nor requirement tain conditions or environments that regularly engender them Although only a couple of examThere are several reasons for such errors ples were given none of the errors cited is in Unfamiliarity with requirements is often one of any sense hypothetical but all were drawn from them another is a fai lure to widerstand the personal experiences observations or specific value and use of COMINT A third one which af studies April 76 CRYPTOLOG • Page 14 71P 60601111 IL Id I 111111 • I 7 PIii JI 8R8RIRI EO 3 3b 3 PL 86-36 50 USC 3605 LETTERS ro THE EDITOR l To the Editor CRYPTOLOG After reading the December 1975 CRYPTOLOG After a short huddle to figure out some way article What Are We AboutT by LCDR Jaaies to cope with our collection system we attacked T Westwood I am compelled to comment our problem I don't know who said it first Overall the article was good food for but our motto was The difficult we'll do right thought But one sentence in one paragraph now the impossible will take a little time prompted the following tirade Our objective was to determine which of the hot signals was the hottest and analyze that one LCDR Westwood stated It follows that if the processing and reporting effort ever catches first So two of us turned to the task of previewing tapes by running oscillograph records up with the collection effort we would be in and observing the signal activity on an oscilloreal trouble because we would certainly have scope The other worker did the analysis I the cart be fore the horse sure would like to have had stock in 3M Kodak 1 have shown the article to several people who were puzzled by that statement Personally and Polaroid One day hwtdreds of rolls of oscillograph l disagree with his philosophy although I somepaper later as the two of us were la in back times think that the NSA CSS and the Intellia ta gence Co111111Unity do not disagree The philosophy that I would prefer to operate wtder is to attempt to keep pace with the collector in processing and analysis Reporting of intelligence in a seieotive manner should follow thus not inwtdating the user with reports he doesn't ' ' - -------------- ' ' 'Pwant or need We bot c eclceil t e paper reco to con rm w at I saw It was there We rewowtd to that part To illustrate the frustrations and potential of the tape and watched the scope There it was danger of that kind of thinking I will tell again and we both saw itl we excitedly showed you of an experience I had in the world of the change in the radar signal to the senior ELINT in 1966-1967 analyst who said that when we got done with the I was one of three analysts assigned to a number one priority signal we would analyze that processing and reporting division We had anew tape first We noted the anomaly on the log collection system feeding us tape-recorded data filed the oscillograph record and went on to at a rate of between Sand 10 tapes per day the next tape We continued to preview tapes 7 days a week We had no automatic processing and almost hated to see nWllber one signals equipment and our job was to find out what sig- because that meant we were getting farther and nals were recorded on those hour-long tapes and farther away from that hot tape analyze the important signals The analysis Months later we ran out of local tape storage results were eventually turned into a hardcopy and moved all but a few dozen tapes to the IRC intelligence report and distributed to users building Still 1110nths later we ran out of Since the data bandwidth and format were dif- paper storage and were forced to get rid of those oscillograph runs We were preparing for a new ferent from previous collectors we had no excollection system because the second system had perience to rely upon and promptly fell behind quit operating The new one went operational in our processing and reporting Some of our and the tape poured in There were new sianals fellow analysts used to stop by daily and joknew recognizers and old #2 signal feli ingly ask us how many pulses we had processed lower in priority We never got to it versus signals or tapes When our collection system went bad it was reThe airman analyst who worked with 111e got out placed with one that automatically sensed when im- of the Air Force and tried to get hired as a portant signals were active and tape-recorded them civilian analyst He was told that analysts The good new was that we didn't have to sort through were not needed so he went back to Ohio and the garbage to find the hot siBJlals The bad news fowtd work there Eventually I lllOVed to a difwas that now every tape would require analysis in ferent branch new analysts were doing what I detail I We were swamped deluged absolutely and had done in the early days of the program The hopelessly buried with good tapes April 76 CRYPTOLOG Page 15 6i0061 EO 3 3b 3 PL 86-36 50 USC 3605 BECKEi old #2' fell lower in priority and reports were pulilished that said we knew virtually nothing about it I haven't forllotten old 12 I To the Editor CRYPTOLOG Thank you for the opportunity to coinment on Mr Greiner's co11111111nt on my short item which appeared in the December 1975 issue In the light of Mr Greiner's conent I regret that my statement about the relationship between collection and processing was puzzling That statement was of course a generalization Accordingly my first obligation is that as a generalization the statement must be accurate and non-exclusive in accordance with the laws of logic If my general statement is discourI did make one more try to get to that tape aging to processors and analysts I regret that Along a out 1969 I was talking to someone as well though I can offer no alternative to about Old 12 and they got excited about the the generalization that the sheer volU118 of colpossibl implications of a change like I was lected SIGINT intercept has got to be greater describing The excitement got to me I took the day off and located the tapes still in stor- in the aggregate than the volume of the same intercept that can be processed per U11it ti'116 age at the IRC building The only way to find I emphasize per unit ti111e because it strikes the tape was to go through each tape and check the int6rcept log for the comment that the air- me that what we face in this context is what Operations Research OR people refer to as the man and I made on the log I spent the day basic inventory model i e looking through tapes in that hot humid storage ro0111 A far cry from SAB-3 where tapes are stO'l ed today I kept think ing It has x· to be t'he next one but no luck The tape I IIUSt eitphosize that I have no OR competency must ha_ve been reused or destroyed as was the and will not be able to defend my point much fate 0£ tapes in those days I did find that beyond this 111 del From the OR standpoint 1110st oi the logs had no analyst comnents we are wont to consider an incoming collection which meant they were never analyzed even where inventory that produces an optimum volume of SOis sJgnals of interest were recorded collection in terms of storage and production That's really sad costs and it has to be a volwae which will But 'how do you measure the value of data un- guarantee 1 m still sorry that processors procesied and not analyzed If nothing bad hap- never catch up with collectors I realize that pens ta vou • lnAc th - Aan it wasn't imnorthis also is a generalization but the converse tant I is even 1110re disastrous to comtemplate than is the generalization -- if only to managers The Chief of Pl Mr William Lutwiniak has spoken often and eloquently to the point that we have to collect a rather large volume to be Things are pretty much the same today Col- in a position to isolate and exploit the relalectors can bury the analysts and computers tively few gems The rationale seems abunwith data We put filters and faucets on them dantly clear I believe this condition exists in an attempt to control them We must restrain because we are the unintended recipients of the them because we don't have the manpower money signals we collect Understandably the intended and 11ay never have the technology to process recipients do not face the problem illustrated and report properly all that can be collected by Mr Greiner inasinuch as their informationhandling systems are inherently and by necesAll through my NSA career I've been trying sity self-serving ln conclusion I appreciate to catch up with the collector I've tried to do my part to improve processing analysis and the old 12 problem It is worrisome but it steps beyond inventory and volume considerations reporting What LCDR Westwood says is that if into the realm of prioritization and that apl succeed we will be in trouble I don't beproaches one of the theses of my article i e lieve that inundation From the maverick point of view the impossiRemember the old 1110tto · it seems to me that the old 12 might well have ble will take a little ti11e been elevated to fl if only temporarily inasmuch as there is no substitute for the on-scene Eugene D Greiner ''judgment calls of the person who happens to be F91 attending the signal -- admittedly another generalization LCDR Fleetwood waa asked 'heth111' he woutd to oOlllll8nt on Mr Greiner 'e LCDR James T Westwood D Chief A732 i ttu to t1ie edito1' He replied as a processing and reponing shop fot c t t A ril 76 CRYPTOLOG Page 16 Etttzt ---------- - - - -' _- _ - --- ·-· ·-- 8R 8RIIT er OHE C• r IPUTER·A OE TRA SCRjPT □ • EO 3 3b 3 PL 86-36 50 USC 3605 f RUSS A lJ CE flTER Al Int1'0duaticm Skeptics have long ago and very convincingly put down the notion that the computer is going to replace the linguist at NSA or indeed anywhere else However a great deal of thought is going into designing ways in which the computer can aid the linguist In this article 1 would like to describe a project currently underway in A Group which is designed to aid transcribers of Russian voice material The project as its na 11e STEPSTONE I implies is destined to be replaced by a more comprehensive system entitled naturally STEPSTONE II Both are part of what A Group regards as REDSTAR Phase I However since interim systems have the tiresome habit of hanging around long after their predicted demise 1 shall not even speak of the future or of greater glories yet to come Instead I shall deal only with the very real present Some time ago it was noticed that highlytrained Russian transcribers were spending at least a measurable portion of their time perfoning functions that were really clerical in nature - - typing proofreading indexing 5 x 8 card files manually searching those 5 x 8 card files etc A wish was born to free transcribers from as many clerical or nonlinguistic functions as possible and at the same time make the fruits of the transcription effort available to other transcribers and to analysts on an online basis The wish eventually became father to the deed in the form of LAYAWAY I a computeraided transcription system The need to generalir e the software for exporting to other transcription problems and the overriding need to provide the intelligence data gleaned from transcription to the intelligence co111111Unity as a whole via regular Prod Data Base PDB methods fathered STEPSTONE I STEPSTONE 1 accordingly is a generalir ed on-line te 'lllinal subsystem designed to aid the voice transcriber in preparing transcripts The software is generalized in t e sense that it can be adapted for use on ot er transcription problems with minimal effort STEPSTONE I is currently operational in A646 and equipment for six more systems in A Group •one in B Group and one in G Group is on or r Interest has been ex ressed in ac uirin • EPSTONE software both by _ _ _ _ _ _ __ and by field sites The term er1111na su sys em implies that data is staged locally for relay to a large data bank in this case the IBM 370 The large data bank computer provided manipulative services and a storage capacity that cannot be expected of a mini-computer Although the main thrust of the STEPSTONE system is data entry temporary storage and retrieval capabilities are provided as well as editing of input and file management Access to the system is controlled by the user's Social Security Number Transcripts prepared on-line are forwarded via magnetic tape to the PDR and thus made generally available The PDB is the primary all-source file for all of A Group Regular procedures exist for extracting infoI'llation from the PDB Briefly they include the following • SELLERS a program which daily selects and distributes data for each organization and for each analyst withi n that organization • TAFBUILD an ad-hoc f i le extract sort and list procedure and • SPECOL a sequential retrieval program that permits retrieval of individual data items using Boolean functions Recently a Cyrillic SPECOL package has been made available to permit retrieval of Cyrillic data A direct link to the IBM-370 via a typed command on the STEPSTONE terminal enables on-line access to the Model 204 retrival system Answer Files The Model 204 system provides the fastest and 1110st complete file-search capability in the building April 76 • CRYPTOLOG Page 23 81 1@RB• i IPOlll l L 08l lll8T 8P61l One of the desirable features of the STEPSTONE system is that for the first time Russian linguists are able to work on a computer in the natural language STEPSTONE terminals are provided with Cyrillic keyboards and Cyrillic display Cyrillic lists can be obtained locally or from the 370 and in general the linguist can remain in the environment where he is most comfortable -- the natural-language environment Advantages and Disadvantaqes Lest I sound too much like Pollyanna I want to hasten to add that there are some disadvantages to a computer system of this type • First there is the matter of computer doim time If you have 48 transcribers say on one computer and the computer goes down you have 48 idle employees Various schemes have been proposed and tried -such as airroring the hardware so that when one piece of equipment goes down its twin steps in and takes over ln addition diagnostic tests are run daily to attempt to isolate potential hardware problems But down time has been a problem in the past is a problem now and will undoubtedly be a problem in the future despite our best efforts • Secondly there is a matter of system tn iining time fortunately this has been minimal -- one day on-line has pretty well sufficed to make the transcriber an effective STEPSTONE teminal user • The third problem for want of a better name can be called Ol d Habits Die Hard Confidence is a fragile thing and slowly built The transcriber may not trust the computer not to do away with that critical piece of transcription or he may fear that the co111puter will be down when he needs it most In the extreme this could cause some duplication -- both entering data into the computer and stickin it in the ri ht-hand drawer On occasion transcribers have been known to retrieve a transcript just after entering it just to be sure it was there This phenomenon tends to disappear with experience on the system Counterbalancing the disadvantages of such a system are the advantages which I see to be priTD rily the following • The abtlity to enter tl'anacripte directl y into ths PDB thus making voice processing a regularlized component of A Group intelligence production • Instantaneous acoessibiZ ity Infonnation that Jones has entered is available to Smith in its entirety immediately There is no need to wait for a grapevine process to get the information to other transcribers and analysts • Perfect readability of entered oopy Asystem editor checks to make sure that dates are valid and that fields are filled in appropriately and that no necessary ones are left blank Erasures and strike-overs do not exist on terminal-prepared copy -- the copy is always clean and readable • Reduced noise Z evel by virtue of the removal of all those clacking typewriters • Economy It is hoped that improved transcription efficiency will lead to an improved ratio in the nufflber of pages transcribed per transcriber Paper costs should be reduced since hard-copy is no longer the medium of data exchange • Rese U'Ch capabiiity With the local system retrieval the transcriber can readily check identifications as he goes along With the Model 204 retrieval capability the transcriber has an im111ediate research-in-depth capability not available previously • Automatic generation of a good deal of what we call managemgnt info1'fflatum instead of via the laborious 111Bnual procedure now used This includes such information as the dates on which the input tapes were received the nUlllber o1 transcription hours required the lag time between transcription and linguist review etc Now that we have considered the why of STEPSTONE including the good and the bad sides let us turn our attention to the how HartAXlre STEPSTONE l is imple111ented on a Digital Equipment Corporation DEC PDPll-40 with 96K words of memory It uses the RSX 11-D operating system The peripherals for each system include • an RP03 disk with 21 5 million words of storage • an RF-11 disk with 246 000 words of storage and • a line printer In addition two ru- 16 magnetic-tape units a card reader and a paper-tape reader and punch will be shared between two systems by means of a unibus switch An additional RP03 disk for each system has been requested to facilitate system backup and recovery in case of disk failure Each PDP-11 can support up to 48 Incoterm SPD 20 20 CRT cathode ray tube intelligent terminals The 20 20 terminal system is a cluster configuration with one controller and 16K of core driving eight teniinals The terminal software is the Incoterm 3270 Emulator a software package that makes the terminals act like IBM 3270 terminals The emulator has been modified in-house to support the Cyrillic alphabet and to translate between ASCII and EBCDIC character sets Fig 1 diagrams the hardware configuration April 76 CRYPTOLOG Page 24 ' Sli18Hfilf ill91ii System ecutive IBM PDP 11-40 3 0 Data Edi ting Entz y Retrievai nie Management 1 1 Cl Printer I J Recove ry Fite 2 Backup Ql E l'- II C MaintenancB • • • RP03 Disk 48 Cul-rent History Ba oHog Fig 1 STEPSTONE I Hardware Fig 2 System FtOliJ The system is designed to support three files which we call the Current File the Backlog File and the History File Briefly all data entered into the system goes into the Current File When the transcript has been reviewed linguistically and analytically it is both sent to the PDB and to the on-line History File It remains in the on line History File 15 days for local access and is then purged In this manner the quantity of data in the STEPSTONE system remains manageable Of course the data is retrievable from the PDB after it is purged from the local STEPSTONE system The Backlog File is for data that hes been in the system for 5 days but has not been quality-controlled The data is transferred to the Backlog File to free up space in the Current File A time permits linguists and analysts can review this data and release it beck to the Current File It then follows the nol'llUll path to the History File and PDB Fig 2 diagrams the files and their relationship to major software coqponents The user's screen is divided into three parts • the system message line for messages from system programs and from theoperator • the system command line to input typed requests to the system and • the data entry area for display of system templates retrieval of data and composition of transcripts April 76 Relationship of files to major software components Fig 3 is a diagram of the user's screen and keyboard The user's keyboard is identical to a Cyrillic English typewriter but has an addi tional 24 function keys which are used to transmit system colllll8llds and to control the screen's cursor Data Entry Each terminal is assigned a workspace in the system All functions -- data entry editing storing retrieving updating and printing -are applied to data in the user's workspace Each user also has available upon request an alternative workspace If he were composing a transcript for example and wanted to interrupt that process to retrieve information on acertain personality he could SAVE the workspace with the partially completed transcript do his retrieval then RESTORE the transcript workspace and proceed from there ln order to prepare a transcript the user requests a blank header from the system and fills in the hend r data and as many screens of text as are required During the process he may page back and forth through his workspace reviewing or correcting the transcript When he is satisfied with the accuracy and completeness of the transcript he uses a function key to call in the system editor The system editor has several important functions It checks to bo sure that the required header fields are present It checks each field to be sure that nUJ11erics are not CRYPTOLOG Page 25 • L SLCICE l Si Gitt S YI TIM MIISAOI LI NI EO 3 3b 3 PL 86-36 50 USC 3605 ll'STIM COMMAND UNI DATA INTI ' AHA SC HIN 100 ACTIVI 0 1 0 • PtOT NUM• 111UIC 0 0 L 0 0 0 T S I·· l l l l lf TPl l l l-1 I I I I Iq IqgIur I 01 ·_•_s•_T_ _•_v__-··_'_'__ l-_ _D_• l ·_··_•_T_l -_'__ 0 l••H I I··· TAI 1 · I SHIPT -i ··z 3 X b C u V B N II 6 H M M SNIFT Ill Dts LAI DIUTI l I NI cu u t IU 10 DILITI CMAI -1 -a INIIIJ LIN I INHIT CNAI Sl'ACI Fig 3 User's screen and keyboard present where alphabetics should be or vice versa In addition it checks certain fields a2ainst each other I 'PoB I l ' ' l l ' '- l '- ' ' ' 'I ' _' ' rne entire transcript 1s edited and 1f any • errors are encountered the transcript is returned to the screen with erroneous parts intensified for easy identification When the transcript is correct the editor releases the record to be added to the on-line ______________________ current File At this point if the transcript were printed •1ia system co11111and it might look There are eight different types of text mod- like the one in Fig 4 Any header field or les which have been designed to be compatible text line may be in either Cyrillic or English as the transcriber chooses with the STRUM field reporting format and the April 76 CRYPTOLOG Page seemzt arenM 26 -- _ l SECRET SPOKE EO 3 3b 3 PL 86-36 50 USC 3605 Fig 4 Sample transcript RetxevaZ As an aid to the preparation of transcripts a sequential retrieval is available which operates against certain of the transcript header fields To use it the transcriber requests a template which is a list of allowable retrieval fields from the system He then fills in the fields and the values he is searching for A hit record is returned to the creen i111111ediately rather than waiting for tbe entire file search to be completed Mmagement Infomiati on As mentioned previously STEPSTONE I also provides certain 1118nagement infol'lll tion to the sponsoring organizations Information extracted from each of the on-line files is sorted and listed on a daily basis for file managers A daily magnetic tape is also created of manageMent information records for historical purposes Follow-on processing of this informatior has not yet been specified It is intended however to dovetail with the A Group CSR Consolidated Scan Record a tape management and accounting information record which is intended to 110nitor intelligence throughput PDB Pl'Ocessinq other lines 1 _ _ _ _ _ _ _ _ _ _ _ 1f are generated from STEPSTONE header fields Once records are in STRUM format they are passed through GENED and FlLB GEN a pair of programs that edit STRUM data for PDB standards and generate the PDB IBM Interface The·STEPSl'ONE t erminal s are interfaced to the 1BM 370 via a typed system command in the terminal comr and line The user will then be able to address Answer Files on the Model 204 retrieval system While in 370 mode the user's tel'lllinal will act just like a Model 204 tel'lllinal and the user must follow 204 protocol Features local to STEPSTONE I 1dli° not be available in 370 mode However since the Incotem is configured with Cyrillic the STEPSTONE terminals will have a Cyrillic capability which the standard 204 terminals do not have Information obtained during STEPSTONE processing will be incorporated into Model 204 Answer Files Sumnati on To swn up I will simply restate the features of the STEPSTONE transcription support s ystem Once a transcript has been released from the STEPSTONE on-line system for PDB processing it must pass through a STRUM-formatting program STRUM is a highly-structured field-reporting format consisting of many different types of lines each with its own individual kinds of data Many of the text lines in the STEPSTONE transcript are al1110st identical in STRUM but April 76 CRYP'l'Ol OG • On-line data entry and editing of transcripts • Local retrieval and file management • Cyrillic data base • Direct access to Model 204 Answer Files on IBM 370 and • Automatic entry of data into POB Page 27 SECit T fJP8HI