I EO 3 3b 3 PL 86-36 50 USC 3605 TSP EGRET CBCE£J 'il0110W ™ m - ' srEPSTONE AND A NON-RUSSIAN VOIC PROBLEM - COLLECTION-SUPPORT T A IS NOT FOR EVERYONE 1 I • 7 I ••• 9 - 11 • 14 I• • • • • • •• • •• I • 14 15 • 17 19 •• 20 11118 B881fl llUT fll lilftillUEJ flODEhORB HKIERME f8P Sl EREI I - - - - - - - - - PL 86-36 50 USC 3605 s •mtrz rrrnz 1tw eeenc ·a a I I f • Iii II il U I i g I a I - X fC I I I Approved for Release by NSA on 12-02-2021 MOR Case# 111019 a fa$ --- · JNon 'P8P 8DORH'f Published Monthly by Pl Techniques and StandaTds for the Personnel of Operations VOL V NO 2 FEBRUARY 1 978 T8P 8D8ttE • - Responsive I PL 86-36 50 USC 3605 I r TEPS TONE ND _·A • NON - RUSSIAN· VOICE PROBLIEIIII S quar P eg Roun H° e • • • • • f the Api-il 1976 issue of CRYPTOLOG t their specific intelligence-nlpduction problems lof A633 introduced and ln our ease we took to heaTt lremarks explained the_STEPSTON I compute sub about adapting STEPSTONE to other transcription system designed to aid the voice transcriber in efforts and we began to look for the most profpreparing transcripts As she pointed out the itable way of using STEPSTONE use of computers can help to free transcribers Naturally there were several logical and from many clerical and unnecessary nonlinguistic reasonable restraints that set bounds for any tasks But to us one of her most significant 11 statements was •• STEPSTONE can be adapted for · fine tuning that we might do to STEPSTONE or •its procedures ·use on other i e non-RussianJ transcription problems with minimal effort 11 While one might eour changes could not interfere with the wonder about the 81110 Dlt of that effort we · normal operation of the overall STEPSTONE in A323 foWld that the adaptation of the system STEPSTONE I concept to our specific requireeThe results of our 01tuning could not in ments was an absolute necessity any way conflict with other existing A3 When STEPSTONE I was first introduced to A Group or Agency computer programs or our non-Russian problem we were somewhat unnecessarily duplicate existing data surprised to see how easy it was to use entry or retrieval programs STEPSTONE data entry and retrieval procedures eAny changes would have to be necessary We had been awaiting a flexible on-line traneasily understood and consistent with scription capabability and after several the design and format of STEPSTONE and disappointments and failures with forerunners of machine-assisted transcription we were • It could not cost the Agency anything ready for something that would work well beyond the normal planned cost in time and money for STEPS'fONE operations To begin with we noted that some people who were using computer programs and software were What Is STEPSTONE What Did We E rpect missing vast opportunities to improve their operations beca4se they_ ver really explored As designed STEPSTONE l is an on-line how computer technology could be applied to terminal subsystem to aid voice transcribers February 78 • CRYPTOLOG Page If QR iECDIJlf lcJPJIBR1t l o 3 3b 3 86-36 50 USC 3605 Ill••• • • f6P S @Bfif tiMBBK • • in pr paring trans - 'E'fpt s with its main thrust being the entry of th• transcripts STEPSTONE also provides the ftansc be r with temporary storage and retrievil Capabil ties The transcribbr enters the• 4ata at- a te J'lllinal near his or he1 transcriptioR positr and the data is relay d to a large 4ata ank •-- t e IBM 370 This tiata then beG omes plrt of the PROD Data Base ' PDB where procedu1es • g • ELLERS and SPECOL exist • fer extnctiha entii e transcrip s or key items withki trallScriptl In the future trans rlbers and ana1 sts Wi l gain access to the Modpl 204 retrievaf system X swer Files•by using theit STEPSTQNE tenn1nals • Tl # above feat res made S'l EPSTO i•sound v sy attractive and wp lmpatientl_y await 4 its • arriwl but whe n twas mea sured aga• st our needs we were qu ickly disapp•in 9d •1• were glad ihat we would be able to create o-i trans ripts on Ifne•-- no more•typewriter produced paper tianlcripts BuJ STEPSTo1i Eis temporary storage aqd limited retrieval cap bilit ' presentcd•a ujor stwablmg block • • After numerous qaes•ions we cam to reali e • that i hat we 'anted • and had expe ted was not' within the scope of'sTEPSTONE's cwrrent purpose STEPSfONE I is mere 3 a portion o the larger • IQOre lexible REOSTAA system that some day will iive us th de9' i red computer s pport • • • m1na es e processing o super lu items and fTees the STAR to perfOl lll hii or her analyst reporter duti s Considering the operational Jrocessi g con cept practiced there are three transcription and reporting factors which have •direct bearing on our approach to STEPSTONE l • · _ 1_____ _ • Because of the natu of our targets i e a random mix of subjects it is not feasible for one STAR to be 't sked with tr nscribin • solel one cate or •of sub·ects veryone processes any su Ject regardless of his or her individual orting respo sibilities • Because everyon may process anything it is more importint to readily identify the subject matte ef the transcript than to identify the transcriber I our W t o teams each compatatively f Because of the li ited number of person assigned to our tra cription effort app i-' 11ately 30 our brancll relies upon the progres9'ive processing concept whereby each person in t'he voice effort must be•a scanner transcriber analyst and reportef -- a STAR In addition to these basic tasks e ch STAR is assi ed re o tin res onsibilities 1--- -r IJL JI L il llif lili 1 1 lUi--transcription OU1' J equirement e realized that our requirements•and expectatians could not be met by the proce8ures and form4ts as presente4 in STEPSTONE t exac ly are th se requirements • PL 86-36 50 USC 3605 as prov1 e are • • • • • are The STAR system not only allows us to approach maximum utility of assigned persoMel but also rovides us wit he is a so an analyst reporter he whether or not a articular i e Modification of the STEPSTONE Forrrnt Because STEPSTONE l was designed for an organization involved solely in the transcription - - - - - - - - - - - - - - - - - - - - - - - • • process and made no specific pr isions _for February 78 • CRYPTOLOG Page 2 If OP 81 18RMlf tilllBRh EO 3 3b 3 PL 86-36 50 USC 3605 or e• @BM • • • i retrieval of transcripts by subject matter we were initially confronted with three options A we could tailor our operations to conform with the given STEPSTONE format B we could enter transcripts using the STEPSTONE format as presented and retrieve them by subject matter using SELLERS SPECOL and the Model-204 system or C we could try to adapt the STEPSTONE format • to meet our specific_ retrieval needs Option A was dismissed as wholly unaccept able • Option B was rejected because at the time STEPSTONE was introduced to our operations SELLiRS SPECOL and Model-204 were relatively inaccessible to the STARs Left with option C we set out to identify those portions of the STePSTONE format templates which lend themselves to modification • We should stress however that _ fully realized that any modifications to tn e STEPSTONE format or the use of STEPSTONE fctt purposes other than those originally intended was solely an interim adjustment until the arrival of fully accessible convenient aata retrieval programs • One of the basic requi ements in our transcription process is th •checking of transcripts' for accuracy In STEP5TONE the checker veri- fies that a particula-1' t int h oe ___ examined for accura yf • • A Our first at tempts to perfoTIII quality con- • trol of the tl llnscripts brought home the need • to differentJ ate between transcripts produced by the two -teams Since both categories of informatioh are I I md because of the large volume of transcripts produced by both teams we found that in order to quality-control QC one team's transcripts that team had to wade through nUJ11erous transcripts prepared by the other te81D f Systemat o Adaptati on of STEPSTONE Format Realizing that the STEPSTONE format could be modified there was no modification of software we began to look for ways to retrieve transcripts bv sub ect matter· ·and location of activity I February 78 CRYPTOLOG Page 3 •or 818 1• IJ IRIIJ EO 3 3b 3 PL 86-36 50 USC 3605 PL 86-36 50 USC 3605 iliRIT l Jftll•t 'IlOP · • 1 I '•' I I I The STEPSTONE system provides three transcript files the Current File CU File the Backlog File BK File and the Hi story File HI File As ' l T' --descri bed the arrange111ent in ' ' ' ' I While our 1110difie1 -- - rrov i ed us with much greater fle i ty rettievi g transcripts we were kill not satisf i ed 1 Jla t all possibilities had 1been exploited •• closer examination of 1t e M'l m Manual we noted that • Briefly all data entered into the system goes into the Current File When the transcript has been reviewed linguistically and analytically that is subjected to QC it is sent to the PROD Data Base 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 other words as the system is designed a transcript should remain on-line for the one to two days required for its transcription and QC plus the 15 days it spends in the on-line History File -- a total of 16 or 17 days The ability to maintain transcripts on-line for 16 illl l ili l -1 lil lll l lil i ilwil W li W MW MM IIMikl -- I or 17 days is apparently more than adequate for • a shop dealing solely with transcription However for our operation dealing with both transcription and anal sis a loner on-line c _a abilit --- '- ' '' ' '71i i ' When tiie ·iiodified ilo _ _ _ _ _ _ _ _ are used together analys scan eas y an quickly detemine the basic content and subject matter of each transcript Some hypothetical examples Our problem therefore was how to deal with the built-in on-line limitations of the system Our first objective was to prevent selected transcripts from being automatically purged after 16-17 days in the system our approach to solving the on-line problem was constrained by the fact that we are obligated to 9 all transcripts on a timely basis to allow February 78 CRYPTOLOG Page WOP SlilElllil P UlfPR t 4 8 '1'8P 66001Jlf wa IDllts · PL 86-36 50 USC 3605 • •• their entry into the PDB As stat above once a transcript has been QC-ed it is automatically transferred to the History File where it has a 1S-day on-line life expectancy We realized therefore that we would have to systematically remove selected transcripts from the Histo Fil · a transcr _pt 1 a 1•r•e•a d•y• i n- t h•e H i s•t•o•r•y F1 ·1 e--w i- l ll be autoutically transferred back to the Cu e F W out a transcript will remain in the CU F i ll•e- o r 30 days and then be transferred automatically to the BK File where it may stay for an indefinite period of time depending upon the disk space available Thus the first two steps we devised for the process of maintaining selected transcripts on-line for as long as necessary were to I EO 3 3b 3 -· • • • At the ed4 of each work day checkKs carry out both tJie QC and the Zero-QC proces-ses As of this writiR selected transcripts have been 111aintain on'-line for over 14 months i periodi- • cally e re_vi-ew the Backlog File and delete those trans r1pts that are no longer necessary Twice•duHng this 14-1110nth period our•Computer • Support Br h notified us that our pa e count in the•Ba klog File was nearing file Cfpacity and wd were•asked to reduce the number of tran- • scripts • e complied by deleting those tran• scri pts relat to completed or nearly•completed eQC transcripts on a daily basis allowing projec s products for which there was o lon1er their entry into the PDB and their • a fieed felr on-line maintenance • • transfer to the History File and • after transcripts have been in the Histoey• sJmnai-y • File approximately 13 days since they • • It lijlllllllary we recogniz e STEPSTONE a nd its disappear after 1S days call-up thoJe concepcs as invaluable transcription tools transcripts that are to be kept on-line Fo Pmany it serves well as is The or i ginal ST PSJONE concept of machine-formattt tranSfription has been left virtuall inta b odi ------------ - P - ll 0 y an unc ec transcr1pt w 1 __ i --r· o• e 1 ems not nee e your peculiar systeM of File into the BK File• operations the STAR system have been modified With the recent developments in other computer programs the experience gained from our exposure - ' --1 to STEPSTONE and the assistance of our Computer This step ensures tat a ero upport Branch we as transcribers and analysts -e transcript will remain in the CU File for are more confident that computer technology an additional 30 days before it is transferred offers much to the voice effort 1 i_ t e B • For us the immediate advantages of the modified STEPSTONE I were • a transcription retrieval capability with better than 6 months of selected material on-line • a systematization and standardization of both the transcription process and the analysis reporting effort • quick convenient and reliable retrieval by subject at the STAR' s immediate disposal and • a significant reduction in the amount of time required to perform quality control -------------------- Perhaps more importantly by being able to group tog ther transcripts related to R specific subject over a given period of t i me through the modified use of STEPSTONE we are now more confident that we are producing SIGINT product based upon all available voice ma terial In this way we feel we are in a better posit ion to comply with the ultimate objectives regarding finished SIGINT as outlined in V-129-175 Intelligence Conclusions in SIGINT Product t February 78 • CRYPTOLOG • Page 5 TOP illiBlllll 6iHBlll4 • filQP BD@DI• f fRIDIM February 78 CRYPTOLOG Page 6 18P 8llf1Rll• Bl'IIBII -- - -·- EO 3 3b 3 PL 86-36 50 USC 3605 I PL 86-36 50 USC 3605 l · COLLECTION·SUPPO RT TA IS NOT FOR EVERYONE · GR F tis often aic that the traffic • analyst working in a collection• operations rea•isn't doing traffiG • analysis t lis•is because very linle in-depth anaJ ysi is being performed Be ter termed the work could be called cQllection-support 'lmal ysis It is a spacial -type of trafric enalysis for which no• all traffic anal 'sts•are well-suited In fact it is so diffuent that s0111e might su4gest it be considered as a unique cateer specialty • perhaps different professional disciplines · within the operations organization The tasking of mission to an intercept station is always changing This causes the CSA to familiarize himself constantly with new targets with no assurances that those targets will remain on mission for long Often the mission assigned to the station will be bits and pieces of several entities That means that the CSA will also need to familiarize himself with many related but non-mission entities that may happen to be intercepted Thus the CSA must have a good information and recall system ecruitm nt of in ivid ls with an aptitude i e working aids Technical SIGINT Reports fop collection-suppo t an41ysis in my opinion sh 1Uld be better defi ned Too often the re • and or a good memory This is one reason why set ms to have been a-willingness to accept any the CSA may frequently complain about the quali tr ffic analyst willi'hg to work rotating shifts ty and timeliness of TEXTA TSRs and other or perhaps willing to'take•a field-station tour working aids The experienced CSA soon learns that the information seemingly critical to him of -duty Unfortunately tl'le skills needed to•be a good traffic ·analy t do not necessarily is not readily forthcoming Often it requites him to establish his own variety of working tr nsfer into making a go collection support aids which he invariably feels the parent oran lyst CSA A certain 111ental persuasion is ganization should have produced Actually needed in order to be a competent CSA TA in a the problem is simply a case of a difference collection-support an lysis environment This in perspective and each has his own view of he person plays a vital ole i the collection priorities cy le He should be he oiL and grease in the ingut output and fe back rocess of the The CSA 1 s perspective is closer to that 0£ collection cycle Th s artitle will identify the collector whereas the perspective of the th• role of the colle ion-s pport analyst analyst in the parent organization is closer and attempt to identif t what his mental persua- to that of the output or reporting phase of si n should be toward nis job the collection cycle The CSA is caught in between the two conflictJng parties As amaiiug •The CSA is the tral c anai st who as always the tasking organization never with the collectors of • at seems to give necessary acquisition data useful t · • area such as in to the collector in quickly identifying and The CSA necessarily re 1e copying a new target Fore le the colrmation as documented from the in-depth lector mi ht ask whethe analysis performed by traffic analysts in he tasking Organization One of his main functions is to serve as a liaison between the t would collectors and the tasking organization and 0 0 February 78 • CRYPTOLOG • Page 7 EO 3 3b 3 PL 86-36 50 USC 3605 tiJQfJFIIHilfPfls1aL seem fundamental to think collector when assigning a target by giving the collector the best acquisition information possible this is often not done The CSA is then left to grope around trying to fill the voids in the acquisition data The irreducible aspect of thinking collector is concerning ourselves with the needs of others and keeping in mind that collection is the starting point of the collection cycle Viewing it otherwise can only cause friction in this cycle This means that the technical in-depth analysis feedback must be given as high a priority as other aspects of a job Then in times of crisis when collection is most essential objectives will aore likely be satisfied As in faming raising a good product requires a good amount of cultivation It seems today that NSA is too often content with the present yield of a product with little or no cultivation The collection-support analyst attempts to fill the vacuum created by the inward reflection of the tasking organization The CSA organizes the technical information in order to • minimize the acquisition time • maximize intercept and the meeting of collection objectives • reduce the learning ti lie needed by the collector to ensure that he will quickly familiarize himself with t e target • document his efforts since people are constantly rotating or transferring and • give feedback through field-station T In turn he must c01111UJ1icate to the tasking organization just what the acquisition and copy problems are The CSA a ist concern himself with next time so that collection may proceed routinely whenever a crisis presents itself Secondly collectors often do not copy the sue target on consecutive days Thus the CSA may need to serve as the focal point of an in-house dialog It is he who should ensure the follow-up on significant intercept The CSA most often concerns himself with the present and the future and this requires a lot of intuitive and innovative action The results of these actions will necessarily reflect his ability to understand his collection environment In addition to understanding the collectorts job he must be able to relate to the person who is working as a collector This involves an understanding of such things as collection techniques signals analysis computer applications language and reporting The better the CSA understands these related disciplines the greater the service he can provide to the operations organization Often the dialog between the collector and the CSA sounds more like confrontation than cooperation And when placed in CSA jobs many traffic analysts feel uncomfortable and some- what frightened at being cajoled In a way it is a sort of game The collector tests the CSA's credentials so that he can evaluate the CSA's use to him Conversely the CSA must test the collector so as to ascertain the collector's proficiency and attitudes Obviously people vary in their inclination to doing various jobs Whereas one collector may love to copy new mission another may be more content with copying a target with which he is f11111iliar But the fact is the CSA and the collector must work as a te1111 and must conaunicate with one another If they do not build a close working relationship the entire mission is likely to suffer If nothing else the CSA must be able to display sincere empathy with the collector's problellS and seek ways to alleviate many of the technical ones The CSA must show restraint and diplomacy toward the collector's work Who in his career has not seen the instance when a collector has laboriously copied a target for 2 hours and then an undiploaatic CSA glances at the traffic and tosses it illmediately in the burn bag When that happens is it any wonder that the next time that intercept is needed the copy may not be forthcoming The CSA must not be afraid of taking charge of a situation Often he must make a decision on whether to continue copying a new target or to have it dropped In the midst of a crisis he uy need to weigh the factors of maintaining mission discipline as given in the tasking or ignoring some mission in order to copy a hot item It is in times of crisis that new targets appear and the decision of whether or not to copy a target may be critical to the gathering ·of needed intelligence information Conversely the improper use of resources by ignoring mission tasking may also cause problems It will probably be the CSA who can best detenine if one target should be sacrificed to copy another -- and then to detel llline which one to drop In these days of optimum tasking some target will have to fall off the tasking ladder if an 1massigned target is copied Therefore the CSA must know where mission discipline must be maintained and where it can be sacrificed The CSA should be able to reco11Rend the direction and help steer coverage Collection-support analysis is where traffic analysis begins and it can often be exciting The person IIUSt enjoy working with the unknown and be willing to deal with the frustrations of a constantly changing mission environment It is not a place for people who do not like change or the unexpected The CSA is an important part of the collection cycle but he is only as good as the support he receives or can 1DUster Since re110te oper ation is becoming an increasing endeavor in the NSA effort it would be beneficial for those who can do CSA work to gain such experience so that they can gain a better perspective of the collection environment In turn they will increase their value to NSA February 78 • CRYPTOLOG • Page GQlll IBDtJII Jl 8 IUllltl rs ss s s 7111 I 1111 I L - -- - ·---- ·· 1 I I Non - Responsiv _ _ _ _ l _ -· - Non - Responsive I EPOL 3 • 3b 3 J 86-36 50 USC 3605 • I niliOP 8R8B BII IJHllll t USC 36 Repl int of A l1 Infoz nrz' Technical Nots 001- February 78 · 18P CRYPTOLOG Page 11 ononn• ttn111n1t EO 3 3b 3 PL 86-36 50 USC 3605 TOR iliRliT UHBB t • February 78 CRYPTOLOG _ __--- __ Page 98P 8B@R 9 Bflllltk _ ___ ___ 12 - EO 3 3b 3 - - - - - ' ' -·•· -- _ PL 86-36 50 USC 3605 • TWP OIODIT WJIDD• February 78 CRYPTOLOG Page 13 1- '- J -' -'1 t M • - - - - - - -- -- -- -- -- -- -- -- - - - - ------ - - · - -·-I Jnn - - ··- - - - -- - ra - - · Non - Responsive I Non - Resoonsi ve I Non - Responsive •1 •- - - - ' · 1Non - Responsive I Non - Responsive