73 Ildikó, Szerényi1 HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) Abstract Introduction: The project detailed in this article covers the indexing of secret-ser- vice themed 20th century archival documents with the help of volunteers and the integration of these data into a database. The State Security Records of the Royal Hungarian Ministry of the Interior form a register, which was created between the two world wars. It contains information on persons considered dangerous to the state, mainly those associated with left-wing movements. In 1945, after the siege of Budapest, these documents were taken to the Soviet Union by the Soviet troops. Method/approach: The 40 archival boxes of original paper documents are now preserved at the Military Archives of the Russian Federation, NAH have pur- chased digital copies of these documents in 2011. The register files are rich in information, so the aim was to transcribe the basic personal data, any other ad- ditional information can be read from the published images. Results: In April 2023 the new database of these 70000 personal records was published on Databases Online2 portal of NAH. Another benefit of this project in Hungary was the huge social activity it generated: 70 volunteers participated in this project, and in this way the cooperation between the Hungarian Society for Family History Research and NAH has been further strengthened. Conclusions/findings: Since the documents were digitised in the Russian Fed- eration, the design of the database structure was the biggest challenge during the project. The registration files often had an additional information sheet, so it was necessary to link the related pages. During the processing work, about 550 register files of prominent persons was discovered in the archival material, some of them are internationally renowned, for instance Josip Broz Tito, Walter Ulbricht, Klaus Mann. Key words: transcribing, volunteers, crowdsourcing, database building, IT solutions 1 Ildikó Szerényi, National Archives of Hungary (NAH), IIAS Member, email: szerenyi.ildiko@mnl.gov.hu. 2 www.adatbazisokonlne.hu. HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 74 COME COSTRUIRE UN DATABASE CON TECNICHE DI CROWDSOURCING – RAPPORTO DI PROGETTO DEGLI ARCHIVI NAZIONALI DELL’UNGHERIA (NAH) Abstract Scopo: Il progetto descritto in dettaglio in questo articolo riguarda l’indicizza- zione di documenti d’archivio del XX secolo a tema servizi segreti con l’aiuto di volontari e l’integrazione di questi dati in un database. I Registri della sicurezza dello Stato del Ministero reale degli interni ungherese formano un registro, che è stato creato tra le due guerre mondiali. Contiene informazioni su persone con- siderate pericolose per lo Stato, principalmente quelle associate a movimenti di sinistra. Nel 1945, dopo l’assedio di Budapest, questi documenti furono portati in Unione Sovietica dalle truppe sovietiche. Metodo/approccio: Le 40 scatole di archivio di documenti cartacei originali sono ora conservate presso gli Archivi militari della Federazione Russa, la NAH ha acquistato copie digitali di questi documenti nel 2011. I file di registro sono ricchi di informazioni, quindi l’obiettivo era quello di trascrivere i dati personali di base, qualsiasi altra informazione aggiuntiva può essere letta dalle immagini pubblicate. Risultati: Ad aprile 2023 il nuovo database di questi 70000 record personali è stato pubblicato sul portale Databases Online della NAH. Un altro vantaggio di questo progetto in Ungheria è stata l’enorme attività sociale che ha generato: 70 volontari hanno partecipato a questo progetto e in questo modo la cooperazione tra la Società ungherese per la ricerca sulla storia della famiglia e la NAH è stata ulteriormente rafforzata. Conclusioni/risultati: Poiché i documenti sono stati di- gitalizzati nella Federazione Russa, la progettazione della struttura del database è stata la sfida più grande durante il progetto. I file di registrazione avevano spesso un foglio informativo aggiuntivo, quindi era necessario collegare le pa- gine correlate. Durante i lavori di elaborazione sono stati scoperti nel materiale d’archivio circa 550 fascicoli di registri di personaggi illustri, alcuni dei quali di fama internazionale, ad esempio Josip Broz Tito, Walter Ulbricht, Klaus Mann. Parole chiave: trascrizione, volontari, crowdsourcing, creazione di database, soluzioni IT HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 75 KAKO ZGRADITI PODATKOVNO BAZO S TEHNIKAMI CROWDSOURCINGA – POROČILO PROJEKTA MADŽARSKEGA NACIONALNEGA ARHIVA (NAH) Izvleček Namen: Projekt, ki je podrobno opisan v tem članku, zajema indeksiranje arhi- vskih dokumentov 20. stoletja na temo tajnih služb s pomočjo prostovoljcev in integracijo teh podatkov v bazo podatkov. Evidence državne varnosti Kraljevega madžarskega ministrstva za notranje zadeve tvorijo register, ki je nastal med obema vojnama. Vsebuje podatke o osebah, ki so za državo nevarne, predvsem tistih, ki so povezane z levičarskimi gibanji. Leta 1945, po obleganju Budimpešte, so sovjetske čete te dokumente odnesle v Sovjetsko zvezo. Metoda/pristop: 40 arhivskih škatel originalnih papirnatih dokumentov je zdaj shranjenih v Vojaškem arhivu Ruske federacije, NAH je leta 2011 kupil digitalne kopije teh dokumentov. Datoteke so bogate z informacijami, zato je bil cilj pre- pisati osnovne osebne podatke, morebitne druge dodatne informacije se lahko razberejo iz objavljenih slik. Rezultati: Aprila 2023 je bila nova baza teh 70.000 osebnih evidenc objavljena na portalu Databases Online NAH. Druga dodana vrednost tega projekta na Madžarskem je bila ogromna družbena aktivnost, ki jo je ustvaril: v tem projektu je sodelovalo 70 prostovoljcev in na ta način se je sodelovanje med Madžarskim društvom za raziskovanje družinske zgodovine in NAH še okrepilo. Zaključki/ugotovitve: Ker so bili dokumenti digitalizirani v Ruski federaciji, je bila zasnova strukture baze podatkov največji izziv med projektom. Registrirane datoteke so pogosto imele dodaten informativni list, zato je bilo potrebno poveza- ti povezane strani. Med obdelavo je bilo v arhivskem gradivu odkritih približno 550 registrskih dosjejev uglednih osebnosti, med njimi so tudi mednarodno po- znana imena, na primer Josip Broz Tito, Walter Ulbricht, Klaus Mann. Ključne besede: transkripcija, prostovoljci, crowdsourcing, izgradnja baz po- datkov, IT rešitve. HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 76 1. INTRODUCTION/PREVIOUS EVENTS In 2021, the National Archives of Hungary published the database of the 1828 national census, which was implemented using artificial intelligence and involved volunteers within the framework of the European Digital Treasures international project. In addition to the publication of the database on the Databases Online portal of the National Archives of Hungary, the social collaboration that made it possible to process the records and publish the name-searchable database proved to be a success story. This positive experience has encouraged the archives to continue working with the volunteer community and to set the goal of creating another database of great interest. The Hungarian Society for Family History Re- search has also become a partner in the second crowdsourcing project, deepening the long-standing and fruitful cooperation between the archives and the society. 2. THE DOCUMENTS USED In the selection of the archival material to be processed, the social use of the ar- chival collection, serving the needs of citizens engaged in family history research played an important role, since social cooperation can be best ensured along this goal. However, the emphasis of the scientific role of the archives was also a priority. The Hungarian language files had been created by the Hungarian police and gen- darmerie between 1920 and 1944 and after the siege of Budapest in 1945, they were taken to the Soviet Union by the Soviet armed forces as booty of war. The 40 archival boxes of original paper documents are now preserved at the Military Archives of the Russian Federation (Российский государственный военный архив – RGVA), previously they were in the possession of the Soviet Secret Ser- vice (Комитет государственной безопасности – KGB). The digitised set of 69,208 files was subsequently inserted by Hungarian archi- vists into the archival system of the Ministry of the Interior’s Archives with the reference code HU-MNL-OL-K 150-VII-9/b. In terms of content, the secret ser- vice-themed documents contain information of persons associated with left-wing movements between the two world wars and as such considered dangerous to the state. This police background, mostly Hungarian language material contains the registration files and data of about 60 000 persons filled in with varying degrees of detail. The personal files also include some 9,000 additional documents. The HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 77 forms, filled in by hand or typewriter, contain details of the person under surveil- lance and often describe the action taken in their case. The richness of the data on persons is outstanding: name, maiden name in case of married women, alias, fa- ther’s and mother’s names, exact date of birth or age from which the year of birth can be calculated, place of birth and residence, religion, occupation, description of physical appearance, distinguishing marks, detailed description of the reason for registration and surveillance, which might be a longer text with additional file numbers that could potentially allow further research in other archival ma- terials, provided they had not been destroyed. A small number of fingerprints or photographs of the person observed may also be found in this archival material purchased from the Russian Federation. Image 1: Additional information to the register file of Jakuboff Mussa The general condition of the documents is satisfactory although earlier they had been transported over long distances in wartime. Some documents are easy to read, some are difficult to decipher. You could also find damaged and torn documents, where data was impossible to unravel. There were also corrected, crossed-out data on the files, and some documents were filled in carelessly and incompletely. The register forms often bear a stamp with the text “Processed in the pocket- book”, which refers to the State Security Pocketbook (commonly called ‘black book’), a printed version of the register with less data, without a detailed case description. Decree No 4400/1931/VII.res. of the Minister of the Interior of 31 July 1931 ordered the publication of this pocketbook. In fact, it was an extracted HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 78 version of the State Security Register and was a periodical publication of the Ministry of the Interior between 1931 and 1944. Even political investigators were only allowed to consult this confidential publication with the permission of their commanding officer. The document was published in small numbers and due to the destruction of many documents during the Second World War, very few sur- viving copies are known to exist. A copy of the Hungarian language pocketbook was provided by PhD Bendegúz Cseh-Gergő, Director General and PhD István Papp, Head of Department of the Historical Archives of the Hungarian State Se- curity. The copy has been digitised in the digitisation workshop of NAH and published as an annex to the database in a double-layer PDF format. The docu- ment has been made text searchable using OCR and can be downloaded from the database information page. 3. IMPLEMENTATION OF THE PROJECT The archives planned to process the large volume of documents by re-engaging the volunteers who helped the work of the archives earlier, so no public recruit- ment was carried out this time. The call and sample pages of the register forms were sent to volunteers from the previous project, and within a week the previous team was almost fully signed up for the new task. Members of the Hungarian Society for Family History Research, with many years of experience in indexing, joined the group, bringing the total number of people who started the processing work to 70 on 1 October 2022. The launch event also served as the closing ceremony for the previous crowd- sourcing project, the 1828 census database. Those who were unable to attend the event in person were able to follow the event live via Zoom, which was later made available on video by the archives. The task faced by the volunteers was to record 9 basic records per master file, which meant transcribing about 450-500 thousand records in total. The archives planned to publish the images of the digitised documents in the database, so that any additional data can be read from the documents and the recorded data could be checked by users. The volunteers of this project were interested in Hungar- ian history and family history research, they already had a certain degree of palaeographical knowledge, had experience with handwritten documents and of HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 79 German, Jewish and Slavic surnames and forenames, and a higher than average level of geographical knowledge to help them read the names of settlements. The candidates were able to test their knowledge and palaeographic skills by using sample pages and to familiarise themselves with the interface of the data entry software before starting the project. The data to be recorded were mostly in the top left-hand corner of the pre-printed register forms. The online accessible Oracle database data entry user interface, which did not require installation, was developed by Helion Ltd. on behalf of NAH. The us- er-friendly interface could be accessed with a unique password. Image 2: Oracle database data entry user interface. The fields of the nine basic personal data to be entered were predominantly free text fields in the database. The data field Religion could be filled in by selecting from a list. Connecting the data fields Place of Birth and Place of Residence to the Geographical Namespace of NAH was an interesting novelty for those doing the indexing work. During the project, the list view provided basic statistical information on the sta- tus of the workflow. The information line (in orange) displayed the following text, which was constantly updated: 12 815 files processed (18.517 %), 56 393 files to be processed (81.483 %). HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 80 Image 3: Screenshot of the database list view filtered by a place name Fegyvernek in Jász-Nagykun-Szolnok County, Hungary. The database offered the images to the volunteers randomly, who had three dif- ferent options available to choose from when the image appeared: a. to fully transcribe the data (green colour code), b. not to transcribe the data (black colour code), c. to skip it or partially transcribe it and classify it as incomplete (blue co- lour code). The records marked with black and blue colour codes later reappeared for another volunteer until someone finalised the solution, i.e. gave a green colour code. What one person couldn’t read, someone else could, so the collaboration was also shown at document level. At the start of the processing work, the indexation task was estimated to take ap- proximately 5 months to complete, but in reality, the data capture was finished much sooner than that; this phase of the project started on 1 October 2022 and ended on 15 November. During the transcription phase, PhD Balázs Kántás, Chief Archivist of the NAH, contributed to a deeper understanding of the period with sci- entific and educational Facebook posts, providing insights into the work of political investigators and the work of the secret services between the two world wars. Volunteers do not work for financial reward, yet the archives would like to express their gratitude for their work in a symbolic way. As in the previous project, prizes were drawn at the end of the programme to reward those who had processed at least 300 registration files. The 1st prize was a 3-month subscription to Ancestry, the 2nd prize was a meal for two at Sir Lancelot’s Medieval Restaurant and the 3rd prize was a 1-year online subscription to a History Magazine called Rubicon. In addition, the top three most diligent volunteers were awarded theatre tickets as a special prize. At the HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 81 end of the programme, all participants received a certificate of appreciation designed by NAH’s graphic designer. The background of the commemorative certificate shows the symbols of the two opposing political forces: a gendarme hat and a Lenin cap. 4. POST-PRODUCTION 4.1. RUSSIAN TRANSLATIONS With the completion of the transcription phase, the post-production phase of the database could begin, which affected both the data content and its structure to a great extent. The Soviet secret service prepared Russian translations of all the files (which were placed alongside the documents) in order to map the Hungarian pop- ulation with communist sentiments as accurately as possible. The Russian partner institution was commissioned to digitise the original Hungarian-language files, but by coincidence some 300 documents translated into Russian have also been digit- ised, these are published in the database as well. The transliteration of the Cyril- lic-lettered files translated into Russian earlier was carried out by Gergő Szikla and András Husvéth, archivists of the Hajdú-Bihar County Archives of NAH. Image 4: A Russian language translation of a registration file HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 82 The data content was checked and corrected using manual and IT methods. The correction of errors detected in the list view was also carried out by 5-6 volun- teers. The sorting of the columns in the list view resulted in similar data being placed underneath each other, which facilitated the detection of incorrectly re- corded information, and other errors. Overall, the volunteers did an excellent job, with very few errors detected in the list view. With such a large amount of data it is not possible to check every detail manually, yet the result is considered to be close to 95% accuracy. 4.2. IT SQL QUERIES The next stage in the workflow was the IT SQL queries, which resulted in much larger lists to be resolved. There were 8,000 cases where the namespace binding was missing for some reason, some of which could be replaced automatically, but there were also manual tasks in this area, e.g. identifying municipalities. The volunteers worked thoroughly and accurately, recording data in many cases in ad- dition to the 9 basic data: in 18,000 cases data were entered in the Comment field resulting in an overcrowded and unclear data field. The solution was to structure and break down the data content of the Comment field. A prefix is a short name or group of letters used in databases to identify fields, followed by a colon, then followed by the data itself, e.g. Nationality: French. In the present case, 27 dif- ferent prefixes were needed to split the data. Some of the prefixes created are the following: Family status, Title, Workplace, Language of document, Address etc.. The use of prefixes has made it possible to display this information in a separate row in the database and to filter on this data. As the use of the Comment field pro- vided each volunteer with the option of recording data on an individual basis, the data recorded in this way is not exhaustive and does not cover all files, so some extra information beyond the basic data is available for some files. Post-processing, data cleaning, sorting and editing was a time-consuming task for the staff of the Digital Services Development Department. This situation will be avoided in the future by creating a form that covers all details. 4.3. PROMINENT PERSONS During the processing work, a large number of prominent persons was discov- ered in the archival material, and the files of about 550 more or less well-known HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 83 persons were revealed. Those persons were considered as prominent, who have an independent Hungarian or foreign language Wikipedia page or are listed in the personal register of the Petőfi Literary Museum in Hungary. Communist politi- cians, writers, artists, doctors and other intellectuals who later became famous, such as Antal Apró, János Kádár, Tibor Déry, Sándor Weöres, Walter Ulbricht, Heinrich Mann, Klaus Mann, John Dos Passos, Berthold Brecht, Martin Anders- en Nexö, Romain Rolland or the Hungarian genealogist Elemér Hanuszik are just a few examples of this category. Image 5: The registration file of Josip Broz, Tito Not so surprisingly, Tito’s registration form can also be found in the database. Some interesting facts, that a simple biography might not reveal about him, can be read from this document. People engaged in secret activities may have used multiple aliases to avoid detection: Josip Broz for example had three aliases: Tito, Brozovics Iván, Tomasek József and he became well known as Tito. According to this document, he was 170-175 cm tall. His physical appearance was described by the agent observing him as follows: dark blonde hair going grey, few hairs, grey- ish blue eyes. The distinguishing marks relate to his behaviour: he smokes a lot, probably in nervousness, he often adjusts his hair with his hands when talking. HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 84 And finally, the reason for registration states, that he was the leader and director of a partisan movement on the territory of the Croatian state, this movement was based on communist principles and was operating under Soviet control. By filtering the occupation, it was also possible to discover lesser-known people in the database. Prominent persons were subsequently entered into the Personal Namespace of NAH; in about 200 cases it was necessary to create a new person- al namespace record. It was a great experience for both the volunteers and the archivists when a famous person appeared in the files, which also added to the attractiveness of the project. 4.4. DATABASE STRUCTURE The design of the database structure was the biggest challenge during the project. As a starting point, the National Archives of Hungary received 69 208 digital image files in .jpg format. It was obvious from the beginning that the database should be based on persons, since the files were basically sorted by persons with- in the archival boxes. The digitization order reflected this structure and the usual digitisation sequence looked like: front page, back page, additional information page(s). At the same time the order of digitization was not always correct, occa- sionally the back page was digitised first. Back pages that did not contain infor- mation were not digitized, it was therefore necessary to find those places where one person’s archival documentation ended and another one’s began in order to link the related information. No logical element was observed in this sequence; a person’s archival documentation could consist of one image, two images, or, in special cases, up to six images. To mark the related images, Zsolt Záros software developer created an image linking program, which displayed the thumbnails of the digitised documents in a list view, where the related images could be linked by clicking on them. When several single items followed each other there was nothing to do, in many cases the front page and back page needed to be connected: the decision was made by checking the physical characteristics of the pages. This method proved to be very efficient as the visual approach is much faster than reading, as opening the image can be avoided. How to link pages based on physical characteristics? It is illus- trated by the following example. There is a damp stain in the lower left corner and a tear in the upper right corner of the images presented here, these character- HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 85 istics are clearly visible on the back of the document in reversed state. If in doubt, the image can be opened and the text, which usually continues the back can be checked by reading. Human intelligence has quickly mastered the new working method, and the 70000 images could be browsed effectively. In some cases, it was impossible to avoid reading the text of the documents, as this was the only way to establish the link with certainty. Image 6: Front side and back side of the same document, where the physical characteristics are clearly observable. The links created by ‘clicking’ on the image thumbnails were distinguished by a variable computer-generated colour coding. To ensure that the links were unique- ly distinguishable across the nearly 70,000 image files, the software used so many different shades of colour that, when placed side by side, the human eye would not be able to distinguish. If you were to express this range of colours in numbers, you would be talking about a 20-digit numerical series. Using IT language, it is a variable computer-generated colour coding system. HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 86 Image 7: The image linking programme in operation. The physical charac- teristics are clearly visible: the backs mirror the appearance of the fronts: there are tears, stains or paperclip marks. The database was published in April 2023, and it was visited by 16433 users in that year. The National Archives of Hungary’s staff involved in the project: Zoltán Sza- tucsek (director), Zsolt Bánki (head of department), Ildikó Szerényi (project lead- er), Zsolt Záros (IT software developer) László Debreczeni (IT specialist), Balázs Kántás (research fellow), István Németh (research fellow, head of department). HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI 87 REFERENCES Bánki, Z., Szatucsek, Z., Záros, Z. (2023). A névterek mint a hiteles tudás forrá- sai. TMT, 70(4), 472–482. Retrieved at https://doi.org/10.3311/tmt.13273 (ac- cessed at February 2024). Tomić, M., Grzunov, L., Dragija Ivanović, M. (2021). Crowdsourcing transcription of historical manuscripts: Citizen science as a force of revealing historical evidence from Croatian Glagolitic manuscripts. Education for Information, 37(4), 443–464. Retrieved at https://doi.org/10.3233/efi-211555 (accessed at February 2024). Simperl, E. (2015). How to Use Crowdsourcing Effectively: Guidelines and Examples, Liber Quarterly. Retrived at https://liberquarterly.eu/article/ view/10758/11830 (accessed at February 2024). Szerényi, I. (2023). A csendőrkarton-projekt. LEVÉLTÁRI SZEMLE 73(2), 15–26. Summary This article discusses a project focused on indexing secret-service themed 20th-century archival records from the Royal Hungarian Ministry of the Inte- rior. These records, preserved in Russia, detail individuals seen as state threats between the world wars, with a significant number linked to left-wing movements. In 2011, the Hungarian National Archives acquired digital copies of the 40 archi- val boxes, which were transcribed with the help of 70 volunteers. In April 2023, a database containing 70,000 personal records was published on the NAH‘s Da- tabases Online portal. Notably, 550 records of prominent figures, including Josip Broz Tito and Klaus Mann, were identified. This initiative also fostered collabo- ration within Hungary’s historical research community. Typology: 1.04 Professional Article HOW TO BUILD A DATABASE WITH CROWDSOURCING TECHNIQUES – PROJECT REPORT OF THE NATIONAL ARCHIVES OF HUNGARY (NAH) ILDIKÓ, SZERÉNYI