Registro De Fatos Históricos

On the practical side of recording history

Most people think you just write things down and you are done. That is not how it works. Registro de fatos históricos is a discipline that lives or dies on the details you choose to capture and the ones you leave out. I have seen too many projects collapse because someone treated documentation as an afterthought instead of the foundation. Here is how the process actually works when you are dealing with real sources, not textbook examples.

A method for registro de fatos históricos

Start by collecting primary sources before you form any narrative. I learned this the hard way about five years ago when I was working on a regional archive project for a municipality in the interior of São Paulo state. The municipal council had been dissolved and reorganized three times between 1964 and 1985, and most of the original records were sitting in a warehouse that had flooded during the 2012 rains. Nobody had cataloged anything because they assumed the water had destroyed everything. What actually happened is that only the top shelf was affected. The lower shelves, which held the bound meeting logs from the original council sessions, were completely fine. The problem was that the warehouse manager had relocated boxes based on visible condition, not on archival order, so the provenance chain was broken. I spent three weeks just rebuilding the chain of custody by cross-referencing ink types, paper stock, and stamp impressions with known municipal printer contracts from that period. Once I had the provenance intact, the actual content extraction took another six weeks using a portable document scanner and a standardized metadata template.

The metadata template is critical. I use a modified version of the Dublin Core standard but strip it down to about fourteen fields that actually matter for local government documentation. Anything beyond that is academic padding that slows you down without adding value. The fields I keep are: identifier, title, creator, date, date_created, type, format, language, location, source_reference, provenance_chain, subject_keywords, scope_note, and related_records.

Technical execution

Scanning itself is straightforward. Use a minimum resolution of 300 DPI for text documents and 600 DPI for anything with photographs or handwritten notes. Save in TIFF format for archival preservation and generate a separate PDF/A for public access. PDF/A is important because it embeds the fonts and metadata into the file itself, so it remains readable decades later without depending on external software installations. OCR, or optical character recognition, should be run on every scanned document. Tesseract is the standard free tool and it handles Portuguese reasonably well if you train it with the ABILIO language model or use the already-trained por model. For handwritten documents, Tesseract will not cut it. You need something like Transkribus or a dedicated handwriting recognition service. I found that Transkribus improved its accuracy significantly between 2021 and 2023 for Portuguese handwriting from the mid-twentieth century, but it still requires manual correction on about 15 to 20 percent of lines depending on the quality of the original script.

Indexing comes after scanning and OCR. Build an index that links each document to its contextual metadata, but do not over-index. A common mistake is creating too many keyword fields with overlapping terms. If you tag a document with both "câmara municipal" and " Câmara Legislativa Municipal" you are creating duplication that makes search results unreliable. Use a controlled vocabulary. The Local Government Thesaurus from Brazil's Institute of Information Science is a reasonable starting point, though it has gaps for very local terminology.

Common pitfalls

The biggest mistake people make is assuming that once documents are digitized, the work is complete. Digitization is the easy part. The hard part is verification. A scanned image does not prove that a document is authentic. It proves that a document existed at a certain point in time and was reproduced in a certain way. Forgers have produced convincing fake municipal records before, and OCR makes their forgeries searchable, which gives them a false appearance of legitimacy. Always verify critical documents against at least one independent source when possible. Another frequent error is neglecting the negative space. What is missing from a record is often as important as what is present. If a municipal council meeting log from 1971 suddenly stops recording discussions about a particular budget item, that silence tells you something. People who skip the step of noting absences and omissions end up producing accounts that look complete but are actually misleading.

There is also the problem of temporal bias in source survival. Government records tend to survive in better condition than private correspondence or NGO documents. When you build a historical account primarily from official records, you are hearing mostly from people who held power at the time. This does not mean you should discard official records, but you should acknowledge the bias explicitly in your scope notes and actively seek out alternative sources to balance the picture.

Tools and workflow

For small-scale projects, the workflow I recommend is: acquire physical documents, photograph or scan them, run OCR, extract metadata into a CSV file, import into a simple database like SQLite, and generate an access portal. Do not invest in expensive repository software until your collection exceeds a few thousand documents and your team has more than two people working on it. SQLite with a well-designed schema handles thousands of records without breaking a sweat. If you need a download link for a metadata template, I can point you toward open-source options. The International Council on Archives publishes template files in both Excel and CSV format that are freely available. There is also the AtoM (Access to Memory) platform, which is free and open-source, though it requires a server environment and some technical setup. For a small project with limited resources, AtoM might be overkill. It is designed for institutions that need to manage multiple collections simultaneously with complex finding aids and publication features.

When registro de fatos históricos fails

There are situations where this entire approach breaks down. If the original documents have been deliberately destroyed, no amount of digitization methodology will recover them. If a community has oral traditions that were never committed to writing, the primary record-keeping practice of the project will exclude those voices by definition. If the political environment is hostile to transparency, you may never gain access to the relevant archives at all. In these cases, the workaround is to pivot to supplementary methods. Oral history interviews recorded with proper consent and archival-quality audio equipment can serve as primary sources themselves, though they require different ethical protocols. Sometimes newspapers, church records, or land registries contain fragments that fill gaps left by missing governmental documents. The key is to be honest about what you do not have rather than pretending the record is complete.

Recording historical facts is not a neutral act. Every choice about what to document, how to categorize it, and which sources to prioritize is a political decision. The best practitioners are aware of this and make their decisions explicit rather than hiding behind the appearance of objectivity.