Timothy Lee has conducted a study of improper redaction in PACER, the US court records system. Sensitive information like social security numbers are redacted in these records, but sometimes the redaction is accomplished by drawing a black box over the text in the PDF; the text is still present in the PDF file, it’s just not displayed, and it’s easy to recover. Out of 1.8m PACER documents, there were ~2000 documents with redaction rectangles. Examining them by hand revealed 194 documents with failed redactions.
ah, the good old blackout doesn’t work with pdf, eh? and another story. this is why lawyers should not run governments: incompetent in the small, incompetent in the large.
We discovered this classified information by opening the slide “Award Actions Trend Data”, right clicking over the chart titled “Award $,” choosing “chart object,” and then clicking “open” to see the hard figures Everett used to create the graph.
heh. i thought that presentation was interesting for its geeky content. but now i guess i will take a closer look at the pdf.
Joost may have accidentally leaked 3 months’ worth of deal plans through hidden data in a PDF.