Hacker News Reader

Comment by callc | Hacker News Reader

> But when the Kentucky AG’s office was preparing to post their brief against TikTok, whoever was in charge of doing the redaction simply covered the relevant text with black rectangles. Even though you can’t see the text while reading the PDF, you can just use your cursor to select each black section, copy it, and then paste it into another file to read the hidden text.

Incredibly hilarious. Only leet hackers can pull this off though, same as pressing F12 in the browser to hack the mainframe!

View on HN

This happened a few times in 2006. I guess we never learn.

https://news.ycombinator.com/item?id=43698326

It happens constantly, just the other day Meta/Facebook's lawyers redacted graphs by drawing a box over them https://www.theverge.com/news/648893/meta-redacted-documents...

Amateurs. Fortunately for them, the case is lost without their efforts.

This seems to happen somewhat often.

Actually, it is quite weird, I wonder if some bad best-practices have been circulated.

It would be really nice if legal documents were prepared in some sort of standardized markup-like language.

I would guess it's the opposite of "bad best-practices", namely that there are no common best practices. Everyone intuitively understands how black-box redaction works, but the ways inadequate redaction fails is a bit technical and not intuitively obvious, so it's a task that's ripe for ignorant overconfidence.

> same as pressing F12 in the browser to hack the mainframe!

so that's why new mac keyboards did away the entire F keys row?

How is this still happening?

Adobe Acrobat even warns you if you pick black as highlighter color!

Maybe it should instead do actual redaction? Or at least ask if that's what the user wants?

I think that's actually hard... because you have to know what the text is you're redacting to get the spacing for the letters right. You want to redact specific words, but keep the rest of the document the same. This means that you need to know the letters (and their order due to kerning).

The best way to redact a PDF is with a printer, a sharpie marker, and a scanner.

Better to draw boxes, print, then scan. I reckon it will somehow be possible to see through the sharpie -- for example change in shine.

No, it's entirely straight forward to simply remove any curves and glyphs that intersect with a region, even if all text was converted to paths; any vector graphics editor can do it.

PDF documents do not reflow.

It's definitely not as simple as you're suggesting. PDF viewers are not vector graphics editors so they'd have to implement the whole intersection algorithms, and even though PDFs don't reflow they can still have text so now you need to figure out if you delete "is" from "this is hard" you need to calculate where "hard" is.

Definitely doable but hardly straight forward.

I think their point might be that some PDF editors had this functionality for years now ?

I don't think so. Which PDF editors have this feature?

Acrobat (not Reader), Foxit, others I tried. It's not impossible, although the result may vary.

PDFs, inheriting from PostScript, do have the concept of fonts and text strings, which can be edited. They're not just a soup of vector shapes.

Preview in MacOS also warns you that black boxes are not really redacting anything, and provides a tool for it.

Because it's a problem with a 50 year old UX design that we still use today and aren't changing any time soon. WYSIWYG UX is easy to use, because you can simply train the users to manipulate what they see until it visually represents the output they want. This is the predominant way we've trained users for most of the history of GUI software.

But, WYSIWYG on a 2D screen is just an inherently leaky UX abstraction when dealing with 2.5D content, e.g. layers.

The vast vast vast majority of humans are not programmers and do not view the world through the mental models programmers do.

Why SHOULD someone using a PDF reader software intuitively understand that placing a black rectangle over a document and saving it doesn't prevent someone from seeing that document?! Why the fuck does anyone doing a job need to learn a set of unique and non-generalizing mental models just to redact a document!

Imagine if paper had unexpected failure modes!

The problem is not professionals trying to do their damn job. The problem is US. The Programmers. Who consistently make garbage systems using garbage leaky abstractions and absurd assumptions and ignoring decades of user interface research because google released a new UI library written by programmers who haven't interacted with non-programmers in a decade.

Why are WE still doing this?

>Imagine if paper had unexpected failure modes!

It does. If you scribble out what you would want redacted in black ink until it's no longer legible, it's not actually properly redacted for just about the same reasons the PDF wouldn't be

Reply on news.ycombinator.com