Hacker News new | ask | show | jobs
by vintagedave 543 days ago
OP, agreed something better than DocuSign is required. Check out Estonia’s digital signature platform — been around and developed for a couple of decades as the entire country runs on it.

In other words, it’s a solved problem, just little known outside the EU. And it works really well! Second, I suspect there are things to learn from the platform they built and the way digital signatures can be trusted within an entire society. DocuSign is astonishingly bad not because of the tech alone but because of the approach.

(Signature spots are mimicking pen signatures. Why not avoid that, or generate them based on the digital signature, eg as a visual / printable representation then digital signature, the way a QR code carries bits? Make something that survives printing!)

1 comments

>(Signature spots are mimicking pen signatures. Why not avoid that, or generate them based on the digital signature, eg as a visual / printable representation then digital signature, the way a QR code carries bits? Make something that survives printing!)

That's the actual unsolved problem of digital documents.

Regulatory technofetishism requires digital signatures strongly tied to identifiable person (natural or otherwise). eIDAS solves that perfectly, as long you have a digital id issued by your government of residence (the one expecting to connect the document through the signature to a primary key in their database). Nobody except a few governments here and there actually use it and when they do it's mostly a closed system with controlled access anyway.

What people actually need (i.e. how people really work with documents) is a PDF with a signature visible in all PDF viewers, that you can also print and bring to a local government office if they for some reason have in-person process for it. Now in theory, you can have PDF fields and fill them with digital signatures produced by your eIDAS-compatible qualified id. Nobody (almost) ever does that, because your typical workflow doesn't require extracting strong identity from the signed document. The fact that adobe viewer operates on a different chain of trust compared to eIDAS doesn't help.

So what people do is adding signatures as content to the document, than signing it digitally with pdf fields sequentially. It works fine, but you can't add another content-signature on top for obvious reasons and you can't counter sign if the field wasn't present before (and document is locked).

Add: to even start properly solving this, one needs to understand how documents generally work, how (federated) governments work, read the actual law, also understand cryptograrphy and PDF format and after all that make a nice UX. And people who can mostly understand all of above have better things to be busy with, as there is no money in it anyway.