[smueller1234]

I keep reading the "two years warning" notion on HN. While that might be technically correct, the real problem was that nobody UNDERSTOOD what GDPR meant (including the legislators) and so to this day, its practical implementation will to no small part depend on the iterative conclusions and learning various implementors (eg. companies) made in an arduous process since.

In other words, the first to think they were GDPR compliant might have had to redo a ton of work to adjust to more recent interpretations.

And let's not forget, for large orgs with complex infrastructure, this is a behemoth of an effort. There's been year long projects in the two large tech companies I've had insight to since.

And while I'm at it, let me comment on the frequently expressed notion of "if you've respected your users in the past, you'll be fine!". Just to pick one counter argument: the right to be forgotten. That can only be implemented thoroughly and in the way the users expect it to work (ie. delete everything but what you're legally required to retain) by finding a way to connect all user data so you know what to drop if need be. That is exactly the kind of action that's caused public outrage at big tech to begin with and it's not only potentially a huge effort, it also increases risk of abuse.

This all being said, I still think GDPR is a good idea at least in principle. And believe it or not, while everyone around me is really of compliance work, GDPR seems widely considered a good idea in principle across engineering in big tech.

[oldcynic]

> the real problem was that nobody UNDERSTOOD what GDPR meant (including the legislators

There we have to disagree. It's not like this is something new and untried.

GDPR is a development from long-standing, and now very well understood, Data Protection. The legislation seems mainly intended to modernise some of the definitions and scope (eg adding biometrics to PII), catch some newer practices, and make very plain and explicit that it doesn't just apply to EU companies.

In 1996 and 97 in the run up to the 1998 Data Protection Directive I recall a couple of common confusions and misunderstandings. Nothing like the ridiculously poor and simply incorrect reporting we have for this.

Any large org should have been fully compliant with DPA for years. They have to add extra mechanisms for explicit opt-in or deletion and get a little less time to retrieve full data and can't charge. That doesn't seem to need a "behemoth of effort", but not to say it's necessarily entirely trivial.

In other words they survived DPA with no apparent effect, yet it's >80% of GDPR with the same definitions. No one should be iteratively fumbling toward an unclear target at all. Even reading the UK ICO's old guide to 1998 Data Protection from a few years ago gets you most of the way there including understanding personal data.

[DanBC]

But there are not massive differences between the laws we've had for many years - for a UK example PECR and DPA implement EU regulations and contain many of the same principles around lawful basis, limiting the amount of data that's held and the length of time it's held for, etc.