[josteink]

I’m constantly flabagastered by what’s considered novel in the world of MySQL.

But then I remember this is the database that lied (lies?) about ACID transactions, and had no problems storing non-existent dates in a DateTime-column because its creators considered ensuring data-integrity and consistency “the job of the application layer” (which back then often was PHP).

Nothing should really surprise me.

[kbenson]

> I’m constantly flabagastered by what’s considered novel in the world of MySQL.

What's that in reference to? The article explains that MySQL implemented group by using a newer standard than many competitors did, by looking at functional dependencies. Not exactly novel, but good for MySQL, they decided to go beyond the competition when finally catching up on this particular feature.

There are, of course, many other problems with MySQL (as you allude to), such as silently ignoring CHECK constraints, but that doesn't seem to make sense as what you were referring to in context of what they consider novel.

[captain_perl]

> this is the database that lied (lies?) about ACID transactions

Time moves on. The Innodb storage engine has had ACID transactions for over a decade.

> no problems storing non-existent dates in a DateTime-column

PostgreSQL checks date formats, but updating (or not updating) timezone databases would affect validity. So it's not as simple an issue as you make it to be.