update employees set salary = salary * 1.10;
It seems to me that this is fundamentally a restricted form of reflection over types which you control (with some type level guarantees).
Am I missing something here?
The original paper could have been better titled: Why not ignore 40 years of database theory and rediscover that network databases are hard to update and then write a ridiculously complex recursive function to traverse the database so we can pretend both to be discovering something new and that it makes sense to consider a database as stateless.
Every time I see one of your comments you seem to present the least charitable interpretation possible of the given post.
The paper presents a way to generalize structural induction, and do so in a semantically sane, maintainable, type safe way.
You’re suggesting that databases could completely replace a program’s data model? Why even have structs, right? Just put it in a database! Ha! Take that, type theorists. Proven useless again.
The “ridiculously complex recursive function” only needs to exist once, in a library, to be used everywhere. Is it a crime that an interpreter or compiler has complex recursive functions with a simple interface that you rely on? Or should we insist that work is useless as well?
Your ability to be smug without actually saying anything at all continues to surprise me.
Meh, don’t let it worry you. The comment doesn’t match anything in the paper, so it’s clearly an attempt at an emotionally charged comment, not an attempt to communicate anything meaningful or helpful.
Some people just want you to feel bad or upset, don’t let ’em get to you.