Awesome explaination. Though example of global counter is pretty simple as we can assume that counter will never decrement and max function can be safely used for merging and it can be considered as source of truth
What if Bird Watch wasn't relentlessly positive, is it significant that this model is grow only? I can't think of a reason why a decrement button and negative values wouldn't work in this model
The original paper also reuses the acronym a bit to mean "Convergent" or "Commutative" replicated data type, when differentiating state-based vs operation-based CRDTs.
Clear talk, but too simplistic and really slow-paced (for me). I was especially disappointed that the big promised "payoff" of helping to understand the Monotonic Join Semi-Lattice was never really explained, because he just gives the same example at the end and vaguely says "that's one of those". To understand what a technical definition like this means, you also need to see an example of something that does *not* satisfy the definition, i.e., a join semi-lattice which is *not* monotonic.
This is suuuuuppperrrr helpful! It gave me foundation to understand other talks like ones by Marc Shapiro. Thank you very much!
Awesome Talk very well explained with an example. Would love to see other CRDT examples :).
Fantastic talk! Thank you.
very clear presentation!! good job!
Awesome explaination. Though example of global counter is pretty simple as we can assume that counter will never decrement and max function can be safely used for merging and it can be considered as source of truth
Best presentation for CRDT
Very welll explained, thank you!
What a great talk!
Great talk, thanks, John!
Isn't the least upper bound like finding lowest common ancestor in a tree data structure?
This is gold.
excellent talk. is the bird watch live somewhere?
This is a nice talk bridging pure math with real life applications.
What if Bird Watch wasn't relentlessly positive, is it significant that this model is grow only? I can't think of a reason why a decrement button and negative values wouldn't work in this model
Neat topic. But at any time did the speaker define what the acronym CRDT meant?
Conflict-free replicated data type
The original paper also reuses the acronym a bit to mean "Convergent" or "Commutative" replicated data type, when differentiating state-based vs operation-based CRDTs.
Clear talk, but too simplistic and really slow-paced (for me). I was especially disappointed that the big promised "payoff" of helping to understand the Monotonic Join Semi-Lattice was never really explained, because he just gives the same example at the end and vaguely says "that's one of those". To understand what a technical definition like this means, you also need to see an example of something that does *not* satisfy the definition, i.e., a join semi-lattice which is *not* monotonic.
I think this would be up to your merge logic. If your merge logic used sums instead of max then it wouldn't be monotonic.
Writing
Why R? Writing