gritzko

24 ms·

331.
by gritzko 7y ago
swarmdb has CRDT-based merge and causal branching which is "super smooth". Lets you recombine changesets rather freely, or at least supposed to. https://github.com/gritzko/ron-cxx/tree/master/d
332.
by gritzko 7y ago
...by adding a single point of failure?
333.
by gritzko 7y ago
Can't say it better. I welcome your thoughts on swarmdb http://github.com/gritzko/ron-cxx That is syncable RocksDB with CRDTs inside. Pre-alpha.
334.
by gritzko 8y ago
In theory, there are incremental paths to achieve that objective, while serving clients all the way along. 1. Define a P2P ready data model and protocol (which they have, Merkle forest and everything) 2. Run a single server/cluster, m
335.
by gritzko 8y ago
What you actually want is to break the universe. It is physically impossible to revoke information unless it happens by a strange coincidence. 24x7, you emit information that races away with the speed of light. You can't chase it dow
336.
by gritzko 8y ago
Me and the other guy. I am Victor Grishchenko, PhD. At the time, we had a CRDT based sync engine which we white-labelled to Yandex. We had been eager to work with Evernote cause all their sync issues have been very clear to us. These days,
337.
by gritzko 8y ago
We contacted Evernote in 2012, 13, 14 begging them to try CRDT based synchronisation. First, it might have fixed their sync glitches. Second, that would make the client a full copy of the data, very much like a git repo. You can't lose
338.
by gritzko 8y ago
That resonates well with my conclusions that led to the Replicated Object Notation project. [1]. If the parser creates an AST tree or some number of dictionaries or some other bullshit... "now you have two problems", that's i
339.
by gritzko 8y ago
> Roosevelt was probably sicker and had to travel all the way to Ukraine. That is a bit sensitive these days, but Crimea wasn't Ukraine at the time. It was transferred to Ukraine in 1954 by N.Khrushchev (without much elaboration).
340.
by gritzko 8y ago
The new x1 extreme / p1 is notably missing.
341.
by gritzko 8y ago
Well. On page 30 par 2 you conclude that OT has time complexity O(1) while CRDT is O(C) or so... I guess, you compare a rope-based OT implementation and a naive CRDT implementation (like, one doing a full rewrite on each key press). At leas
342.
by gritzko 8y ago
My personal experience is different. OT tends to be quadratic in terms of computational complexity and codebase size. As long as you consider a simple case, 1^2 is 1. Then, add more data types (not just text), faulty servers, faulty network
343.
by gritzko 8y ago
Recently one guy reverse engineered Apple Notes. It is a CRDT. I personally made a CRDT sync for Yandex. There is also list on Github... https://github.com/ipfs/research-CRDT/issues/40
344.
by gritzko 8y ago
RON CT/RGA may consume inputs in arbitrary order.
345.
by gritzko 8y ago
An article of disappointing quality from well known OT authors. Like,algo x has issue X, algo y has issue Y, z has Z, so CRDT has issues X, Y and Z... and many things like that.
346.
by gritzko 8y ago
I see that often in smaller hotels: they sell rooms in several online systems, then reconcile it all in a copybook. Personally, I had interesting booking.com situations too. Like, I arrive to a booked apartment, dial the password, the door
347.
by gritzko 8y ago
I am the author of the CT vector CRDT, also the author of the Replicated Object Notation http://github.com/gritzko/ron . I tried to look into the code, and I am not sure which way they implemented list and vector. The e
348.
by gritzko 8y ago
So what? Chrome was based on WebKit (Safari), Safari was based on KHTML (Konqueror).
349.
by gritzko 8y ago
That's very interesting. Thanks!
350.
What’s Next for Mobile at Airbnb(medium.com)
1 points by gritzko 8y ago|0 comments
351.
ACID 2.0 – Associative, Commutative, Idempotent, Distributed(swarmdb.net)
3 points by gritzko 8y ago|0 comments
352.
by gritzko 8y ago
> I hadn't seen RON It looks really interesting It is like the Lamport event model married to a LSMT database. And their children are CRDTs. > That makes sense to me except for calculating diffs efficiently. Are you saying that c
353.
by gritzko 8y ago
1-3 is generally true for ORDTs (in archagon's terms). I wonder, how much metadata noms needs to achieve those features? CRDTs brute-force all those distributed/concurrency issues by adding some metadata. With CRDTs, two hard topi
354.
by gritzko 9y ago
1. 7574 is not "novel" indeed. About 10 years old, if you count the original draft. 2. 2^64 is quite a lot, for a single file. The variant I described deals with a single static file. Also, it is a part of a network protocol, so t
355.
by gritzko 9y ago
RFC 7574 Merkle tree solves that rather strictly (I am the author of the scheme). Max file size is 2^64 bits, so the hash tree is defined as a binary tree covering the entire 2^64 range, leaf layer made of 1KB pieces. A hash of an empty ran
356.
Replicated Object Notation: a CRDT-friendly data sync format(github.com)
3 points by gritzko 9y ago|0 comments
357.
by gritzko 9y ago
It's good you welcome contributions to your open platform. In case your next demo is based on my work, please notify me in advance.
358.
by gritzko 9y ago
It is not a drop-in thing at the moment, that's for sure.
359.
by gritzko 9y ago
It's the new economy, man. It's perpetual growth on the blockchain. https://news.ycombinator.com/item?id=15169613
360.
by gritzko 9y ago
You didn't make a point, sorry.

More ›