cjqf

@carsonfarmer

15 Following

26 Followers

cjqf pfp

Runs on coffee https://www.youtube.com/shorts/rk8wnnhuSDw

0 reply

0 recast

1 reaction

cjqf pfp

A coprocessor based design metaphor is extremely useful when you consider how great blockchains are for pipelining computation. But it also works super nicely in an async context. Async coprocessor stacking provides a nice way to think about layering compute and other operations over data/transactions.

0 reply

0 recast

2 reactions

cjqf pfp

It is a coffee and French hip hop kinda morning... ☕️

0 reply

0 recast

2 reactions

cjqf pfp

I would wager that a good `fmt` tool for an ecosystem (rust, go, deno, zig, etc all have one) saves *weeks* of dev time for some orgs... not thinking about formatting etc means more thinking about logic and code design.

1 reply

0 recast

0 reaction

cjqf pfp

Turns out, calendar invites work best when you actually include the other person in the invite... 🤦

0 reply

0 recast

0 reaction

cjqf pfp

Gotcha, thanks for the lulz! So unless something extreme has happened in the past ~year, it is safe to assume I can keep moving. I'm currently exploring nim and zig...

0 reply

0 recast

0 reaction

cjqf pfp

So what is the current state of vlang? Looking at https://vlang.io, it looks amazing. But I hear a lot of complaints also. What, in 2023, is the latest objective take? Worth exploring?

1 reply

0 recast

0 reaction

cjqf pfp

Damn, really sorry to see #protocollabs layoff announcement today 😢. Good luck to all those affected!

0 reply

0 recast

0 reaction

cjqf pfp

Really Safari... no look-behind in your RegExp engine? It is 2023?!

0 reply

0 recast

0 reaction

cjqf pfp

have used, can confirm... noice!

0 reply

0 recast

0 reaction

cjqf pfp

only one way to find out for sure! pop over here and let's figure it out: https://discord.gg/dc8EBEhGbg. but seriously, i'm guessing we'll want to bench writes... that's where the performance hit is going to happen if anywhere. also, adding nodes to the network can improve read throughput

0 reply

0 recast

0 reaction

cjqf pfp

no read does't need to go thru rpc. open network, so access control on reads would just be lying to ourselves 😜, so it is quite fast. And actually, writes are pretty fast when pushing via rollups/L2s.

1 reply

0 recast

0 reaction

cjqf pfp

Yeh I mean this is an amazing super power. We like to think of this as "mutable data with immutable (access) rules". I love that. Smart Contracts to control contract-level behavior (access control) and off-chain nodes to support storage and adherence to said contracts. Beautiful!

0 reply

0 recast

0 reaction

cjqf pfp

But having a blockchain in the mix makes data availability and consistency a breeze, so this is pretty nice. Frees up engineering creativity for much important problems like verifiability etc

2 replies

0 recast

0 reaction

cjqf pfp

Yeh nice to have very specific requirements like this! I think Tableland hits all of those, save for maybe "private" per se. We made a conscious decision to leave encryption/privacy to the application layer, and I'm really happy with this tradeoff.

1 reply

0 recast

0 reaction

cjqf pfp

Not really in any practical sense. Two very different approaches to the problem of distributed data. ThreadDB was designed for pure p2p sync, whereas Tableland network will end up being more federated. ThreadDB relies on peers to make data available, whereas Tableland relies on blockchains for data availability...

2 replies

0 recast

0 reaction

cjqf pfp

Hey, member of the tableland team here. Would love to hear your thoughts on alternatives to on-chain mutations. We've been exploring alternatives like ipfs hashes and other forms of verifiable off-chain data "dumps". Would love to hear your usecases for this type of thing!

4 replies

0 recast

0 reaction