Shane da Silva

@sds

493 Following

25418 Followers

Shane da Silva pfp

One of the most important lessons from this film.

0 reply

0 recast

1 reaction

Shane da Silva pfp

Fair point—abstraction is a better word. I’m not aware of any tool that treats it as a primitive, but when thought of conceptually as primitive it seems inevitable that we’ll see FSMs described in code which will create additional output that can automatically be verified by tools like TLA+.

1 reply

0 recast

1 reaction

Shane da Silva pfp

FSMs as a core data primitive are going to get more popular. Combined with a runtime library for your preferred language, and you have a way to formally verify statements about business logic in your system. https://raphael.medaer.me/2019/06/12/pgfsm.html

2 replies

7 recasts

60 reactions

Shane da Silva pfp

The more direct support for switching JSON to tables seems interesting, not that I currently have a pressing need. The perf improvements to WAL locks are probably the most exciting change overall for apps with high write concurrency.

0 reply

0 recast

1 reaction

Shane da Silva pfp

We do! But in that case the order of processing events matters, whereas if you’re treating it like a job queue ordering of job execution is good enough as long as it’s roughly FIFO.

0 reply

0 recast

1 reaction

Shane da Silva pfp

It’s a project we’ve been keeping an eye on. We’re taking advantage of AWS’ ElastiCache offering to handle orchestration. Are you running yourself or using their cloud offering?

1 reply

0 recast

0 reaction

Shane da Silva pfp

Stoked for this.

2 replies

9 recasts

60 reactions

Shane da Silva pfp

Whatever amount that does not exceed the total global storage limit. See https://optimistic.etherscan.io/address/0x00000000fcCe7f938e7aE6D3c335bD6a1a7c593D#code#F1#L438 CC @horsefacts.eth to check me on this.

1 reply

0 recast

4 reactions

Shane da Silva pfp

Some jurisdictions have protections around titles. “Engineer” is a protected title in Canada, but almost no one that programs in Canada has an engineering degree. Hence “Developer” as the title.

0 reply

0 recast

9 reactions

Shane da Silva pfp

12 replies

1 recast

47 reactions

Shane da Silva pfp

The level of detail and thoroughness is impressive.

0 reply

1 recast

2 reactions

Shane da Silva pfp

Impressive tool from Xata providing a solution for zero-downtime, REVERSIBLE (!!!) migrations—the holy grail. Excited for this to eventually reach v1. Current feature set is very impressive, but it's not yet ready to handle all possible migrations just yet. Something to keep an eye on. https://xata.io/blog/pgroll-schema-migrations-postgres

3 replies

2 recasts

21 reactions

Shane da Silva pfp

This is a factor, and also we did not expect the system to reach this many hubs this quickly. Ethereum, for example, has around ~7K nodes at time of writing. https://ethernodes.org

0 reply

0 recast

4 reactions

Dan Romero pfp

Please spread the word: there are 0 economic rewards—currently or in the future—for running a Hub. Only run a Hub if you: 1. Find it fun 2. Find it useful for your project or your business There's no surprise or gotcha that comes later. All of the YouTube video tutorials are lying to you about a potential reward.

16 replies

33 recasts

129 reactions

Shane da Silva pfp

Would prefer questions be asked in /farcaster so that everyone benefits from the answer. We do expect you to have done at least a little reading of the existing docs, though.

1 reply

0 recast

4 reactions

Shane da Silva pfp

My knowledge of distributed systems locking theory is way out of practice, but feels like it might make it harder to detect deadlocks (or perhaps increase the chance of livelocks) if you allow lock requests to “jump the queue”? That said, someone has probably come up with something clever to mitigate.

0 reply

0 recast

2 reactions

Shane da Silva pfp

Thanks for putting on my radar!

1 reply

0 recast

2 reactions

Shane da Silva pfp

Great post on how PG lock behavior during schema changes. It's surprising (but obvious in hindsight) how a statement that obtains a lock which doesn't conflict with a schema change could still be blocked by another statement that does obtain a blocking lock, due to the FIFO nature of the lock queue. https://xata.io/blog/migrations-and-exclusive-locks

2 replies

1 recast

31 reactions

Shane da Silva pfp

Anyone ever used a higher-level language like PQL for expressing complex SQL queries? https://pql.dev Looks like heavy use of CTEs under the hood to make it easier to express queries as pipelines rather than declarative statements. Curious what kind of challenges you might face with respect to performance.

1 reply

0 recast

14 reactions

Shane da Silva pfp

Love Book Club Radio.

0 reply

0 recast

2 reactions