amw-zero.bsky.social
@amw-zero.bsky.social
Working on database observability at Datadog. Personal blog at https://concerningquality.com/.
"What does it mean for a system to be correct? One can
informally enumerate a set of properties and hope they are
sufficient to provide correctness..."
February 18, 2025 at 3:48 AM
This is a fantastic idea (comparing observed application traces with an offline specification) and a form of runtime verification:

www.linkedin.com/posts/marc-b...
Marc Brooker on LinkedIn: One aspect of the work we describe in Systems Correctness Practices At AWS…
One aspect of the work we describe in Systems Correctness Practices At AWS (https://lnkd.in/gBSfXebN) is the work Ankush Desai and team have been doing on…
www.linkedin.com
February 13, 2025 at 1:05 PM
To this day, Ruby is my favorite language to _write_. I don't use it at work, but it's still my go to language for "thinking out loud".

For example, I wanted to calculate the powerset of an array:

`0.upto(arr.length).map { |i| arr.combination(i).to_a }.flatten(1)`

I'll always love method chaining
February 7, 2025 at 3:22 AM
The Collatz conjecture is a concise example of why branch coverage doesn't mean correctness: concerningquality.com/collatz-conj...
Branch Coverage Won't Prove The Collatz Conjecture | Concerning Quality
concerningquality.com
January 27, 2025 at 12:20 AM
A coworker reminded me of this post - Queues Don't Fix Overload: ferd.ca/queues-don-t.... It does a great job framing a system as a set of interacting queues. This has become my mental model for sure.
Queues Don't Fix Overload
ferd.ca
January 6, 2025 at 8:59 PM
Here's a post looking into the effect of response time distribution on various queueing characteristics. It uses SimPy to simulate several minutes of operation in a queue in various different configurations.

concerningquality.com/queue-simula...
Simulating Some Queues | Concerning Quality
concerningquality.com
January 6, 2025 at 12:11 AM