This is a list of articles that I found interesting and want to note down for myself and share with others. You can also subscribe via RSS.
Finally I found the missing link to set up EDID overrides on Big Sur and later, where System Integrity Protection (SIP) is enabled.
In short, MacOS also takes overrides from the path /Library/Displays/Contents/Resources/Overrides, which is admin writable.
Most companies and projects are by far not big enough to benefit from microservices, and not good enough to deal with the implications and repercussions. There’s a reason “distributed systems” are hard: it’s the next difficulty level after multi-threaded concurrency — harder to observe, harder to reason about.
The aspect of “transaction cost” for doing a particular thing once or multiple times is interesting.
Starting a bunch of things in parallel will often lead to many being finished at a similar point in time, often all at once and leading to the dreaded ‘big bang integration’, which even in short sprints may be painful enough already.
That said, I love starting multiple things at once. Sometimes being ‘stuck’ on the same thing and not having some other outlet or diversion to put your mind to makes a task take longer. Having the ‘diversion’ often gives me more energy to breeze through the other task… and sometimes it doesn’t. It’s not scientific.
A follow-up to the question whether we do REST wrong that provides examples of what constitutes a fully RESTful service.
And here I thought I knew full well what RESTful APIs had to look like. The constraint in my head was that individual resources (e.g. items in a database) should have their own URL and you used the HTTP verbs (GET, POST, DELETE, etc.) correctly.
What was missing from that is the Hypermedia aspect, where each response defines the appropriate URLs for the possible next steps.
More...“Be an engineer, not MacGyver.”
Always work at a place where you and your colleagues give a shit about the product(s), the work, and the people around them.
An interesting look at how long lead times from idea to specification to implementation to release can cause waste in the software development process.
More...A deep dive into the protocol powering HTTP/3 and others, implementing a lot of TCP and supporting encryption at all steps.
An excellent guide I found via Tim Bray’s blog ‘ongoing’ - Making Code Faster.
At work I develop in Go and this is a good list of things to look out for and avoid.
I‘m learning about eBPF. It’s an absolutely fascinating technology.
“The only purpose of a plan is to ensure better results, it has no value in itself. And as James wrote, we must know what to overlook, as sometimes it is better to have no plan at all.”
I value knowledge and being able to respond or adapt when it’s needed over obsessive planning of details that will be wrong the second you get to working on something. Many people are able to adapt and figure out the details for themselves — maybe not in the way you would but in a way that also works.
A broad strokes direction as “Plan” is very often useful, though, which is exactly “planning is for doing”.