LinkedIn

Everybody says they do CI&D. I have not had nearly enough mojitos* for this to be true.

On one hand, this is about the daily work of engineering. What does daily reproducible progress look like? Are your people getting something done every single day, even if it's something small?

On the other, what's actually getting through to your customers? How long does it take to get new features and bugfixes all the way through to the point where customers are using them?

*I don't drink. Certain butler-themed applications used for these purposes have occasionally made me revisit this decision.

There's a lot going on here. The field is maybe a decade old or so, and there are some really neat success stories where people have mixed acres of data with the right raw statistical theory and distilled useful knowledge out of that reaction.

On the other hand, sometimes it explodes.

I'd really like to get this kind of work done safely, but we're still in the early phases. I'd settle for doing it reliably, and, someday, we might even have some idea of how well our applications are working.

One microservice is easy. Multiple microservices are not so easy. Sometimes, when microservices depend on each other, it can enrage people to the point of arson.

HTTP and JSON aren't too bad by themselves, but pretty soon you'll start needing to use mock services, and after that, you can end up needing dynamic mock services. If that sounds suspiciously like another microservice, that's because in point of fact it is another microservice.

I'm pretty sure the next step involves hurling the computer off the roof, and that's before SOAP and XML get involved.

I can spell it, at least. I've gotten to use it some, and, while it's certainly not the most intuitive piece of software on this green and verdant earth, I have high hopes. That said, if I run across one more tool that tries to elevate YAML to a programming language, I might be the next thing to dump core.

I am one of the few people I know who has saved a family member's life with open source software. My son is a type 1 diabetic, and this is what we've used in the past to monitor his blood sugar at school.

I keep thinking there has to be a way to make the setup and operation of this easier. Somewhere, in all my dev/ops experience - both days of it - there has to be an idea or two about how to staple together some infrastructure-as-code so that setting up a new site becomes more "push the button" and less "slay the hydra with the minotaur horn, take the twelve stones of power, and activate the medallion at the exact moment of total eclipse".

• nightscout on GitHub
• NightScoutFoundation on GitHub

I'm an old terminal jockey at heart. I speak HTTPie to microservices, jq to objects in JSON, and q to CSVs.

I've also been very impressed with a lot of the stuff coming out of the Rust community and crates.io lately, but my newest, bestest friends are exa and Bottom.

In my recent work with Python, I came across pipx and Glances, which have also been very helpful.

I'm very happy with my subscriptions to the JetBrains suite of products and O'Reilly Learning.

• ingernet
• dogonthehorizon

I've been fortunate enough to have a wonderful family, good friends, and I've run into lots of people who I've learned from. Some of them even put up with me voluntarily! I'm grateful to all of them, and more besides.

• anuraghazra, for github-readme-stats