2022#12 Readings 🇺🇦🌻
3 minutes read | 564 words by Ruben BerenguelI took some days off for Easter, and I definitely needed them.
The Problem with Tech Leads
This is a well-written piece about how to become a tech lead or team manager and how to improve at it.
The best engineering interview question I’ve ever gotten, Part 1
Once you are done with this one, you probably will want to read Part 2. I don’t think I’d pass this interview (my C++ abilities are close to 0, and my C knowledge is basic) but the approach to solving… I’ve been doing this for more than 15 years!
“Stories” Don’t Tell a Story: Good Sprint Planning Uses Milestones
I like the premise here: seeing a JIRA board tells you nothing, because it’s hard to know what 100-piece puzzle you are building when you can only see 10 pieces (and sometimes, some are upside down).
Subclassing in Python Redux
I’ve had this article from Hynek as pending to read for close to a year, because it looked very long but I was sure it was going to be great. Surprisingly, it wasn’t that long. But it was great.
Deep dive into Apache Kafka storage internals: segments, rolling and retention
Probably more than I wanted to know about how Kafka stores data in disk when rolling up partition segments.
Minto Pyramid
I first heard of Barbara Minto’s pyramid structure for business writing in this talk: Executive Communication w/ Harrison Metal. This “untool” gives you the gist of it, which is what you usually need.
A playbook for managing & leading in difficult times & crises
Worth bookmarking. You never know what is the next black swan we live.
On “doing a good job”
You have to have your own definition of doing your best work. You can do your best work with a client and they can still struggle to make progress. You are not in control of their behavior; you are only in control of your own behavior.
Modern Data Stack: Which Place for Spark?
I have been thinking (for more than 2 years) that Spark can be considered just another SQL engine that just happens to have Scala, Python and more lower-level APIs. At a previous company we migrated a SQL workflow from running in Redshift to (directly) running on SparkSQL in Databricks and it was painless, just making sure the same data source tables were present in both.
Two-way writeups: Coda’s secret to shipping fast
This looks like an improvement to the writing culture Amazon advocates for (the 6 page memo). I’m a big proponent of more and better writing and the problems mentioned here are why some people (and companies) hate writing.
