Does EXPLAIN confuse you? You’re not alone. This talk will help you understand what EXPLAIN output means, how to interpret it, what’s good/bad in a given EXPLAIN plan, and how to change queries, indexes, or schema to get better query execution plans.
Last week I “finished” a new medium-sized book (an ebook—that is, a PDF). But even as we downloaded the final PDF from Google Docs and uploaded it to our company website, I already had a file full of errata and omissions. Some of these are substantial problems, but my opinion remained almost unwavering: ship it.
I was raised in an environment where feelings were minimized, denied, brushed aside. In some circumstances, acknowledging feelings could even be risky. As a result, I adopted a mostly unconscious perspective on feelings which didn’t serve me very well. Recently, however, I’ve become more aware of my feelings, and I’ve been surprised and pleased to learn what they have to teach me—and to rediscover thoughts, which I now also see in a new light.
I’ve written a new 65-page ebook, titled DevOps for the Database. Drawn from what I’ve observed over the last 15 or so years in my own experience as well as that of many others, it tries to answer why and how some organizations deliver value to customers fast, run data-intensive applications with high quality, and have great engineering cultures—and why some don’t.
Do you know how to be a great conversation partner? I know I’m making progress, but it still feels like I have a long way to go. I’ve had a lot of teachers—my uncle, great books, friends, my spouse. But Celeste Headlee has earned a place high on my list of influential teachers of conversation.
This is the last in a series of scheduling- and calendaring-related topics. This one’s about my overall strategy of using scheduling as a tool to manage my time. Clearly, managing my time isn’t the same as managing my calendar, but managing the latter does help with the former. The essence of my strategy is this: I care about my time, but others care about my calendar, so I can forestall a lot of intrusions on my time by using my calendar intelligently. To do this, I block off time, keep an eye on how many meetings I’m in, schedule my work when I need to, and tell people how I prefer to use my time.
Today’s blog post is about—you guessed it—more calendaring and scheduling stuff. This time I want to focus on one of the most unbelievably wasteful uses of time: time spent scheduling meetings. It’s a necessary evil, but can we at least optimize it, and get the lesser of the evils?
Following up on my post about the benefits of speedy meetings, here’s a collection of tips that I think improve calendaring etiquette. Some of these are personal preference, but all of them feel like genuine improvements to me in real life.
Having recently written quite a bit about email and communication, I want to turn to some of my backlog of thoughts on calendaring, scheduling, and meetings. I have learned a lot of techniques that I believe make me and my colleagues more effective and efficient, but there’s one simple tip that rules them all. It’s this: don’t schedule things back-to-back—leave gaps. The reasons why are a lot more important than I thought at first, and I encourage everyone to do this.
After writing a series of posts on various email clients, it should be clear that I’m still an email user. But I am doing my best to move as much communication from email to Slack as possible. I like to have as few “systems of record” and “official communication channels” as possible, and when I can, I’m making Slack my preferred one. Why do I want to do this? How is Slack better than email? There are actually a few fundamental differences; it’s not just the same distraction in a new place.
I’m using Outlook again on my iPhone and iPad. It’s the latest in a series of email clients I’m evaluating to replace Inbox, which Google is discontinuing. I used Spark for about four weeks and took pretty detailed notes, then switched to Outlook. This isn’t the first time I’ve used Outlook on the iPhone; I used it in 2015 until I learned to love Inbox. It’s interesting to revisit it after those years, and I’ve taken some notes to share about that.
As I wrote yesterday, I’m sad that Inbox is going away, and it’s hard to bear the thought of using GMail again. Thus, I’ve been evaluating a few different email clients to see what’s available. One that I tried recently is Spark, an email client for iPhone, iPad, and Mac. Spark feels not-quite-mature, yet still pretty impressive. In this article I’ll summarize what I see as its main attributes and areas for improvement.
Google’s Inbox email client is going away in a couple of months, and I’m sad. As I’ve been looking around for a replacement, I realized that I never wrote about how I came to appreciate it. In fact, the last thing I wrote about it was quite disparaging, so I’d like to set the record straight.
A couple of weeks ago I recalled a thought from long ago, and with it the feelings. I was able to access, with some degree of fidelity, the state of my mind and body at a point in the past. Nothing unusual; something I’m sure we all do a lot. But this ability, which seemed ordinary until that moment, suddenly struck me as extraordinary, with far-reaching implications.
My Twitter profile currently includes the phrase “silence is violence.” People ask me about this occasionally: what does it mean? I don’t know where I picked up this phrase, and maybe it has a different meaning for me than for others, but what I mean by it is that if I’m a bystander who witnesses something I disagree with, I’m actively supporting the behavior to which I object.
Years ago I was recruiting a replacement for myself in a leadership position, and I interviewed a Canadian man. I remember being impressed with the candidate’s thoughtfulness and people skills. After a while I asked him, “what is something you’ve learned from leadership?”
I wrote yesterday about Vitess, a scale-out sharding solution for MySQL. Another similar product is Citus, which is a scale-out sharding solution for PostgreSQL. Similar to Vitess, Citus is successfully being used to solve problems of scale and performance that have previously required a lot of custom-built middleware.
