someone wrote a 680 page interactive book on cs algorithms

Late notice perhaps, but we're back in the loving embrace of the City Pride on Tuesday next week (20th January)! Come along to chat with your fellow humans about working with humans and machines. meetup.com/extreme-tuesday-c…

This startup made $10,000 by allowing people to add custom colors to Google Calendar events. Stop overengineering.

What a w̶o̶n̶d̶e̶r̶f̶u̶l̶ delusional world ... I am just speechless

Heard something today that will stuck with me for a while: Nowadays, People cannot tell the difference between “hate speech”, and “speech they hate”. So true …

Look at these 2 UIs. One of them will 100% outperform the other for one reason. Miller's Law. Here's how this psychological principle works & how to use it in your products:

I was told that, of course, Agile is all about project management because that phrase appears in the Manifesto. The Agile Manifesto was written a quarter century ago. How agile would we be if we hadn't learned and adapted over those 24 years? Back then, everybody thought in terms of projects. The notion that a product focus was better appeared on the scene about 7-8 years ago, if I remember correctly, and it's now dominant. In fact, the Manifesto signatories that I know don't use "project" anymore. Similarly, we've learned to deliver work every few days—a "couple months" is now considered suboptimal. Very few orgs outside of megacorporations or government contractors create "comprehensive documentation," and they're not in the least bit Agile. So, the details have changed a bit, but the thinking that underlies the values and principles is still quite valid (and valuable). To me, an Agile shop embodies that core thinking. In any event, the Manifesto was never intended to be anything other than a snapshot in time. It was never intended to be fixed-in-concrete doctrine. To be Agile, you need to be—agile;

Nobody wants data (particularly in tabular form). Everybody wants answers to questions. Don't foist the work of figuring out the answer onto your users. The best software gives them the answer, often with no data in sight. There's a middle ground of showing the graphs you need to get the answer, but that's still not providing an answer. Do the work. Don't present the wherewithal to do the work.

You can argue all day about what customers want or need, what they'll love or hate, but if you don't ask them, you know nothing. Guesswork has no place in product design (unless it forms the basis of a controlled experiment like an MVP that you'll validate with actual conversation & feedback).

Do you remember all those rushed changes that your developers implemented three years ago and how they complained about the design damage they caused to make that happen? It's all still in the codebase. It never disappears. Only YOU forgot it, they still pay for it.

As for putting one of those mythical 10x programmers onto an average or low-performing team, local optimization doesn't work unless you're optimizing at a bottleneck/constraint. It's the speed of the entire system that matters. A trail runner in the middle of the pack who can run 4x faster than everybody else will slow down when they bump into the slower person in front of them. In a software-development system, what matters is when the last person crosses the finish line. Form a team comprised entirely of high performers and then put them at the bottleneck. Everywhere else, they're wasted. Better yet, put them to work teaching others how to become "high performing."

Certain workflows require sharing the entire screen, if the presenter has a much larger display than the audience it can be hard to see what is happening. DeskPad creates a virtual mirrored display that is easily shareable and customisable: github.com/Stengo/DeskPad

"Developers are drawn to complexity like moths to a flame, frequently with the same result." - Neal Ford

Timefold Solver now has 1000 stars on Github! github.com/TimefoldAI/timefo… Try our open source solver for planning optimization in Java, Python or Kotlin: timefold.ai/open-source-solv…

Headline... "Going green cost more than they bargained for" Facts: 1. It was a schedule 30mins wrong (poss hardware issue) 2. THE TOTAL COST DIFFERENCE WAS £1.47 and refunded immediately 3. In 3 months, total cost £100 vs £300 on std tariff 4. IT'S 6x CHEAPER THAN PETROL

The difference between free users and paying users is about 1000x from my experience Anyone can get 100,000 free users But converting those to paid, you can divide than free number by ~1000 and you'll be left with 100 paying users (rough numbers) via @gregisenberg

Flexibility in code is a _cost_ not a _benefit_. "Let's make it configurable!" puts the burden on the consumer, requiring more decisions to make, resulting in more effort, whereas an opinionated setup might be good enough, allowing the consumer to spend that effort elsewhere.

Passwordless authentication 🤣

If you happen to be in London, the SDD Deep Dive (sddconf.com) conference is coming up in a couple weeks. I'll be doing an in-person, 3-day hands-on Incremental Architecture workshop. Learn how to build an architecture that welcomes change and can grow incrementally as you learn. I don't do these publicly very often! @sddconf