Yesterday, Katerina commented, “What happened to your keyboard? It sounds so much better than usual.” By better, she meant quieter. The answer is that I was using a Surface Laptop instead of my usual MacBook Pro. And, I have to agree. The keyboard is a nice change, and lovely to type on.
I just installed the latest Windows 10 Insiders Fast Ring build (18346) with the new Windows Subsystem for Linux updates that allows you to access Linux files from Windows.
Before this, to use Visual Studio Code or other Windows apps on the same files you were using with Linux tools, you had to be careful and make sure that you were working in the Windows filesystem, exposed via the
/mnt/c path. This worked pretty well, but it wasn’t quite ideal. I felt a performance hit when running Linux-based tools on a large number of files. And, it felt like the main user story it solved was to enable using Linux tools on Windows files. While that’s a good use case to solve, it wasn’t my use case.
Now, you can open up any Linux path through what looks like a network connection to Windows apps. This means that you can open up your Linux home directory in Explorer and work directly on files that are located there with Windows applications. The net result is a lot better for the way I work. And, any potential performance hit happens on one file at a time instead of when I run a compile on a whole bunch of source code.
Better yet, this is the developer experience I really want. It solves the use case where I want to use Windows apps in the GUI — primarily Visual Studio Code as a text editor — to work with files on the Linux side. It lets me treat the Linux command line as one layer, and Windows as another layer.
Not bad. Not bad at all.
Rewrites of a codebase are bad, right? Maybe it’s more complex than that. Herb Caudill looks at six big rewrites and the lessons they teach us.
Molten salt reactors, prototyped back in the 1960s at Oak Ridge, may hold the key for combating climate change and doing something about the 300,000 tons of spent nuclear fuel that will otherwise be hanging around for the next 240,00 years. M. Mitchell Waldrop writes in Knowable Magazine:
Admittedly, it would take centuries for even a large network of molten salt reactors to work through the full backlog. But burning it would eliminate the need to safely store it for thousands of centuries. By consuming the long-lived isotopes like plutonium-239, molten salt reactors could reduce the nuclear waste stream to a comparatively small volume of fission products having half-lives of 30 years or less. By the 10 half-life rule, this waste would then need to be isolated for just 300 years.
The nuclear power question isn’t whether we’re going to continue to run our current reactors with their known deficiencies. Those need to be shut down, and they will be over time. We can’t keep running them forever. The real question, however, is if what we do next will be something that can help with both the current spent nuclear waste and climate problems.
The stories coming out of Revolut are really bad. This kind of toxic behavior is never acceptable, regardless of what kind of pressures a company is under.
I gave into the .dev domain madness and registered duncan.dev for my primary website domain. I’ll move things over the next few days, hopefully not breaking too much along the way.
I’m such a geek. It really didn’t help that our group’s internal Slack was talking about nothing but .dev domains. There was just way too much encouragement there from folks like @bketelson and @film_girl. Really, it’s a miracle I didn’t buy more than one domain.
In an op-ed published in major European newspapers, Emmanuel Marcon argues for a renewal of the European project in the face of challenges from nationalism and Brexit.
Never since the second world war has Europe been so essential. Yet never has Europe been in such danger. Brexit stands as the symbol of that. It symbolises the crisis of a Europe that has failed to respond to its peoples’ need for protection from the major shocks of the modern world.
With Brexit due this month and the European elections in May, this seems to be make or break time for Europe as we know it. I hope that Marcon’s arguments help steer the discussion in the right direction.
The way Stephen Wolfram organizes his life is next level. Especially interesting to me is how he runs his company as a remote CEO.
A relic of the Industrial Revolution, the eight-hour workday’s insistence on working long blocks of time with frew breaks isn’t great for how we work now. Dr. Travis Bradberry writes at Quartz about a better way to structure your work day to be productive:
the length of the workday didn’t matter much; what mattered was how people structured their day. In particular, people who were religious about taking short breaks were far more productive than those who worked longer hours.
The recommendation: Work an hour, then take a 15 minutes break. A real break, like going for a walk. Repeat as needed. It sounds like a lower frequency version of the pomodoro technique.
Azure is opening two new datacenter regions in South Africa, the first from a major global cloud provider to arrive on the continent.
Nikita Prokopov proposes a redesign of the GitHub repository page. I really like it. I also like how he worked through the process in a step-by-step iterative fashion.
For International Women’s Day, I want to highlight Katerina’s work as a speaker coach and speechwriter. The event she just helped curate in South Africa had six speakers, and all six of them got standing ovations during their presentations. She takes particular pleasure in helping other women tell their stories, such as Keavy McMinn telling her story about identity, Katrina Owen talk about mentorship and Melissa Fleming telling the story of two Syrian refugees.
Katerina, you’re kicking ass, and I’m proud of you each every day!
It’s payback time for requiring ESTA visas from Schengen country citizens. Starting in 2021, US citizens travelling to Europe will need to get an ETIAS visa in order to enter the Schengen zone.
The problems outlined in Elizabeth Warren’s proposal to break up Amazon, Google, and Facebook are real. Many people I know are increasingly concerned about how big tech has played out over the last decade. I find myself really uncomfortable, however, with the proposal to use the anti-trust breakup hammer to address the problems.
Some part of my discomfort is probably irrational and likely comes from when I worked at Sun back in the late 1990s. Most of my management chain was consumed by the lawsuit the company was pursuing against Microsoft, and it wasn’t pretty for anyone. In the end, not much good came of it. In fact, I think the lawsuit changed the Sun’s culture for the worse and in a way that it never recovered from.
Of course, the Sun case wasn’t the same as the Department of Justice anti-trust case against Microsoft, but they were concurrent, informed each other, and were settled within months of each other. The two, at least in my head, are intrinsically tied.
Who knows. Was it was a mistake for the DOJ to settle short of breaking Microsoft into two companies? Should they have gone further than they did? Or did the bust of the dot-com bubble and the resulting re-distribution of tech talent in Silicon Valley have a big part to play in the rise of Google and Facebook? And, what about the phoenix rise of Apple from it’s near death experience to industry dominance driven by spectacular execution more than anything else?
It’s tough to say. I don’t think it’s as clear cut as we’d like it to be.
When I push myself to think really objectively about it, I think my unease really centers around the quality of the conversation between technologists and politicians. It’s just bad.
Every time there’s a public hearing, no matter how good or bad the technologist on the spot performs, most of the politicians just come across as clumsy and clueless in the extreme. Of course, there are exceptions and the latest crop of fresh politicians seems like they understand what’s going on. But, the old guard? Not much clue there.
More regulation is coming, of that, I have little doubt. But as long as both sides are talking past each other and understanding very little of what the other is saying, I fear that it’s going to be implemented in a way that’s clumsy and that will hurt as much as it helps.
Today I learned about the Dutch Reach and am going to try to put it into practice. Getting doored isn’t the only danger urban cyclists face, but it’s high on the list.
What’s a product roadmap and how do you build one in the early days of a startup? James Turnbull writes about one way to do it on the Microsoft for Startups blog.
My group at Microsoft is going on tour to twelve cities around the world. This is a big part of my current work and I’m pretty stoked about it.
First stop: New York on April 9th.
Helping companies through the Microsoft for Startups program has reinforced my feeling that many problems companies face as they grow are really people problems in disguise.
To help figure out what a startup needs to focus on at each phase of the startup lifecycle, Wendy van Ierschot gives us a road map:
We’ve identified five stages, each pegged to a certain number of employees. For each stage, we’ve established which areas an organisation should put its HR focus on.
The diagram in the article showing what to focus on at each stage feels about right, and it’s a better starting point to implementing HR strategy in a startup than having nothing at all.
Kubernetes is the new hotness, for sure, and pretty awesome to boot. But is it really needed, especially when you’re bootstrapping a new idea? Maybe not. UK-based Freetrade started out with a decently designed Kubernetes-based stack, complete with all the bells and whistles, but ended up scrapping that plan and launching with Firebase functions.
Even now at 20k users, it’s been the right decision for them.
In fact, not only did we not need it - but if we’d launched with this stack with just two engineers (our launch team size!) I’m confident our customers would have been very unhappy.
As they continue to grow, I’m sure that Freetrade may end up putting some functionality into on-demand container instances or even spin up a Kubertnetes cluster to handle the parts of their app that need it. When they do, however, I’m sure they’ll be able to leave a significant portion of their application surface in functions.
Steven Sinofsky’s tweet thread about going through the Microsoft security crisis is related to Facebook starting its own pivot, but is useful thinking for taking any kind of organization through a significant shift of strategy:
All you can do as an organization executing on this sort of pivot is to do the work. Keep people informed. Build incremental trust across all stakeholders by actions.
Nobody knows if Facebook can pull off what Mark Zuckerberg says that he wants to do, but for sure it’s going to be messy, take a lot of time, and it will be a hard time for all involved to build up trust.
There’s not much that’s reasonable to say about the horrific news from Christchurch today. The actions of the gunman, a self-proclaimed fascist, being amplified and repeated by so many others online…
I just can’t…
I’ve been stuck most of the day. Stuck in thoughts that don’t go anywhere, at least not anywhere good.
The last few years, the worst side of humanity has been winning in a big way, and while there’s nothing new about white supremacy, fascism, violence, or hate, we’re seeing how those old human reflexes have adapted to the tools that we’ve built in and for our online world.
What are we going to do about this?
Beto O’Rourke was a member of a group famous for hacktivism back in the days of the Apple IIe and 300-baud modems? Alright.
Not everyone likes untitled blog posts. Dan Steinberg argues:
As a reader, I remember a post I liked from an author I admire and I look back through their posts and I see a sea of posts without a title. It makes it hard for me to find the one that resonated with me.
Titles are a straightforward affordance for scanning through posts in an archive page. And, it’s not just a human reader that might appreciate the hierarchy a title gives. Search engines really lean on them as well, both for indexing and for displaying search results.
Going title-optional in a meaningful way means thinking about both of these issues.
Trevor Sumner’s tweet thread about why the 737MAX tragedies aren’t a software problem is an insightful read.
Hint: it’s probably a systems problem.