All posts by Benjamin Vulpes

Tests are for getting rich slowly

Meeeeester James Steinberg claims:

Until you are making a million a year in revenue, it is a waste of time to write tests.

We must accept this argument from authority! Señor Steinberg has imploded a startup, and sold others off (although with no mention of total internal return, but let’s not let boring old notions from finance get in the way of opinions from lottery winners)!

If you’re the kind of person who likes to play scratch-its with your career and production stability, it’s entirely reasonable to test everything by hand every time you change anything; that’s the mindset I’d expect from someone playing with scratch-its.

Those of us who make calculated bets instead of playing the startup lottery, who understand alien technologies like expected valuetotal internal rate of return, and compounding returns take a markedly different approach from the throw-up-a-landing-page-and-sell-a-startup-on-the-basis-of-mailing-list-signups crowd. In that we make investments that pay accelerating dividends. Continuous Delivery as a superpower, you know? If only we could inquire with Herr J. S.: “What about continuous delivery? Is that for millionaires as well? What tooling is worth investing in? Should I just tar up my php dir and shit it directly onto my Digital Ocean VPS?”

Not writing tests is a false economy. “Run through your service/app/site manually as a user, sure”, the man says. But how many times are you to do that? If you’re working on a change deep in a user flow, will Steinberg’s Methodology of Virtuous Sloth force you to click through ten screens to validate each change? If that’s your workflow, what amount of time have you lost being unable to programmatically determine if your changes are working? Moreover, how do you develop confidence that the rest of your system remained functional? Do you click through that as well?

Investing in tests is daunting, and if it’s not a set of muscles one’s in the habit of flexing (and been forced to live without), the upside is entirely opaque. It’s the less-than-glamorous engineering side of our work. One has to have intuition for what to test, and how to test it effectively. Setting up testrunners is a laborious and sometimes arduous process, all the more so the longer the project’s been delayed. It’s the work we do when we’re not delivering features so that we can deliver features at a serious cadence. It’s the work we do in the engine room, emerging greasy and glorious in the eyes of our fellow engineers, but a complete mystery to our colleagues in Product. A failure to write and automate testing at the outset is just as expensive as failing to write deployment automation at the outset.

And no matter how fast you feel like your engineers are going today, on a greenfield project, with no tests, you will hit the ceiling on throughput quickly, as manual testing and regression fixes begin to absorb your entire engineering team’s attentional budget.

Tests, just like deployment automation, is an investment with compounding returns. If your engineers universally have confidence that they can make changes and confirm the change works in very quick loops, they will universally move more quickly than if they’re testing every change by hand. If you want to onboard new engineers and they have to boot your codebase laboriously and manually every time, and start clicking around to see if their changes work and to understand if they’ve broken anything, you’ll lose time there. Your new engineers likely won’t even know the critical flows to ensure they haven’t broken, and your engineering leads will have to assign them low-risk tasks because you don’t have guards against fixing things, and then your merge gates will similarly have to involve manually testing everything. For every single change.

“Tech debt” feels like a cutesy name, but it’s also more accurate than a tightly-collimated beam. As you rack up debt (in terms of unreadable and cross-cutting implementations), the cost of making each change will increase. Adding another engineer to the project will slow the whole thing down as they extract tribal knowledge from their colleagues. As your featureset expands, the amount of manual testing necessary to validate every change increases, and you have a combinatorial explosion to keep an eye on as your engineering team grows and wants to merge and validate more changes on an ongoing basis.

Perhaps the good mister is right after all, but let’s give it some more color. Tests are for engineering teams who want to become millionaires slowly, and then quickly. This is precisely the Get Rich Slowly approach to becoming a high performing engineer. It’s not enough to just have 5 years of experience in React, that just makes you a React developer with 5 years of experience riding a surfboard on the shifting sands of JS. Strong engineers a) have experience in multiple contexts (and no, Next in addition to vanilla React doesn’t count), b) know how to collaborate on software in a team context and c) build systems to take their code from development to production without tears.

A strong engineer is a millionaire in competence. They codify what would otherwise be implicit project knowledge in tests. They start the project with tests, so that it’s already cleanly factored and prepared for automated testing throughout its lifetime.

But perhaps this perspective is too alien. If you only work with folks with 3 years of writing JQuery and 2 writing React, it’s likely that they won’t be competence millionaires out of the gate. You’ll hire them for their ability to smash out changes in JS, not for their broad mastery of the collaborative software development process. Is this really what you want, though? Just because you’re smashing features out to get your product to the magical flywheel point where user-to-user recommendations bring more users into your funnel than you lose in the churn doesn’t mean that you can afford to add features that reduce churn. Failing to consider the impact of quality and reliability on churn is that classic myopic startup mentality, that eg “only MRR matters” or “only signup rate matters”, or that “if we just fix the top of the funnel, our churn rate won’t matter”.

Churn matters. The experience of quality matters. Users lost to a broken experience that nobody realized was broken because there’s no automated testing matter. High velocity in software development matters, and making investments that compound their return over time is the only way to get there. There are no silver bullets (in software or any other production process), there are only a complexly interrelated set of tools that leveraged together unlock successively more and more magical engineering workflows. You can’t have preview environments until you have deployment automation. It’s hard to stomach the risk of continuous deployments to production without automated tests and robust monitoring. QA staff will balloon in count unless you teach them (and insist that they!) automate their work; and you’ll have to invest in tooling to support their automation until you have the budget for dedicated testing engineers.

So put it another way: tests are necessary but insufficient to become a millionaire, unless you’re playing the lottery. Just like compounding your personal savings over time is necessary to have a hope in hell of retiring before 75 in the modern American economy, unless you’re into playing the lottery with your career.

Just remember: lotteries are -EV, and winners always overvalue their own influence over outcomes.

TarPit – a Proof-of-Work HTTP Ratelimiting Scheme

I’m pleased to announce that my employer, SeaWorld Parks and Entertainment, has graciously granted me permission to open-source a research and development project that I wrote on their dime called TarPit: a proof-of-work-based API rate limiting .NET Core middleware.

It’s not a protocol, it’s not a library (technically it’s a library in the sense of the source code being open for use and reuse, but I don’t really encourage that at all), it’s a notion that’s been floating around for some time in several different contexts that I finally scraped down the cycles to actually implement. I’m not a .NET guy (I’m barely qualified to call myself a software guy), so when I found myself with some breathing room, I undertook to boot up on the CLR in earnest, and C# in particular. Some goals, in no particular order: write a load of C#, write tests, work on something intellectually stimulating, do something that had never been done before, solve a real problem, familiarize myself with the .NET Core HTTP stack, familiarize myself with .NET Core middleware nuances, lean into Dependency Injection, and wrap my head around the popular (in .NET-land) Clean architecture. I also wanted to get to a point where I could load test the same application with both a Redis and Postgres data store, but decided to park and publish this code before I got to the load testing. Continue reading TarPit – a Proof-of-Work HTTP Ratelimiting Scheme

Fine points of rowing your own

(Disclaimer: drive at your own risk. I am not liable for anything you do in your own car)

Being that driving stick (or “rowing your own”) is a skill soon to be relegated to the dustbin of history, what with the 10-gear automatics in the 5L Mustangs (although the high-end Lexii still sport that old-school, low-fidelty 6-speed auto…) it is fitting and appropriate that I should elect this moment to write up how one is to approach managing gears in a standard transmission.

For my entertainment, this piece assumes a working familiarity

We have three goals when operating a vehicle (on roads, in the execution of our daily duties, the world has not enough caveats for this parenthetical): have fun, deliver our passengers an enjoyable ride, and do the whole thing in a safe fashion.

Goals dictate operational constraints (rules for our solver, if you will), and we use tools to satisfy our constraints.

Safety is a metaconcern that must be addressed at a higher level of abstraction than gear selection and transition, so we can dispense with it forthwith.

The goal of having fun provides a convenient operational constraint: we must always be in the correct gear. “Correct gear” is highly situationally dependent. For fuel economy, aim for the highest gear possible that keeps you going down the road at the speed you’re trying to keep. For power, you want your tachometer to indicate that you’re in the “power band” for your engine, that sweet range where it’s optimally converting essence du Motoя into vitesse du voiture. At any point in time, you must be able to demand that the engine give it all that you has, and you may not simply mash the gas pedal into the floor and expect anything to work. No, you must actually downshift, and in doing so, a) not rattle everyone around in the car and b) not wear your clutch out.

The goal of an enjoyable ride for our passengers demands that we shift as smoothly as possible between gears, such that our passengers don’t even know that we’re shifting. It is, I contend, possible to deliver a driving experience that exceeds the automatic transmission for all velocity derivatives.

Your tools (should you choose to avail yourself of them) are rev-matching and toe-heeling. Alors.

Rev matching is simple. When you decide that you need to be in a lower gear, because there’s a corner coming up that you want to power out of, because you want to cruise into the subdivision at something less than a blistering 45 miles per hour, or because the ding-dong in the left lane going sixty-five in a fifty has burnt through what few stores of patience you have left after dealing with dipshit vendor “customer success engineers” all day, or whatever reason may compel you at any time, you must blip, nay, caress the accelerator such that the engine revs up into the bottom of its power band, and ease the clutch in so that as the engine slows down from the wee skosh of fuel you shot into it, the clutch catches and everything matches up perfectly. Learn to effect this move, and your passengers will stop flinching when you downshift. The real joy comes from when it’s completely automatic, and you can blip your revs up and let the clutch out super fast because you can feel the precise right moment to make it all happen.

The poorly-(or more generously historically-)named “toe-heel” technique involves manipulating both your brake and gas pedals with the same foot. In human-sized cars, you can make this happen with the left and right sides of your foot, but in my pickup I actually have to keep my heel on the brake pedal and metatarsal on the gas. My goal is always to come in as hot as I can, and brake at the last possible second, dumping as much speed as possible in the shortest runway possible. I generally pop out of gear a few seconds before I’m trying to apex through a turn (turns are one of the few times that toe-heeling actually makes sense), apply the brakes, drop one or two gears, and if I’m good with my timing, blip the gas without letting off the brake, let the clutch out quickly without disturbing vehicle balance (messing with balance will ruin your day at speed, if you’re anywhere close to the edge of the performance envelope with your tires), and gas my way from right before the apex onto the straightaway (or whatever’s next in my daily urban obstacle course).

Go forth and drive well. Downshift with blips to rev match (if the clutch drags the engine up to speed you dun fukked up), let the clutch 80% of the way out before applying gas when upshifting, and toe-heel your way through corners where you’re downshifting. None of it is hard, and it all takes practice, so get out there and hail motor!

Tooling and workflows for high-output engineering teams (and their stakeholders)

[1]I like that “stakeholder” imparts this image of someone holding a pointy wooden bit to ones chest…My team at SeaWorld is, depending on how you look at things, either stretched ridiculously thin, or incredibly efficient at what we do. We’re responsible for: branded toolchains on around iOS and Android codebases (producing builds for each of our brands on each platform from two distinct codebases); a handrolled API gateway frequently described as “the mobile API”, which caches CMS content and abstracts over several different; and a Next.js application that gets variously webviewed into the native apps and delivered on its own. We get all that done, delivering not no features, and all on a team of about 10 engineers, 1.5 product people, 1 designer and 1 QA (and no project managers, if you can believe it! Jokes, I want a PM like I want a static compiler for Python). I contest that the only thing that makes this remotely possible is our continuous deployment automation, and I further assert that focusing on engineering productivity is an excellent locus for IC attention as you move up in seniority and into management/the chain of command. Join me for a tour of the last two years of evolution in CI and CD with my brand new team.

Continue reading Tooling and workflows for high-output engineering teams (and their stakeholders)


1I like that “stakeholder” imparts this image of someone holding a pointy wooden bit to ones chest…

Basic Olympic Rings Routine

3 circuits. Sets of 5. 2:30 between sets.


A day:

  • mountain climbers (knee to outside elbow, between arms, cross to opposite elbow. repeat on other knee. that’s one rep)
  • push ups
  • leg lifts above the rings (from standing, arms fully extended and gripping the rings from above, raise both knees to chest. progress to fully extended legs. progress to l-sit hold)
  • dips (progress to dips with knees held at 90deg to body, progress to full l-sit dips)

B day:

  • pull ups (progress to muscle-up)
  • leg lifts below the rings
  • chin ups
  • v push ups (legs close to hands, arms fully extended into a V, bring head to ground and then push up to full arm extension)

This is about 30 minutes, and gets the majorly important areas. Stretch before and after, consume protein after.

Get strong, fam!

Happy Imbolc!

I personally turn into a dead zombie of death over the winter months. Living at the 45th parallel in the stupid little microclimate of the Portland metropolitan area guarantees that I can’t get a decent amount of UV in my face and so my vitamin D levels drop and I have to (literally) dose myself with reptile lamps in order to keep my immune system functional and from falling into a deep well of emotional pain for 6 months (it takes ~3 months after summer ends for me to deplete my vitamin D reservoirs and for my energy levels to begin drooping).

While some folks like Christmas, for me it always spells the beginning of the Season of Doom and Gloom, and I much prefer what happens at the end of the holiday season, after the solstice and as we go into spring with days lengthening (but before the sun god even starts thinking about showing us his hallowed face again.

Which is why I light the Christmas tree on fire every year.

Happy Imbolc!

Rocket Stove

I built a rocket stove. Tallulah helped me cram it full of firewood.

I’m not super pleased with how little flame it produces but that’s the efficiency/vroom tradeoff at work.

The perfect cup of coffe

The perfect cup of coffee is perfectly repeatable. Variables are constrained, and you enter a regime of hunting local maxima. Small and controlled changes to understand their impact on your brew.

There is no perfect cup, everyone has their own preferences on oils and particulate matter and so on. Some folks love the rich mouth-feel of a cup from the French press, other folks like it strong out of the Aeropress, some folks won’t touch it unless it came from a ten-thousand dollar espresso machine lovingly kept in tune all day by a barista with a Master’s of Social Work.

Here’s how I make coffee, in the most repeatable and reliable fashion. Minimal tooling, minimal overhead, minimal complexity, maximal repeatability.


  • burr grinder
  • Clever coffee dripper


  1. Measure out and boil 12 ounces of water.
  2. As soon as the water boils, turn the kettle off, weigh out and grind 20 grams of fresh beans.
  3. Line the Clever with a paper filter
  4. Dump ground beans into the Clever
  5. Pour in just enough water to soak the grounds (this allows the CO2 to “bloom” off, and prevents the grounds from floating and not soaking fully in the brewing medium)
  6. Bloom the grounds for 1 minute
  7. Pour the rest of the water into the Clever
  8. Soak grounds for another 3 minutes
  9. Decant

Simply, perfectly, fool-proof-edly repeatable.

Get 3 Clevers, because they’re guaranteed to go out of production in the next decade, and then you’ll be stuck doing pour-overs again, with all the asinine variance that imposes on the process.