I have some questions about mobile, geo, and platforms

February 7th, 2016

At some point blogs professionalized to the point where they became the province of answers rather than questions. This blog post isn’t an answer, it’s a question.

At the first (and only?) O’Reilly P2P conference I remember seeing Ward Cunningham give a lightning talk on how to ask a question on a wiki (and by extension the Internet), “You write down everything you know, and the Internet will correct your mistakes.” I’ve heard many folks repeat this, cynically, but as I recall Cunningham’s telling it was joyful, what a generative and generous way to ask a question.

I’m a product of the web. I started working professionally in software development the year the web was created, and I’ve been working on the web in one capacity or another for 20 years. It shapes my imagination. And one of the truths on the web is the switching cost is low. I don’t have to discover and download software, learn a new tool, navigate some gated garden of experiences. If I want to use an experience, or switch experiences it’s a simple click away in the universal client that has come installed on every computing device for most of the last two decades. This obviously doesn’t describe either desktop apps, or mobile apps.

Swarm (and Dodgeball before it) briefly had a feature for messaging folks you know who are in your general geographic vicinity. It’s not a common behavior. Especially as we’ve aged. And it probably isn’t a feature that ever makes sense in areas less dense than a major city. And certainly not useful if you don’t know a large number of folks using Swarm. Any way, the feature was removed due to lack of usage. Which as a once-and-former CTO I understand, but as someone with the problem, “I have some rare and unexpected free time and I’d love to see someone who is also at loose ends”, I’m frustrated.

In late high first bubble “portals” arose. Websites dedicated to being able to satisfy all of your needs: weather! sports scores! stock tips! tv schedules! horoscopes! email! what else is there to life! Even pre-Google that was a pretty dumb strategy. Even then switching costs on the Web were low and the portals were doomed.

Now however there are a number of things about portals that feel like they’d make sense. If you’ve managed to get someone to download, install, configure and authorize your native app it feels like the natural pressure would be to accumulate more and more features, your crappy email/stock/horoscope features in the installed app vs a better version in an app you haven’t installed (and can’t discover due to the paucity of our current generation of app stores — definitely pre-Google there)

Why haven’t we seen this? Why do we see something like the Facebook app demonstrate comparative discipline? Why doesn’t Google Maps offer to fulfill all my possible geo needs, the rare behaviors as well as the common ones? (the obvious answers around cost of development are irrelevant in this late-stage stacks era we find ourself in) Why hasn’t Twitter tried turning itself into a portal in its darker hours?

I want the feature of being able to message nearby friends who are at loose ends for the once or so a month that I find myself inclined to use that feature, and the maybe 2-3 times a month I might be able to join someone else’s broadcast short notice plans.

My first thought of course is, if it isn’t going to be part of Swarm someone should build just that app. And now we run into the challenge of high switching costs in native mobile. There is essentially no chance that anyone other than Foursquare will be as effective at getting my friends to use their purpose built app, especially for a use case which just doesn’t come up that often. So scratch that.

Is it possible to build it on top of the web, or somehow recreate the web’s low costs model? It doesn’t seem obvious to me that it is?

Could you string together enough use cases to make a mobile geo platform a thing? A sort of IFTTT for a thing in my pocket that knows who my friends are, and where we’re all located? This feels like an idea from a more optimistic and earlier era (call it Web 2.0)

Maybe you could lower the costs by federating, if you don’t need the app I’m using and yet we can collaborate, that might work? We obviously have a few examples of this (email), but in the last few decades this has been something of an idealist’s fantasy.

Except you know with geo we’ve already agreed on a pretty well defined set of abstractions for talking about the shared space we’re describing, namely latitude and longitude and the planet as it exists today. Platforms are hard, federation has failed significantly more often then it has succeeded, but geo feels like a place it could work.

Of course that’s all a lot of work for a feature that would be trivial for a number of lumbering mass installed apps to add. (thus does conscience does make cowards of us all)

What am I missing? We don’t we see more portal like behavior? Can federation be made to work for real humans? Can platforms overcome the inherent inertia around installing yet-another-app?

(I tried to fit this all into 140 characters, but obviously I failed)

Tagged: Uncategorized

When does conversational commerce make sense today?

January 28th, 2016

Morning coffee, with hummingbirds, Monteverde

If you’re interested (even just professionally) in ecommerce, and you’ve been paying attention, than you’ve been thinking about chat based UIs and commerce for a while. For a quick refresher, you could read the blog post Chinese Mobile App UI Trends.

It all got a little frothy this week (I don’t blame Chris, he’s part of the gestalt), enough that someone asked me my thoughts on it, so I wrote them down.

  1. One of the key promises that seems to be exciting people about chat based UIs and conversational commerce is the ability to route functionality over widely installed communication tools (e.g. Messenger or WeChat) and thereby avoid the hassle associated with building an app, and dealing with Apple/Google gatekeepers. Outside of WeChat however there is no indication that gatekeepers of the communication tools will be any more flexible or less extractive than the gatekeepers for apps. Slack is a bit of darling to hack on at the moment because the API and integration is comparatively generous. (reprising one of the most successful strategies at Flickr)

  2. Pushing a button is always going to be easier than using a conversational interface. Therefore conversational interfaces are going hit their sweet spot when the range of possible actions is large. If I’m doing something simple, transactional and repetitive (like summoning a car), a button is the right UI. (I’ll acknowledge this point is debatable in the context of both home and driving, which while important are arguably growing less so)

  3. Following on the idea of a button being a more straightforward UI, if a given task (or purchase) is something I do frequently enough to be an expert, even if the tasks is complicated, purpose built tools are going to be more powerful, e.g. ordering weekly grocery delivery.

  4. Language is lossy, and trust in new mediums develops over time. I expect that conversational commerce, especially bot mediated, will be dominated by lower price point (sub-$30) purchases in the near future.

  5. On a side note transactions delivered via chat/communication tools will also probably quickly evolve to be much more structured, this isn’t a returning ascendancy of NLP meets command lines.

So what’s the sweet spot in the near future?

We’re looking for casual transactions, at a moderate price point, that aren’t repetitive, and where we aren’t an expert. And we need to be offering this service in a channel which independent of our offering is used frequently enough to be habit forming.

If I was trying to launch something in this area I’d ask myself, “What are the transactions I make, sub-$30, where I still immediately ask someone working at the business for help?”

For more interesting thoughts on building them conversational UIs, this survey by Matt Webb is quite good, especially the Jack Principles from 90s game “You Don’t Know Jack”. (PDF warning)

Finally Daniel’s ideas on “search as a conversation” has affected much of my thinking on search, commerce and communication for years.

Tagged: Uncategorized

Towards an understanding of technical debt

January 10th, 2016

Somerset tire fire

I’ve spent the last few years rather flippantly stating, “Technical debt doesn’t exist.”

What I was trying to say was, “I’m deeply uncomfortable with how our industry talks about this thing, I think it’s probably harmful, but I don’t know quite how to express what I mean. Also, given that everyone seems to be really focused on this tech debt concept, I’m kind of worried that the problem is me, not the term or the industry”.

When I first heard Peter Norvig say, “All code is liability”, that felt closer to right.

The other week I was sitting in the car, waiting for the baby to wake up, and bantering on Slack on the topic of tech debt when my pocket computing device died, and I was forced to finish my thoughts on paper. These are those notes.

Technical debt exists. But it’s relatively rare. When you start arguing with someone about technical debt, you’ll generally encounter a definition like: Technical debt is the choices we made in our code, intentionally, to speed up development today, knowing we’d have to change them later. Hard coding a variable because currently there is no plan to change it is a common example of technical debt. Similarly not modularizing a function.

This is a fairly clear, succinct, and easy to reason about definition, that describes a phenomena that exists relatively rarely. Relatively rare compared to what? Compared to the amount of technical debt we ascribe to the codebases we work on. How then do we explain the overwhelming prevalence of technical debt we encounter when we talk to people about code?

The term is being abused, or at least dangerously overloaded.

“tech debt”: an overloaded term

There are at least 5 distinct things we mean we say “technical debt”.

  1. Maintenance work
  2. Features of the codebase that resist change
  3. Operability choices that resist change
  4. Code choices that suck the will to live
  5. Dependencies that resist upgrading

1. Maintenance work

As technologists we often intentionally or unintentionally use the term “technical debt” to describe to other teams (technical or otherwise) the necessary ongoing maintenance work a codebase needs. In this case the wide recognition and acceptance of the term is useful shorthand for buying breathing room. But it comes at a cost. We end up pathologizing something which is normal, often casting an earlier team as bumbling. A characterization that easily returns to haunt us.

2. Features of the codebase that resist change

When we program we’re writing down a very explicit and exact description of a solution to our current best understanding of the problem we’re trying to solve. Over time both the problem we are trying to solve and our understanding of it will inevitably change, often dramatically. When that happens some percentage of our very explicit and exact description is now wrong, and a liability. Taken to the logical conclusion, given time, every line of code is wrong and a liability that will need to be dealt with.

Therefore the second common meaning of “technical debt” is the features of the codebase we encounter in our work that make it resist change. Examples of features that can make a codebase resist change include: poor modularization, poor documentation or poor test coverage. Just as easily though an abundance of modularization (and complexity) or an abundance documentation, and tests encoding the now the incorrect old behavior can apply a strong downward pressure on change.

A little discussed and poorly understood design goal for code is disposability. Given change, what design patterns can we follow that allow us to quickly expunge incorrect behavior from our codebase? Interestingly it is a much more tractable metric for measuring as opposed to more popular criteria like “elegance”. (a post for another day)

In my experience these features that resist change are one of the most common meanings of “technical debt”. However we often lack self-awareness about it because we don’t tend to think of our work in terms of the life time of the codebase.

An example of maintenance work might be dealing with data that accumulates over time. An example of inevitable change might be that at some point that data that accumulates over time no longer fits on a single physical device. And the resistance will come from how straightforward it is to add a more sophisticated storage solution.

3. Operability choices that resist change

Related to the code choices we’ve made that resist change: what are the operability choices we’ve made in the design of our systems that put downward pressure on change?

If the site goes down every time you make a change, you stop making changes. If you don’t have metrics you can’t deploy changes confidently. Similarly if your tests are flakey, if extensive coordination is required for a release, if you don’t have staging environments, if you can’t run product and operational experiments, if people don’t have access to the information they need to make decisions, if incentives are misaligned etc, etc.

Attempting to improve an environment either under conditions of constant crisis, or in the face of inertia often leads to the energy drained state that programmers attribute to working on systems with large amounts of technical debt. These are failings of operability.

4. Code choices that suck the will to live

Operational inertia and/or crisis isn’t the only phenomena that can lead us to feeling tired and drained. A significant percentage of what gets referred to as technical debt are the decisions that don’t so much discourage change but rather discourage us from even wanting to look at the code. The aspects of the code that suck our will to live as it were.

We often describe this code with the suck-the-will-to-live quality as messy (spaghetti), unmaintainable, or amateurish. Often what we’re describing under the surface is fear and confusion. We don’t understand the code, and when we don’t understand things, as human, we tend to get scared, tired, and angry. Often we find this pattern in teams who’ve inherited a codebase. The code that was originally written by a small tight knit team with a clear vision of the problem is now being worked on by (often much more senior) separate teams working in some degree of silo. What was a productive lack of indirection to the code becomes a poorly considered lack of abstraction resisting change.

Hence the paradox: how is it that a team of brilliant senior engineers need 6 months to clean up after that one early programmer’s weekend kludge job?

It makes no sense, unless you factor in and address the emotional impact of working on certain types of code. (and to be clear different types of code have this impact on different types of people, just another reason to be thinking about having a diverse team)

The other time I see teams struggling with this scared-tired-angry experience is when they’re asked to work on technology that runs counter to their identity politics. I struggle to describe this phenomena with an empathic view point and so I’ll defer, but this recent post on contempt culture is a good starting point.

5. Dependencies that resist upgrading

Finally we use technical debt to describe technical decisions that bind a codebase to a technology that due to the passage of time has become a liability: it has stopped receiving updates, expertise are difficult to find, upgrade paths become convoluted. Often a single dependency pegged to an older technology cascades across your infrastructure holding back important upgrades.

Archaic dependencies are often a symptom that we weren’t able to prioritize ongoing investment and maintenance of the codebase (see #1), and is the thing most reasonable to refer to as technical debt. Even in this often both clear cut and debilitating situation it is still a mistake to characterize the technical debt as a failing, moral or otherwise. We may be in this situation due to a failure of planning, foresight, or expertise. Or maybe someone made a very correct choice to focus on delivering business value knowing that building for the future is only worth worrying about if you’re actually going to survive.

This is an area where focusing on a small number of well known technologies, and being diligent about your new technology adoption process can really pay off.

Why it matters

The first step to addressing any problem is understanding it. When we conflate these 5 different phenomena we confuse ourselves about possible solutions. When we take it a step further and turn these conflations into a judgement on the intellect, professionalism, and hygiene of whomever came before us we inure ourselves to the lessons those people learned. Quickly we find ourselves in a situation where we’re undertaking major engineering projects without having correctly diagnosed what caused the issues we’re trying to solve (making recapitulating those issues likely) and having discarded the iteratively won knowledge that had allowed our organization to survive to date.

The two most common solutions both often make the problem worse:

  1. Declaring bankruptcy and rewriting from the ground up.
  2. Papering over the issue with a layer of indirection/proxy/wrapper

Rather if you are addressing a problem you understand and have correctly diagnosed you can put together a reasonable strategy for iterating and measuring your way to improvement. (and if you aren’t measuring when making changes you aren’t doing engineering, but that’s an essay for a different day)

And finally you should especially worry if your team believes they’re “fixing” or “paying off” technical debt. All code is technical debt. All code is, to varying degrees, an incorrect bet on what the future will look like. You can address issues that are damaging to productivity, operability and morale, but only way to “fix technical debt” is “rm -rf”.

Tagged: Uncategorized

Optimizing for Learning

November 8th, 2015

I gave two similar keynotes back to back in San Francisco in September. One at Github Universe, and one at the Code for America Summit. I felt like I was bit rusty after taking over two years off to focus on the inward facing work at Etsy, but folks enjoyed them.

This is the CfA version:

And some photos from the GHU one:

Tagged: Uncategorized

Five years, building a culture, and handing it off.

August 31st, 2015

After 5 years, the time has come. This is my last week as Etsy’s CTO.

Five years ago I joined a tiny site, with a struggling software product. I wasn’t sure how long I’d stay. I was coming off 5 intense years helping scale Flickr. But I had some theories about engineering I wanted to test, and Etsy had a nascent team with some great folks who were up for the challenge, though a number of the odds were stacked against us.

Theory 1: Nothing we “know” about software development should be assumed to be true.

Most of our tools, our mental models, and our practices are remnants of an era (possibly fictional) where software was written by solo practitioners, but modern software is a team sport.

Theory 2: Technology is the product of the culture that builds it.

Great technology is the product of a great culture. Culture gives us the ability to act in a loosely coupled way; it allows us to pursue a diversity of tactics. Uncertainty is the mind-killer and culture creates certainty in the face of the yawning shapeless void of possible solutions that is software engineering.

Culture is what you do, not what you say. It starts at the top. It affects everything. You have a choice about the culture you promote, not about the culture you have.

Theory 3: Software development should be thought of as a cycle of continual learning and improvement rather a progression from start to finish, or a search for correctness.

If you aren’t shipping, you aren’t learning. If it slows down shipping, it probably isn’t worth it. Maturity is knowing when to make the trade off and when not to.

I had some experience with this at Flickr, and I wanted to see how far you could scale it. My private bet was that we’d make it to 50 engineers before things broke down.

Theory 4: You build a culture of learning by optimizing globally not locally.

Your improvement, over time, as a team, with shared tools, practices and beliefs is more important than individual pockets of brilliance. And more satisfying.

Theory 5: If you want to build for the long term, the only guarantee is change.

Invest in your people and your ability to ask questions, not your current answers. Your current answers are wrong, or they will be soon.

Five years ago, continuous deployment was still a heretical idea. The idea you could do it with over 250 engineers was, to me at least, literally unimaginable.

Five years ago, it was crazy to discuss that monitoring, testing, debugging, QA, staged releases, game days, user research, and prototypes are all tools with the same goal, improving confidence, rather than separate disciplines handled by distinct teams.

Five years ago, focusing on detection and response vs prevention in order to achieve better, more reliable, more scalable, and more secure software was unprofessional.

Five years ago, suggesting that better software is written by a diverse team of kind people who care about each other was antithetical to our self-image as an industry.

Five years ago, trusting not only our designers and product managers to code and deploy to production, but trusting everyone in the company to deploy to production, trusting engineers to meet customers, and trusting teams on the edges to use data to make their own decisions went against everything we’d been taught about how we were supposed run a software development company.

Five years ago, rooms of people excitedly talking about their own contribution to a serious outage would have been a prelude to mass firings, rather than a path to profound learning.

And five years ago no one was experimenting in public about how to do this stuff, sharing their findings, and open sourcing code to support this way of working.

Five years ago, it would have seemed ludicrous to think a small team supporting a small site selling crafts could aspire to change how software is built and, in the process, cause us to rethink how the economy works.

These last 5 years testing the theories, making discoveries, growing a team who has changed how software is built, supporting the flowering of a tech scene here in New York — I’m incredibly proud and grateful to have been a part of it. And I’m proud to have been part of growing Etsy from a tiny, not-always-online website to a billion dollar public company that challenges conventional notions of whether you can be led by values, other than selfishness, and be a successful business at the same time.

The goal was always to build a culture — a culture of learning, a culture of generosity, a culture of values. Culture infects everyone. Successfully building a culture ensures when you leave you can hand your work off to people you trust and they will run the thing without you and make it better than you could have imagined.

And building a thing that can be handed off means … you eventually hand it off.

I took a sabbatical this summer. It was awesome. Burn out is a real and pressing phenomenon in our industry and I was certainly feeling its effects. I spent the summer doing things I enjoy: spending time with family, traveling, meeting new people, talking to smart folks about the future, and writing code. By the end of sabbatical, I felt better. I felt stronger and more present.

So I’m handing it off. To my team, you’re all awesome, I love you and I’ll miss you.

And now it’s time for me to find something new.

Tagged: Uncategorized

More thoughts on the Keyboardio Model 01

June 15th, 2015

I’m super excited about the Keyboardio Model 01, and glad to see their Kickstarer is live (go back it now!), even if I squirm to see my bits in the video. Given that I recorded nearly an hour of video about my experience using the keyboard, I tend to think the best bits of the interview didn’t make it into the their final draft.

Here are some more thoughts

I met Jesse nearly 15 years ago at Diesel Cafe in Somerville, MA when I walked up and introduced myself and asked him about his crazy keyboard setup. I’ve always preferred working out of coffee shops to hermetically sealed fortress of solitude like setups and ergonomics is one of those things that gets compromised. Jesse was already bridging the gap, but 15 years ago the results weren’t pretty. (sort of Snow Crash gargoyle-ish)

Most popular ergonomic keyboard designs are worse then useless, and the ones that aren’t seem to assume you’re very stationary spending 8-10 hours a day in your designated parking space (probably with a fancy chair, desk, dual monitor setup, etc) — they’re big, heavy, clunky things.

And at the end of the day they’re still designed for typing prose.

The Model 01 is different. It’s small enough that I can throw it in my low profile Osprey day pack which is my standard navigating the crowded streets of New York bag, along with a laptop, a book, and a notebook. And it’s lovely from the smooth milled maple, the custom keycaps, the rainbow backlit keys, and the solid metal backing. (that said it isn’t light with all that solid maple and metal)

The Model 01 is a tool for a craft, the craft of programming. When I think about great craft people, I think about how their tools become extensions of their hands. You keep your tools sharp and well oiled, you invest the time into customizing your editor (except for the folks always flitting to the latest and greatest which feels like the opposite of craft), but a keyboard that is a sharp, well designed tools, that’s designed to be customized and repaired, and also takes into account how hands actually work is something new.

I’m sure folks will get value from the Model 01 for tasks other programming. But it’s hard to overstate how poorly thought out the current offerings of keyboards are for programming, with our statistically unusual range of characters, forcing a ton of stretching and reaching, the activities that cause RSI.

I’ve been using the keyboard for about 2 weeks. I’m definitely not entirely comfortable with it yet. For some reason my right hand slips left and I have a habit of hitting the Return key instead of H meaning my sentences look like t
is wit
weird spaces in t

I do love the functions keys, the no reach navigation and bracket keys. The delete and space keys come very natural. Still getting used to reaching for the shift keys.

Looking forward to having time to do immersion as I’m right now flipping back and forth between the Model 01 and my laptop keyboard. It’s been a while since I’ve learned a new layout, so many of the alternative layouts like Dvorak promise typing speed increases which is one interesting. I’m looking for typing sustainability not typing speed.

And despite threatening to, I haven’t hooked the butterfly programmable button up to deploy … yet.

Tagged: Uncategorized

“Engineering capacity for an economic revolution”

April 19th, 2015

“Engineering capacity for an economic revolution”

That’s what my LinkedIn profile says I’m up to if you check it. That means a lot of things. One of the things it means is I believe that software is important, doing software well is important, and when we invest at getting good at human centric, thoughtful, sustainable software development we’re increasing capacity in the system, and we should share what we learn. Internally we call this a spirt of generosity.

I think at Etsy we’re currently doing the best work we’ve ever done in digging into the how and why of software development and leadership, and when I look across the team I see an amazing pool of folks who someday (hopefully far in the future) will flow out and change how this industry works.

But this week I’ve been feeling reflective, and it got me thinking about all the folks who are already out there on that journey.

A totally non-comprehensive list of companies where folks who used to work at Etsy are now either founders or senior engineering leaders (managerial or otherwise) includes:

  • Abacus
  • AltspaceVR
  • First Look
  • floship
  • Handy
  • Hinge
  • Julia Computing
  • Justworks
  • liwwa
  • Radico
  • Ringly
  • Signal Scienes
  • SoundCloud
  • Stripe
  • Tumblr

For a small company with a low turnover rate that feels like a high impact list, and I’m proud of the capacity we’ve added to the system.

If I missed some feel free to add them in the comments.

Tagged: Uncategorized

A little bit on the technical track

January 11th, 2015

I’ve got a post I want to write in response to Dan’s post on the technical track which I think does a decent job of laying out the issues, but doesn’t cover all the challenges and I think is poor on the solutions. Barring significantly more focus and free time dropping from the sky, this clip from a recent fireside chat I did has my recent, if somewhat opaque, thinking on the topic and is the closest I’m going to get to a blog post. 2 minutes and 25 seconds.

(also this is about 90 minutes into the fireside chat sitting on really uncomfortable chairs, going to get those replaced before the next extended QA session we do)

Tagged: Uncategorized

2014 End of Year Booklist

December 29th, 2014

Read Rick’s, Rachel’s and Diana’s year end book list, and got inspired to write up my own.

2014 was a year of extenuating circumstances, the most obvious being the arrival of Anwen 9 hours before it began, but numerous others large and small abound. Reading this year was centered around escape, even oblivion, in short doses, and re-reading featured prominently.

(also I finally forwent the urge to build a Kindle exporter as part of this process and just used clippings.io which did the job well, though seems to have hit a problem after downloading nearly 2400 highlights)

Additionally I’ve probably missed a couple which I didn’t highlight anything in (or read exclusively in dead tree)

  1. The Old Ways: A Journey on Foot by Robert Macfarlane

    Pilgrim paths, green roads, drove roads, corpse roads, trods, leys, dykes, drongs, sarns, snickets – say the names of paths out loud and at speed and they become a poem or rite – holloways, bostles, shutes, driftways, lichways, ridings, halterpaths, cartways, carneys, causeways, herepaths.

    In Ireland there are hundreds of miles of famine roads, built by the starving during the 1840s to connect nothing with nothing in return for little, unregistered on Ordnance Survey base maps. In the Netherlands there are doodwegen and spookwegen – death roads and ghost roads – which converge on medieval cemeteries. Spain has not only a vast and operational network of cañada, or drove roads, but also thousands of miles of the Camino de Santiago, the pilgrim routes that lead to the shrine of Santiago de Compostela. In Scotland there are clachan and rathad – cairned paths and shieling paths – and in Japan the slender farm tracks that the poet Bash? followed in 1689 when writing his Narrow Road to the Far North

  2. The Cuckoo’s Calling (Cormoran Strike Book 1), Robert Galbraith (aka JK Rowling)

    displayed a multitudinous mess of life’s unnecessities.

  3. The Silkworm (Cormoran Strike Book 2), Robert Galbraight (aka JK Rowling)

  4. Startide Rising (Uplift Trilogy Book 2), David Brin

  5. The Uplift War (Uplift Trilogy Book 3), David Brin

  6. Tamsin, Peter S. Beagle

    She’s got a daughter, too, my cousin Barbara, and we were always supposed to be lifelong buddies, but the first time we met, when we were maybe two years old, we tried to beat each other to death with our toy fire engines, and it’s been downhill from there.

  7. The Peripheral, William Gibson

    “Terrorism,” said the rental. “We prefer not to use that term,” said Lowbeer, studying her candle flame with something that looked to Netherton to be regret, “if only because terror should remain the sole prerogative of the state.”

  8. Ancillary Justice (Imperial Radch), Ann Leckie

  9. Ancillary Sword (Imperial Radch), Ann Leckie

  10. Country Driving: A Chinese Road Trip, Peter Hessler

  11. The Slow Regard of Silent Things (The Kingkiller Chronicle, Book 1), Patrick Rothfuss

    You might not want to buy this book.

  12. Memory (Vorkosigan Saga Book 10), Lois McMaster Bujold

    There’s no place like home. I didn’t say there was nothing better. I just said there was nothing like it.

  13. The Winter Long: October Daye #8, Seanan McGuire

    “The day I get a vacation is the day the world ends,”

  14. Chimes at Midnight: Book Seven of Toby Daye (October Daye), Seanan McGuire

    “A knight of the land Courts asking a Duchess of the Undersea to save a King of Cats,”

  15. Fool’s Assassin: Book One of the Fitz and the Fool Trilogy, Robin Hobb

    “I will always take your part, Bee. Right or wrong. That is why you must always take care to be right, lest you make your father a fool.

  16. The Girl Who Circumnavigated Fairyland in a Ship of Her Own Making, Catherynne M. Valente

    I’m afraid the whole thing moves around according to the needs of narrative.

  17. The King’s Blood (The Dagger and Coin Book 2), Daniel Abraham

    The day you throw me in a ditch and take control of the company?” “Still not today.

  18. The Dragon’s Path (The Dagger and Coin Book 1), Daniel Abraham

    if everyone benefits, you’ve overlooked something.

  19. The Widow’s House (The Dagger and Coin Book 4), Daniel Abraham

    “The world is burning. Anything that doesn’t end in ashes is worth doing.

  20. Captain Vorpatril’s Alliance (Vorkosigan Saga Book 14), Lois McMaster Bujold

    “The ones still inside, you’ll want to commend. The ones outside, those are the ones I’d promote . . .”

  21. Komarr (Vorkosigan Saga Book 11), Lois McMaster Bujold

    He’s not even a mad scientist. He’s merely a very upset engineer.

  22. A Civil Campaign (Vorkosigan Saga Book 12), Lois McMaster Bujold

    You won’t regret this seemed a much too optimistic statement to add.

    “What about acts of ineptitude?” “A gray area, and don’t tell me you haven’t lived in that twilight before.” “Most of my life, sir. Not that I haven’t leaped up into the blinding light of competence now and then. It’s sustaining the altitude that defeats me.”

  23. Rogues, George R.R. Martin and Gardner Dozois

    the very image of a person with just the average amount to hide.

  24. The Rhesus Chart (Laundry Files Book 5), Charles Stross

    “Hurry up! We can’t stop here; this is black pudding country!”

    The five stages of bureaucratic grieving are: denial, anger, committee meetings, scapegoating, and cover-up.

  25. Farthing (Small Change), Jo Walton

  26. The Wise Man’s Fear (The Kingkiller Chronicle, Book 2), Patrick Rothfuss

    “It’s the questions we can’t answer that teach us the most.

    “Maps don’t just have outside edges. They have inside edges. Holes.

    It was the same scolding any child receives. Stay out of the neighbor’s garden. Don’t tease the Bentons’ sheep. Don’t play tag among the thousand spinning knives of your people’s sacred tree.

  27. The Name of the Wind: The Kingkiller Chronicle: Day One, Patrick Rothfuss

    “I thought you would be older.” “I am,”

  28. Be Slightly Evil: A Playbook for Sociopaths (Ribbonfarm Roughs 1), Venkatesh Rao

    We are low-entropy creatures trying hopelessly to swim upstream in a universe that’s gradually winding down towards a maximum-entropy heat death.

  29. My Real Children, Jo Walton

    She had never felt older than those years when the children were small and so demanding of her attention. She had felt it a new lease on youth when they were grown and gone,

    In November of 1966 there was a flood in Florence, killing six people and damaging some property. Fortunately the weather computers had predicted it well in advance, so most people had evacuated and most works of art were moved to safety. Some frescos were damaged. Pat wrote articles about their restoration and sat on a committee to raise money for it.

    Just the cold contingent universe where things happened for random reasons nobody could understand.

  30. Capital in the Twenty-First Century, Thomas Piketty

    Expert analysis will never put an end to the violent political conflict that inequality inevitably instigates.

  31. Words of Radiance (The Stormlight Archive, Book 2), Brandon Sanderson

    “If I toss something upward, it comes back down.” “Except when it doesn’t.” “It’s a law.” “No,” Syl said, looking upward. “It’s more like . . . more like an agreement among friends.

  32. The Way of Kings (The Stormlight Archive, Book 1), Brandon Sanderson

    Shallan moved her eyes down to the bottom of the page where—separated by a line—the undertext was written in a small, cramped script. Most books dictated by men had an undertext, notes added by the woman or ardent who scribed the book. By unspoken agreement, the undertext was never shared out loud. Here, a wife would sometimes clarify—or even contradict—the account of her husband. The only way to preserve such honesty for future scholars was to maintain the sanctity and secrecy of the writing.

    When he passed, the timid grass pulled back, but the fingermoss was bolder. The clumps would only pull into their shells if he tapped the rock near them.

  33. Moneyball, Michael Lewis

  34. Liar’s Poker (Norton Paperback), Michael Lewis

    Why did investment banking pay so many people with so little experience so much money?

    By the mid-eighties, however, all things black, yellow, and female have disappeared from the photographs. There isn’t a trace of anything but white men in the annual reports.

  35. Healthy Sleep Habits, Happy Child, Marc Weissbluth Md

  36. Flash Boys: A Wall Street Revolt, Michael Lewis

    Every systemic market injustice arose from some loophole in a regulation created to correct some prior injustice.

  37. The Hard Thing About Hard Things: Building a Business When There Are No Easy Answers, Ben Horowitz

    “When we drove off the cliff, we left no skid marks.

    The people who stay will care deeply about how you treat their colleagues.

    Yes, of course. The reason that people leave our service and don’t come back is that we have not been sending them enough spam. That makes total sense to me, too.

    Shock is a great mechanism for behavioral change.

    Hiring scalable execs too early is a bad mistake. There is no such thing as a great executive. There is only a great executive for a specific company at a specific point in time.

  38. Half-Off Ragnarok: Book Three of InCryptid, Seanan McGuire

    Nature’s not talking, possibly because even nature realizes that giving perfect camouflage to apex predators is sort of a dick move.

    (Australia. The only continent designed with a difficulty rating of “ha ha fuck you no.”)

  39. Turn the Ship Around!: A True Story of Turning Followers into Leaders, L. David Marquet

    Most of what we study, learn, and practice in terms of leadership today follows this leader-follower structure. This model has been with us for a long time. It is pervasive. It is the structure depicted in The Iliad, in Beowulf, and in other Western epics. But this model developed during a period when mankind’s primary work was physical. Consequently, it’s optimized for extracting physical work from humans. Those who take orders usually run at half speed, underutilizing their imagination and initiative. While this doesn’t matter much for rowing a trireme, it’s everything for operating a nuclear-powered submarine.

  40. The Privilege of the Sword (The World of Riverside Book 2), Ellen Kushner

    “I do not make the rules,” he said creamily. “This annoys me, and so I comfort myself by breaking them.

    Getting the long sword into the soft scabbard was a bit like getting a bootie onto a baby’s foot: neither was very interested in helping,

    And suddenly, as the night air turned cold and the day sky burned a bright and gallant blue, the world was full of apples. The air smelt of them, sharp and crisp, then underlaid with the sweet rot of groundfall.

    Serious sword-practice made me forget to think in words, so that I didn’t always understand when people spoke to me.

  41. What’s Going on in There?: How the Brain and Mind Develop in the First Five Years of Life, Lise Eliot

  42. Influx, Daniel Suarez

    “Mankind was on the moon in the 1960s, Jon. That was half a century ago. Do you really think the pinnacle of innovation since that time is Facebook? They designed the Saturn V rocket with slide rules.

  43. Abaddon’s Gate (The Expanse Book 3), James S.A. Corey

    She had a sudden vision of Jesus, who’d asked His disciples to keep doing this in remembrance of Him, watching her little congregation as they floated in microgravity and drank reconstituted grape beverage out of suction bulbs. It seemed to stretch the boundaries of what He’d meant by this.

  44. Caliban’s War (The Expanse Book 2), James S.A. Corey

    Owning your own racing ship wasn’t even wealth. It was like speciation.

  45. Leviathan Wakes (The Expanse Book 1), James S.A. Corey

  46. The Tyrant’s Law (The Dagger and Coin Book 3), Daniel Abraham

    “Job is we kill a goddess and save the world. Let’s not complicate it.”

    The citizens of the city wrapped themselves in coats and cloaks, scarves and mittens, until they all seemed part of a single unified race of the chilled.

    “Good that we have the powers of chaos and madness on our side sometimes.

  47. The Golem and the Jinni (P.S.), Helene Wecker

    Gradually the well-kept brownstones gave way to Dutch clapboard houses, their trellises heavy with roses.

    He supposed that eventually Matthew would grow older and lose interest, and take his place with the feral young men who slouched on the neighborhood stoops. Or—even worse—he’d become just another streetcar rider, dull-eyed and unprotesting.

  48. The Boy Kings: A Journey into the Heart of the Social Network, Katherine Losse

    Logging on to Facebook that first day, in retrospect, was the second, and to date the last, time that any technology has captured my imagination.

    “California is a place in which a boom mentality and a sense of Chekhovian loss meet in uneasy suspension; in which the mind is troubled by some buried but ineradicable suspicion that things better work here, because here, beneath the immense bleached sky, is where we run out of continent.

    While they abhorred the idea of being a wage slave, the young men of Silicon Valley were not trying to tear down the capitalist system. They were trying to become its new masters.

  49. Neptune’s Brood, Charles Stross

    Which means I need to brief you on the politics of mermaids.

  50. The Everything Store: Jeff Bezos and the Age of Amazon, Brad Stone

    “How do you plan to handle the narrative fallacy?”

    “When a company comes up with an idea, it’s a messy process. There’s no aha moment,”


remembered at least one book missing from above

Half a King, Joe Abercombie. YA but looking forward to the sequel

Tagged: Uncategorized

Annoyed by “The Peripheral”

November 11th, 2014

I enjoyed reading William Gibson’s latest, as I’ve enjoyed reading nearly everything he’s written, and as I anticipate enjoying seeing him and James Gleick speak tomorrow night.

However as I finished “The Peripheral” the overwhelming sensation I was left with was annoyed.

Below are spoilers.

There were a number of moments that bothered me, but in 3 quick bullets.

  1. The literature didn’t need another implacably evil bearded Arab villain. Even is he is half Swiss.

  2. The literature didn’t need another heroine whose happily-ever-after is signified by getting pregnant.

  3. It goes against everything I look for in a Gibson novel to basely indulge a maybe-the-future-will-take-pity-on-us-and-save-us fantasy. It left me with the sensation that the author is deeply scared, which while reasonable, isn’t what I expect from the typewriting steely eyed prophet of the future.

Lots of good moments, and interesting world building, again the near future work, especially the near future rural poverty in a cheap printing/fabbing future was interesting.

Tagged: Uncategorized

Pumpkin Pudding

October 26th, 2014

Back when my dad drove his big yellow truck, delivering food for the food bank, he sometimes dropped me off at the Live Oak Senior Center industrial kitchen.

And every October I remember peering over the edge into the bowl of an industrial stand mixer taller then I was at a truly giant vat of pumpkin pudding.

And while I’m telling myself this pumpkin flavored Chobani yogurt is kind of a pumpkin pudding, it really isn’t.

Tagged: Uncategorized

“Wolf” narrative considered harmful (also biologically unlikely)

September 2nd, 2014

Rands’ “The Wolf” post floated across my Twitters this morning. My immediate reaction was that it was misguided to the point of being potentially harmful. That was 16 long hours ago, but I’m going to see if I can pull the strands of that thought back together.

Michael mentions he’s had the fortune to work with several “wolves” in his career. I’ve worked with a bunch, including teams compromised of nothing but (I guess you’d call that a “pack”). The problem is the post plays into the fantastical narrative of the lone wolf. But here is the secret. No one is a lone wolf. Or at least not completely. “Wolf” is the story that many people want told about them, and that on good days they tell themselves, but in the dark of the soul, we all need help and have doubts. By celebrating the narrative construct rather then reality you perpetuate the dangerous tendency towards isolation that is inherent in this archetype. I’m playing the long game, and in my perspective sooner or later one of two things is going to happen to your isolated “wolf”: either they’re going to run into a hard enough problem that they’re going to need more help then they know how to ask for, or in their bid to maintain their Harvey Keitel-like facade they’ll eventually isolate themselves from new challenges and dead end their usefulness.

So let’s pull back the curtain a bit and understand the archetype:

  • they’re passionate about solving hard problems, and generally they have a few favorite problems they like to gnaw on. That deep personal investment in the solution is what pushed them to develop their skills past the point of their peers in the first place, and also where they find the personal clarity they need to opt out of the clarifying bureaucratic structure of the larger org. With that passion can come narrowness, often expressed as hyper rationality. Pro move: keep supplying hard problems off center of their primary focus.

  • they’ve developed a toolkit. A preferred language, set of scripts, investigative processes, checklist, whatever. And they come back to it over and over again. That disproportionate impact they have, it isn’t magic, its craftsmanship. They’ve mastered their tools to the point where the tools are an extension of their hands. They might see everything as a problem to be solved in Lisp, or via experimentation or machine learning or wireshark. They might have written their own text editor that lacks features what you and I consider non-negotiable, but doesn’t seem to impair their progress. The best of these folks literally generate tools, turning their own idiosyncratic but highly effective approach into tools that the wider org can use. Pro move: create a space and a culture that celebrates investing in tools. Everyone should hone their craft, and a few great tool builders makes everyone if not a 10x programmer, maybe an 8.5x.

  • they’re convinced they’re right. Or at least, they’re convinced often enough that they’re right that they appear convinced all the time. A banal outcome of this is hill climbing. The toxic outcome is stagnation with a hugely talented individual stuck because they don’t know how not to be right. You owe your high impact individuals to surround them with people who challenge them. If they’re really a “wolf” they’re likely to try to wriggle out of this. Pro move: they care about impact, so make the road to impact littered with smart challenging people.

  • the “well defined IC track for non-managers” is probably the project I’ve found most elusive and difficult to execute on. Rands gave me my first key insight into it late one night after much tequila: you have to make sure the role comes with super powers. Management comes with super powers by default, you have a team to force multiply you, and by convention, most corporations share information with managers at a greater rate then ICs. I haven’t found a one-size fits all approach to super powers, but a thing we’ve done is make sure that while the “senior eng org” (super powered engineers) have their own discussion space, they can also ease drop on manager discussions if they so choose.

If you’ve been maneuvered into calling someone “a wolf”, then I hypothesize you’ve let someone who could be transformational to your organization off the hook too easily. Imagine instead of a wolf you had a Intrinsically Motivated Full Stack Product Hacker.

Tagged: Uncategorized

AWS Redshift: “Missing newline: Unexpected character 0x4f found at location 3”

July 21st, 2014

I was trying to bulk load some data into Redshift this weekend using the copy command and I got this error message that no amount of Googling turned up a hit for. It was particularly strange given the hex value specified was for that exotic character the capital-O (as is OMG).

stl_load_errors listed the offending raw_line as being “.LZO”.

Anyway, this particular error message is documented here for the next person who forgets to toss lzop on to end of their copy command and tries to Google it

to be explicit you’ve got:

copy some_table from 's3://some_bucket/data.lzo' CREDENTIALS 'some_credentials' delimiter '\s' timeformat 'auto';

and you need

copy some_table from 's3://some_bucket/data.lzo' CREDENTIALS 'some_credentials' delimiter '\s' timeformat 'auto' lzop;

Tagged: Uncategorized

Communication Skills

May 17th, 2014

A great deal of negativity has been directed lately at the practice of hiring for culture fit. I find this troubling because its a central tenet of Etsy’s hiring. This pithy insight was shared with me yesterday by one of our senior most engineers, and I slept better last night.

“As time passes, I focus more and more on interviewing for communication skills over technical skills.

Even the most brilliant engineer can get off course. But a great communicator might move slower, but they will always be moving in the right direction.”

(actually I didn’t sleep better last night, but that was the baby’s fault, not anxiety about contributing to oppression.)

Tagged: Uncategorized

Shell trick for the perpetually distracted

May 12th, 2014

Manager time is the schedule that sub-divides your day into 15 minute blocks. It’s the internal clock in your head that knows when you’ve reached the end of a 30 minute meeting without checking the clock. It’s a trained form of ADHD, and it’s the opposite of maker time.

One of the things I find I need in manager time is a quick way to regain my context. I’m way behind on adopting a proper getting things done framework, but it wouldn’t help with this one thing I wanted to post about today anyway.

I keep two copies of most of the Etsy code base locally. The one I use constantly, which is for quick reference when a technical question comes up, and the one I use almost never which is where my various half hearted attempts to compress a quick hack into a 15 minute block live.

git gives me the tools to make it very painful when I get confused about which directory I’m in.

So in prep for hoping to finish up a quick hack during hack week (didn’t happen) I shaved a small attention starved yak. Namely changing the color of my terminal when I’m in workspace, which is the hacking directory. It’s a parallel to prompt hacks I’ve had for years that change based on host, but it uses osascript (command line AppleScript) for simplicity in scripting Terminal.app

set_bg () {
  osascript -e "tell application \"Terminal\" to set current settings of window 1 to settings set \"$1\""

function cd {
  builtin cd "$@"
  if [[ "$pwd/" =~ /workspace/ ]] ; then
    set_bg "Red Sands"
  elif [[ "$old_wd/" =~ /workspace/ ]] ; then
    set_bg "Basic"

Override cd, if we entering workspace, use the profile “Red Sands”, if we’re leaving workspace switch basic to “Basic”.

Tagged: Uncategorized

Explaining to the phone how to be a CTO

December 5th, 2013

The other morning I was sitting stuck in traffic … which an odd experience in New York, but for a few reasons I’ve actually been driving to work lately, profoundly surreal as a New Yorker, and not recommended.

Back to the story.

I was sitting stuck in traffic, and the last thing I’d seen before being stuck in traffic was the question, “How do I learn enough to become a good CTO?”

So below is the first blog post I ever tried to write via Siri. I thought about leaving in all the misidentified words, but the point wasn’t comedy. It still came out quite brief and almost certainly incomplete and incoherent:

It’s a leadership role, which means you need to develop a theory of leadership. It’s a spiritual role which means you need to be able to speak and write. It’s a role that fundamentally trades on credibility, so having been part of a team that did something that actually worked is key both for the reflected credibility but also because it’s important to have the certainty about what something working feels like.

It’s a role we learn by trial and error so you need to find a trusted community to compare notes with. You need to develop an empathy for the customer and interest in the business so whatever interdisciplinary pursuits do that for you is part of your training.

Finally being right more often then you’re wrong helps, but is less critical then you’d imagine.

It isn’t the 1st engineer either literally or metaphorically though that person is sometimes called the CTO. It isn’t the engineering manager in chief, though often you end up doing that job as well (but that’s a VPofE).

Software development scales non-linearly with clarity and excitement so your job is removing unnecessary ambiguity and drag. holding the narrative of what engineering means at YourCompany.

That’s how you learn to be a CTO.

And now that I’m typing this in to a textbox here on web, I’ll add, get good at managing your psychology

Tagged: Uncategorized

Slow Twitter

November 17th, 2013

My brother tweeted recently, “making sure I have a working implementation of Logo is part of setting up the nursery, right?”. I feel the same way about building the just the right photo sharing site and Twitter client. But I’m RESISTING. And I’m very proud of myself for it.

The Twitter client I want (and have wanted for years, but I want it more now) is at least fairly straightforward, and I’m hoping someone has either already built it or will get inspired. I think of it as “Slow Twitter”.

Twitter works great when you’re constantly checking it. Or when you don’t follow that many folks. But once you’re distracted you start missing stuff. In fact you only see the people who tweet constantly. And they’re often the most banal.

What I’m looking for is an app (and I include websites in that word, I’m old fashioned) that shows me who has tweeted since I lasted accessed it, with access to what they’ve tweeted, sorted in reverse order of the likelihood that they would have tweeted during the period of elapsed time aka probably sorted by ascending rolling average tweet frequency.


Tagged: Uncategorized

5 Approaches to Handling Bugs We Tried Before Our Current Approach

November 1st, 2013
  • don’t write bugs
  • once you write a piece of code you own it forever and your top priority is fixing the bugs in it
  • assign teams to own every section of the site, allow them to make sure no unapproved changes happen to those areas
  • have a bug fixing team
  • delete all code over a year old, switch to a new language, rewrite it
Tagged: Uncategorized

A quick knot in a few Flickr threads

October 19th, 2013

Paul Mison’s Maximum Viable Product feels like a more clever (if more cheeky) way to get at what I was trying to say with my notes on Flourishes and Minimal Competence (and Competence found), and threads the needle of recent 4up from Aaron and The Flickr We Lost from Dan.

I don’t have much conclusion as much as a sense there’s a thing there that I’m still trying to get my hands around.

The other piece I wonder at about this is the role of sophisticated testing and measurements and being the user.

I hear that engagement is way up in the wake of Flickr’s changes, and I know I’ve been gone a long time and things have probably changed, but I can’t help but be struck how unsophisticated we were at measuring things at Flickr — in part because we were a tiny team, and in part because we relied on our instincts. It makes me wonder what’s guiding the current work.

Tagged: Uncategorized

Paths to Production Confidence, Part 1 of N

August 8th, 2013

This started out as a short note explaining the unifying theory behind the Etsy development practices. Then it got out of hand (see also, Mark Twain’s “If I had more time…”). As such I’ve made it “Part 1 of N”, where “Parts 2 .. N” will cover the actual practices and how they relate to the philosophy.

Why Write This?

We believe in being as open as possible about how we develop and run Etsy — our current best theories, learnings, practices, and tools. Given that openness I often get questions about the hows and whys of given subsets of our engineering practice, e.g. “How do you do testing? How do you know you have enough testing?”. Or monitoring or deployment or what not.

At times, it can be tricky to do the questions and answers justice when taking them piecemeal, because underlying them all is a single shared philosophical premise, that isn’t necessarily obvious. And while I tend to be a pragmatist, favoring the rough and ready over the theoretical, without understanding the theory you can’t reason correctly about trade offs. This post is an attempt to surface that underlying philosophy and the practices it informs.

The theory is a theory of change, and the philosophy is about finding paths to move from risk to confidence.

A Theory of Change

Etsy is in the change business. As are, definitionally, all startups, nearly all businesses, and most human projects. We’re attempting to add new capacities to the world, and influence behavior around them. And we’re attempting to do it in an uncertain and complex[1] environment; we neither know the exact recipe for success, nor we do expect that recipe to stay the same over time. In fact, as a startup, we believe that our ability to respond to a changing environment is the key success factor for our engineering organization. It’s natural to read that sentence and think of change in terms of product changes, but more prosaic examples might include the ability to add new server capacity when the site slows down, or to replace a hard drive when one fails.

But change is risky. This is something most of us believe intuitively, and it’s worth examining the sources of risk in change.

Why is change risky?

As humans and practitioners why do we associate change with risk? Doing new things inherently contains the risk of doing the wrong thing.

We may for example have reached a ready state in our project. Through a combination of good luck and planning we find ourselves running a system that we understand sufficiently to keep running indefinitely, while a change would implicitly contain the risk of moving from a state of working to a state of not working. Steady state systems are so rare and so often illusory, it’s almost not worth mentioning except we’re fervently entranced with the possibility. Generally the illusion of steady state simply means the needed changes are non-linear, and often the cost of ignoring them will be high.

More practically, very few of us are employed to maintain systems in a stable environment. Even if we hold the pieces we control constant it’s unlikely that our systems will remain stable forever, at which point action is required. Still change is often associated with surprise in a system that hadn’t previously surprised us, and surprise is definitely risky.

The second reason change is risky has to do with how we think about causality, intention and culpability. While we can agree that the ability to choose not to make a change is an illusion, often fear leads people to approximate avoiding change, by avoiding making choices. If I personally avoid making changes to a system and instead wait for outside pressure to force change, or if I simply play the odds and hope that disastrous failure happens rarely enough that it won’t happen on my watch then I can avoid the personal risk of being labelled the root cause of failure. Forced choices avoid the necessity of stating a hypothesis before acting, thereby reducing significantly the opportunity to be personally wrong.

Software development is a complex system existing as it does at the intersection of people, systems, good intentions, confused and changing goals, and overly literal state machines. Past behavior isn’t always an indication of future behavior, and humans are terrible at reasoning about complex systems. As such we’re unlikely to know or have good visibility into whether we’ve reached a steady state and our hypotheses are likely to be wrong. In this uncertain and complex environment we initiate change only when the cost of not making a change overcomes the fear of making it. (e.g. “The server is down” or “You’ll be fired if this feature isn’t done by April 1st”)

As an industry this means though we’re in the change business, often we aren’t very good at it, and we avoid it out of fear.

Different groups attempt to address this tension by:

  • raising the cost of not making change (“you’ll be fired”)
  • distributing those costs broadly (this is one of the key functions bureaucracy and process serve)
  • gaining confidence by addressing the fear

We see Etsy’s engineering practices as spectrum of tools for increasing our confidence in our ability to make change.

Going back to the opening idea of this post, the attempt to answer a question like, “How much testing do you do?”, the answer becomes, “Enough to gain confidence. But testing is just one of the tools we use to gain confidence, so less then a strong testing shop might.” Similarly if someone asks, “How much monitoring is enough?”, the answer is, “We add monitoring until we feel like it gives us confidence, and we’re comfortable striking an 80/20 balance, particularly upfront, because we’re confident if we don’t have the balance right we have other ways of finding out.” In fact how many and how much confidence boosting techniques you need is situational, and depends on how risky your change is. Which speaks to another fundamental piece of our process, small and iterative changes.

Hopefully that starts to explain why, while I think our testing infrastructure (with its try-servers, “Bobs”, static analysis, integration tests, and quality metrics) is awesome, just telling you how we do testing isn’t necessarily going to be useful. Or perhaps it just speaks to my personal penchant for holistic post-modern explanations.

So given a theory(-ish) of risk, change and confidence, what’s the philosophical premise we derive to underly our development practices:

To be able to consistently deliver the kind of resilient and ongoing change the business requires, we deploy a spectrum of confidence gaining techniques.

Or jokingly what we call, “Making failure cheap and easy.”

Before moving on it’s worth calling out that the goal is NOT to be careful. The goal is to be confident. Careful would imply we’re trying to avoid the risk which is fundamental to the change we’re trying to make. Attempting to avoid risk often leads to paralysis, favoring the short term risk avoidance while compromising long term goals. Instead confidence implies you believe to the best of your ability that you understand and have mitigated the risk involved in your change, and are now going to act.

Now, with a little shared theory and philosophy, what does that spectrum of confidence gaining techniques look like?

Our Paths to Confidence:

  • Small and Frequent (and Iterative)
  • Testing
  • Ramp Ups
  • Controls
  • Default Access to Open
  • Monitoring, Metrics and Anomaly Detection
  • People / Culture / Brains

Each of which I’ll talk about in subsequent future posts.

1. Complex systems as defined as something that has many diverse, interdependent, adaptive and connected parts points to the uncertainty. Small perturbations can produce large results, and those results could be failures or successes, but in either case: the potential for surprise is high.

Tagged: Uncategorized