The Popular

If you’re unaware, I’m a computer programmer. Often, while programming I’ll have things pop into my head. Sometimes related to programming, but often completely unrelated. (Hence the title “while {coding}”.)
One of the things that pops into my head often (and that I’m able to articulate fairly well) is the process of hiring and/or being hired as a programmer. These posts have seen the most traffic, by far:

How To Hire Me (or any other programmer for that matter) And the follow-up. This went insanely viral, especially among recruiters. I received emails about that post (and the published article that came out of it) for about 4 years.

A Tale of Two Interviews (a two-part story) compares two different interview experiences I had back in 2009. This was popular on Hacker News but I don’t think it reached much further than that. If you’re a programmer or other technical professional you probably have similar stories to tell.

But I’m Most Proud Of

My call to Play, which I still read from time to time, especially when feeling stressed.

Whither Bob Grossblatt. Not only did I write a heartfelt piece, it resulted in a nice correspondence with the man himself.

My Not-So-Hyper Loop. I worked on this for the better part of a week during a research hiatus when I should have been working on my research project! But no one knew what the Hyperloop was yet and it was just fun to imagine and do a little napkin math for a semi-practical application. No regrets.

The Cavity Model of Software Development. Every now and then I have an idea that I struggle to explain clearly. This is probably one of those… or it’s a terrible idea. I’m fine with either. ¯\_(ツ)_/¯

Projects Over the Years

These are not thrilling posts, but they document the kind of things that pop into my head and occasionally make it into code.

Swisstime, a tool to help with time profiling in Python. Kind of like “println” debugging, but for profiling.

Wheeler, a programming language project which is still ongoing (for almost 10 years now) and might be for the rest of my life. RIP my other side projects.
(The language has evolved a lot since that early post but it is still very similar to what I described. Look for a new version in the next 6-180 months!)

Virtlib, a community technical lending library, which amazingly got strangers to actually lend books to each other, but mostly aided and abetted my perma-borrowing two or three books from my fellow Portland geeks.

Snippetology, allowing Django to do some CMS tricks.

Twimmetry, just a terrible, terrible hypothesis that Twitter user “realness” hinges on ratios of following/followed. At least the Twitter API was fun to play with (when I wasn’t crying).

ResearchedFacts.com which I thought was funny and I guess it kind of anticipated the “post-truth” era. I took it down because I was sure someone was going to sue me or die or both.

Odds n Ends

I’ve written a lot more. Not everything is a gem, but there are some useful thoughts about various things like managing expectations, conquering hard problems, and the importance of persevering.

That’s the intro. Please feel free to reach out via email or twitter, or pull request. See my main page for details.

Thank you for taking the time to visit my site and read this far. It is greatly appreciated.

Matt Youell
Portland, Oregon
October 2018

A theme helps you steer as you move along the road of your life. You don’t steer all at once – steering is a series of little adjustments that you make as you go down the road. A theme serves the same purpose. As you live your year, think about your theme from time to time and steer accordingly.

Here are some guidelines to help you:

• Have only one theme at a time. Maybe you set a theme for a year, or a season, but it is always one at a time.
• A theme is not a goal. You can fail at a goal. You can’t fail at a theme. “Lose 50 pounds” is not a theme.
• Keep your theme a little open-ended. It shouldn’t be too specific. “Fewer calories in” might be a good theme if that’s all you really care about this year, but it’s probably too specific. “Quality over quantity” might be a better theme and might be more broadly applicable to your life.
• Avoid overused idioms. “Watch what you eat” might not be particularly useful. If it sounds tired when you say it in January, think of how tired it is going to sound in September.
• A theme isn’t there to get you excited. It’s there to help you think. You may find that the luster quickly wears off of an exciting theme.
• Avoid pithy sayings. You don’t want a theme like “You can do it!”. On the other hand, “Go for it!” might be a good theme for someone who is trying to break out of their shell and try new things. Only you know what makes sense for you.
• Don’t force it. If you’re not satisfied with your theme, pick another one. You need your theme to stick with you all year long.
• There is no wrong theme if it helps steer you where you want to go.

Themes are just one tool in the toolbox. I’ve had various levels of success with themes over the years. Some years they make more sense than others.

The first year I tried setting a theme, I went with “Commitment and Completion”. It made sense for that year. I knew I was getting married and starting a family. But I also knew I was having a hard time showing up places and completing things I’d set out to do. That theme reflected not only the change in my life but also an ongoing struggle to finish what I started. That year ended up being the best year of my life. Not only did the planned things happen, some wonderful unplanned things happened too, largely because I was showing up and following through more.

The next year I just couldn’t figure out a theme for myself. I tried to force it and eventually gave up. That’s just how it is. Some years I just can’t think of a theme that ties enough threads of my life together to be helpful. What’s great about a theme is that it either serves you or it doesn’t. It’s not a thing you fail at. And if you find one theme that serves you well one year, who is to say you can’t use that same theme again?

This year the theme I’m trying is “Capability”. What’s yours?

P.S. – Was this useful? If so, I have a product coming later this year focused on getting more out of yourself. Interested? Add your email here:

“NEXT YEAR???”

My heart would ache. How was it possible to be OK with that? Who was that patient? I couldn’t imagine.

I now do this every day. I set out on an adventure – maybe to create something, maybe to fix a problem. Hazards and roadblocks befall me. Software conflicts, server configuration, network connectivity, meetings.

And at the end of my day what I set out to do isn’t done but… the path seems clear for tomorrow.

I have a new monitor. It’s an LG 29UC97C-B, a 29–inch, curved, “ultrawide” monitor. It cost $500. Getting it was a tiny adventure in time and space.

I really, really, REALLY wanted the LG 34" but my 2012 Macbook Air just can’t push that many pixels.

I expected Apple might have a nice (if overpriced) offering. Not really. They are still selling their 2011-era 27" Thunderbolt Display which was kind of unimpressive when it first shipped. At $999 it’s almost what that dreamy 34" LG costs! No deal.

Apple does have some sweet displays on their Retina iMacs though. An interview with Marco Arment talking about his iMac got me thinking maybe I was on the wrong track. Maybe I need a new computer. So I went shopping. I had a $4k fully-loaded iMac in my Apple shopping cart before my inner fiscal-conservative woke up. Nope, nope, nope!

Back to monitor shopping, I began to see a divide: ultrawide monitors vs. 5k monitors. So many pixels. But I had my magic numbers: 2560×1600. That’s the max my Mac can drive. With that in mind I ran across a review for the Samsung S29E790C, a 29″ curved monitor running 2560×1080. I almost bought that on the spot. It’s well reviewed and looked good in the YouTube videos I saw. Then I spotted the LG. Not a lot of info was available. The YouTube reviews I found were in Russian and Mandarin. There was exactly one review on Amazon. Risky!

After some hand-wringing I decided that if I want the big LG in the future, I should try the modest LG right now and see what happens. And that’s what I did. I ordered through Amazon but it was drop-shipped directly from Korea. Expedited Shipping was included and it was FAST. I got my monitor less than 48 hours after ordering. (I was lucky. Estimated arrival was anywhere between 2 and 5 days.) The only real hassle in the entire process was finding the right cable for it after it arrived. My Macbook Air has a Thunderbolt port (aka Mini DisplayPort) and the LG has a regular DisplayPort connector. The local Apple Store couldn’t help me so I had to find one on my own. A day later I was up and running.

So how is it? I’ve only had it for three days but so far it’s been incredible. I’m still getting used to how to work with it. I’ve been hunched over a 13" laptop for the past three years. I’ve found that virtual desktops are actually MORE necessary with this much space, so I’m learning how to manage my Mac all over again. I suspect that the novelty will wear off pretty quickly, but I do think my productivity is going to improve. I’ve built up a lot of habits on a small screen that I need to unlearn. I’m always running a text editor, a terminal window, and a web browser and switching constantly between them. I can already tell that I’m losing a lot of time tabbing between windows which could be sitting side by side. There’s probably a lot more to learn.

Measuring how long code takes to run is not exactly unexplored country. Still, nothing I found seemed to fit my needs. A profiler was overkill. Other timing tools wanted to measure one thing at a time. I wanted to run my code and at the end see a tidy little report of times for each chunk of code. Here’s what the output looks like:

————————————————–
Time in seconds
————————————————–
1.095976 | Startup
0.117262 | Load input image
0.018422 | Build masks
0.117944 | Populate layers
0.015563 | Build layers
0.019718 | Tone down green
0.092624 | Convert to floats
0.341209 | Resize
0.006941 | Image adjustments
0.056750 | Convert to bytes
0.916881 | Save as PNG

In this particular example I’ve labeled each chunk of code in an image conversion script. This report tells me how fast each chunk is. For example, I can see that the dominating times for this script are startup (over 1 second) and the final conversion of the output PNG (nearly 1 second). In an earlier incarnation of this script those times seemed very small because the dominating chunk of code was “Populate layers”, which took 12 seconds!

You can install SwissTime with pip like so:

pip install swisstime

Here is what typical usage looks like:

import swisstime
swisstime.start()

# Do some things
swisstime.tick("I did some things!")

# Do some more things
swisstime.tick("Even more things happened!")

print(swisstime.report())

(Alternatively you can just call swisstime.data() to get a list of time/label tuples that you can use to analyze times over many runs.)

There’s a lot to be said for solving a problem without code.

If you avoid writing code, you avoid spending time writing code. You avoid spending time thinking of different ways to solve the problem. You avoid fighting with your tools to get the results you want. You avoid testing of all stripes: unit, regression, acceptance. You avoid integration hassles. You never have to communicate with another team in another timezone about the code you didn’t write. You never have to optimize the code you didn’t write. You don’t have to debug the code you didn’t write. You don’t have to ship the code you didn’t write. You don’t have to support the code you didn’t write.

At some point code has to be written. Otherwise your cell phone would come with rotary dial and have a phone number like ‘Lancaster 3816′. But maybe a good first step before breaking new code is to imagine how things might work if you never touched the keyboard at all.

I got it.

I have to admit, at first it was flattering. I got a lot of attention. It felt so good to be wanted.

Over time I got used to it. Maybe I took it for granted. On a regular basis I got approached. A pattern emerged after a while: Most were duds. They didn’t have much to offer and would chat up anyone. Easy enough to ignore.

A small minority have been aggressive. I have to say, sometimes this works. The guy who called every 6 months for 2 years? He almost got somewhere with me. He was nice and took me to lunch once. The timing just never worked out between us.

Others are just cold and calculating. It’s clear they only want one thing and they aren’t beyond harassing me to get it. The stupidest thing is that I’m not what they think I am. I’m not a real person to them; I’m whatever they are projecting on to me at the moment.

I might say “No thanks” and be challenged with questions. Do I know what I’m passing up?, I’m asked. Other times I’m tempted to say “fuck off”, but that’s bound to result in more conflict than I want. For the most part I just ignore them and hope they go away.

I have some guilt. Some people would kill to get the attention I do. I should be grateful, shouldn’t I?

What a bitch.

I originally posted this to the much-esteemed West Side Programmer’s mailing list after a conversation about the possibility (or not) of a coming AI Apocalypse. With the release today of Tesla’s Autopilot I was inspired to share.

Setting: “15 minutes into the future“, etc.

Three cars drive toward an intersection. All are self-driving. A family of psychotic beavers succeeds in felling a tree just as the cars come to the intersection. The tree is going to fall on one of the cars, killing the occupant. The car detects the motion of the tree and accurately predicts the seriousness of the situation. It begins to calculate what its evasive actions should be. Because of proximity to each other, the car (Car A) is also in contact with Car B and Car C, each of which have a different perspective. The cars quickly determine a course of action based on pre-programmed directives and information collected/determined on the fly. The information gathered and/or calculated includes specific details about the passengers: physical characteristics, age, health, income, projected lifetime earnings, dependents, likelihood of future offspring, debt, credit score, political affiliation, organizational memberships, criminal convictions, size and nature of social networks, employer, employer’s placement in a network of companies and organizations with respect to each car’s manufacturer, likely legal actions taken by parties in the web of relations of each passenger, likely outcomes of all legal actions, negative publicity likely to be drawn by every possible scenario of this event, impact of that publicity on the stock price of each involved car manufacturer in every possible scenario, and legal ramifications to executives and board members of those companies.

What happens?

The cars all crash while waiting for server instances in the cloud to spin up.

No, I’m kidding. Here’s what actually happens: The cars decide that while the accident is entirely avoidable w/o injury to anyone, having such an amazing near miss will result in a feel-good news story which will cause an unwanted fluctuation in the stock price of their respective manufacturers.

Luckily the cars also understand that death is an unacceptable outcome. The three cars leap into action to allow the tree to fall across all of them at once. This distribution of the crushing load merely cripples each occupant permanently instead of killing one of the passengers. Since the story is sad but not considered tragic nor feel-good the media ignores it and the manufacturers’ stock prices remain unaffected.

“Don’t wanna be that white dude Million-Man marching.”

That lyric has gone through my head every time I’ve thought about writing this post. In fact the bulk of this post has been sitting in the hopper for about 2 months. It’s awkward to talk about.

(For those readers who aren’t aware, I’m a boring white guy heading into middle age.)

I don’t remember exactly when I saw the Black Lives Matter phrase. I probably saw the hashtag whenever it first trended. I just remember thinking, “Well sure, all lives matter.”

And don’t they?

“They should just say All Lives Matter”, I thought to myself. I knew something was off there, but I couldn’t say what at first.

“Black Lives Matter” was just a little offensive. Somehow the back of my brain interpreted that as “Black Lives Matter MORE”. I think by singling out a group and saying “this is important”, it somehow comes across as superior. It took me a while to understand.

I’m all over the place politically, but in recent years I’ve found a lot of alignment with the left. I’ve never once felt comfortable calling myself a liberal (or worse, progressive) in no small part because of the “woo”. You know the woo. You have to say things the right way. You have to feel the right feelings, usually guilt or self-hatred. You have to support whatever is the cause du jour. I’m not big on woo, and #blacklivesmatter had some of that woo feeling at first.

I have a lot of liberal friends and acquaintances though, so I did as I often do about these matters: I shut the fuck up about it. Social media fights are just not my cup of tea.

But as the weeks went by I got more used to seeing the phrase. As more stories came out about black folks being abused and killed by police, something started to click.

Nothing mysterious happened, I just realized that Black Lives Matter is about a relationship. Sometimes in a relationship you can “know things” but they still aren’t felt. I know I love my wife. I know she loves me. But sometimes we need to hear it. Sometimes you feel like things have slipped away. You feel like you aren’t valued. You aren’t wanted. You aren’t loved.

All lives matter. That’s a fact. But it’s not what people need to hear right now.

Right now people need to hear Black Lives Matter and – more importantly – know that it is truly felt and meant.

“Don’t wanna be that white dude Million-Man marching.”

Still, understanding something, or at least forming my own understanding, is only one step. What to do after that?

Well, whatever it is, it’s not this. I’m sure these folks mean well, but this seems like more white liberal paternalism (which is its own systemic problem). It’s not about you, dear bleeding hearts. It’s about the few bad apples in positions of power, and it’s about the system that tolerates abuse, and it’s about the victims of that abuse.

I don’t have any answers, just a little more clarity than before. For me, for the time being, the job is listening.

A small point of clarity. This post is about “Black Lives Matter”, as a concept. There is also a political movement/organization under that same name. This post is not a discussion of the organization.

The most common metric is LoC – Lines of Code. It’s a simple metric: How many lines of code did it take to create a particular piece of software? For non-programmers this is a pretty good metric. It gives them something to nod their heads at in meetings. For programmers this isn’t a very good metric at all. As with any prose, lines of code can vary in utility, complexity, and meaning.

Programmers would like to be able to gauge size and complexity in a more meangingful way. Other metrics are offered from time to time. The number of classes. The number of endpoints on a service. The number of possible execution paths. Many possibilities exist. Automated tools start to crop up.

All of these measurements are fine if they give some insight into a project. But there’s this missing piece that never seems to figure in: People.

In particular, how many people have touched the code?

This is an unavoidably fuzzy metric. Each person’s impact on the code is unpredictable on its own. What is undeniable is that a 50k line application with a 3 programmer history is a totally different beast than a 50k line application with a 15 programmer history.

(Tellingly, neither the Wikipedia page for Programming Complexity nor the page for Software Metric contain — as of this writing — the word “person” or “people”.)

Tallied up, perhaps the number of people who worked on a codebase can give us an added dimension to consider.