Phillip's Notes

Tasty Blended Coffee Treat for 50¢

2026-04-14T08:34:00-05:00

Do you like Frappucino®s¹ from Starbucks? Do you not like how expensive they are? Here’s a recipe for a tasty blended coffee treat that costs less than 50¢ to make:

1 cup milk
5 tbsp sugar
1 tbsp instant coffee
1 cup ice

Blend and enjoy!

Frappuccino® is a registered trademark of Starbucks Corporation. This site is not affiliated with, endorsed by, or sponsored by Starbucks Corporation. ↩︎

AI is making things better, but worse

2026-04-11T20:45:00-05:00

We thought AI was going to make life easier, giving us time back by automating away the monotonous work. But just like the Industrial Revolution, factory automation, and robots didn’t lead to giving workers a life of ease and luxury, knowledge work automation through agentic and generative AI is no Caribbean vacation; it tends to leave us with dopamine “crashes” and mental fatigue. Both of those lead to increased brain fog, exhaustion, and irritability.

Before we go any further, let me preface what follows with the caveat that I do not consider myself a neuroscience enthusiast, let alone a professional. This is my rudimentary understanding, and it is likely missing nuance or is simply incorrect. If you have more experience in this field than me, and you find my explanations lacking, please send me an email and I will update this post accordingly.

When you click the button to send off the generative AI to go do its work, it’s a little bit like pulling the lever of a slot machine. What you get out of it could be terrible, or it could be great. Life changing, even. More often than not, it’s just good enough that we feel like we are winning: The report numbers are crunched, the software is written, the email is polished, the artwork is generated—in quantities far beyond what we could previously accomplish. But there is a cost. We pull the lever, getting close to what we really want, but never quite reaching the peak of excellence. It’s good enough that we go back for more, but not satisfying enough to get us to quit. And so we pull the lever again, and again, and again.

Generative AI work is a lot like a slot machine—sometimes you hit it big.

My understanding is that this is called a variable reward schedule¹. After a prolonged period of operating in this mode, our brains calibrate themselves to this, and the rewards for normal routine activities are dull by comparison. Irritability, exhaustion, lack of motivation, and poor or impulsive decision-making set in, and it takes time to recharge and let our brains calibrate back to normal before we can be decent human beings again.

A related second problem is simply mental fatigue. When you make a series of high-impact decisions, perform qualitative analysis on a large volume of content², or attempt to multitask³, it has a similar effect, at least in my experience. I know less about the biological and psychological mechanisms on this one, but they’re real.

AI work involves all of these leading causes of mental fatigue. Entrepreneur and software engineer Steve Yegge, in his widely circulated Medium post, The AI Vampire, writes:

I’ve argued that AI has turned us all into Jeff Bezos, by automating the easy work, and leaving us with all the difficult decisions, summaries, and problem-solving. I find that I am only really comfortable working at that pace for short bursts of a few hours once or occasionally twice a day, even with lots of practice. … I’m convinced that 3 to 4 hours is going to be the sweet spot for the new workday. … Assume that exhaustion is the norm. Building things with AI takes a lot of human energy.

Because we automate the busy work, it doesn’t give our brains time to recharge from the fatigue. Furthermore, we’re constantly switching from task to task as our agents happily work away. I’m currently learning how to protect that energy and keep some reserve in the tank for my family during evenings and weekends.

You could try to sustain your multiplied output at full capacity and end up disconnected from reality, exhausted, angry, and unhappy. I’ve recently been filling my evenings with LLM conversations, music generation, and vibe-coded side projects only to yell at my kids and say rude things to my wife because my brain is too tapped out to be compassionate.

Too tapped out to be compassionate.

We have to find the right balance between productivity and rest—both at work and at home. We can’t go full bore at work and then expect to be able to have the energy and motivation to accomplish the things we need to get done at home. What is that balance? It obviously will be different for everyone. But we have to be intentional about it. We can’t just let it happen to us, or we’ll burn out and be left with nothing.

B. F. Skinner showed that when rewards come after an unpredictable number of tries (he called it a variable ratio schedule), people keep going much longer. Because the next reward might come at any time, it encourages repeated behavior. This is why slot machines, doom scrolling, social media reactions, and now generative AI are so addicting. This is different from the variable interval schedule, which is time-based. Because the number of tries is proportional to the number of rewards—albeit variably—the variable ratio schedule is much more likely to increase engagement. ↩︎
John Sweller’s Cognitive Load Theory, developed in the 1980s, is helpful here. It’s the idea that we have a limited capacity for working memory, and when we exceed that, it increases stress, reduces learning, and our attention fragments. It’s the same reason a small code change is easier to review than a large one. Generative and agentic AI defaults to producing large changes. There are probably ways that we can tune our context so that the models return smaller changes at a time, but then we drift further into rapid task switching (see footnote #3) and reduce our overall productivity. ↩︎
Psychologists generally distinguish true multitasking from rapid task switching. When people switch back and forth between tasks, performance slows and errors tend to increase because the brain has to reorient each time rather than doing both demanding tasks at once. Furthermore, while the research is still in the early stages, there are indications that the negative long-term effects include weakened memory, reduced concentration and increased anxiety. ↩︎

Vibe coding a native macOS app to edit my blog

2025-12-29T23:22:00-06:00

I have an awesome new editor for my website. It’s a native macOS SwiftUI app that I can boot up on my Mac and write and edit posts. I don’t know Swift. And I don’t know the SwiftUI API. To me this is the perfect candidate for blind vibe coding. It’s a totally offline, unsupported app that improves my quality of life and inspires me to write, because it feels… well, it just feels good.

Collapsible sidebar, date pickers, posts sorted by date and grouped by month.

It’s not perfect. And I have lots of ideas for improvements. It will just keep getting better and better. If you’re interested, you can look at the source in the standalone philtr/blog-admin repo.

I’ve been using the OpenAI Codex CLI for my personal projects, and I find it roughly equivalent in features and quality to Amazon’s Kiro CLI that is provided to me at work, which uses the Anthropic models. I think I prefer Codex just a hair; it does a better job understanding my poor job of explaining what I want it to do, and asks more clarifying questions if my instructions are too vague.

UPDATE (2026-04-17): I split the app out into its own repository at philtr/blog-admin. It still targets the same content layout in this site repo, but it is now developed and versioned separately.

Wanting Back What We Can Never Have

2025-12-27T09:00:00-06:00

On any given September afternoon in the late 1990s, you might find me racing home from school on my bicycle. I was overweight, and not a particularly strong cyclist. The way to my house from Sheridan Green Elementary (which, I learned recently, was demolished and will soon be replaced with a city park) was mostly flat, until you got to the hill on 112th Place, which didn’t really look much at all like a hill until you tried to ride your bike up it, and, upon reaching the summit, you found yourself out of breath and your legs all wobbly.

It was mostly flat.

After catching my breath, greeting my mom, and having a quick snack, I’d ask permission to use our family’s phone line to connect to the internet to check my email and play an online match of Age of Empires, firing up our trusty 56K dial-up modem connected to the family computer. The «familiar screech of the modem making contact» with the digital universe was like the secret knock on the door to a magical world, where nothing was impossible. It was the Information Age, and we optimistically looked forward to the dawning of a new millennium that would keep us more informed, more connected, more efficient, than ever before.

«The internet followed me out of the family room and into my pocket.»

The internet kept getting better. Our connections kept getting faster. First, we installed a dedicated second phone line in our house so we wouldn’t tie up the line for phone calls. Before long, we had DSL. I could download a song from mp3.com in mere minutes, not hours. ICQ and AOL Instant Messenger connected me to school and church friends, but also my friends from around the world. I had several guys I chatted with often whom I had met on a youth group mission trip. The opening door sound of my friends coming online gave a rush of excitement—I didn’t even have to leave my room or stop what I was doing to talk to my friends.

Slowly, the center of social life faded from schools, clubs, and churches—it was now Myspace, Facebook, and Twitter. My addiction became asynchronous, scrolling for the next cool post while waiting for the reactions and comments on my own. Before I knew it, and with my full consent, «the internet followed me out of the family room and into my pocket» with the advent of the smartphone, which acted as an accelerant for all the best—and worst—things the internet had to offer. I didn’t stop to consider the consequences.

It was the Information Age, but we didn't stop to consider the consequences.

The Information Age failed to deliver on its promise. The flood of news, awareness, and connection around the world combined with addiction only led to more anxiety and more wasted time. Keeping up with every latest news story, every current internet trend and meme, has become increasingly exhausting. Those entering adulthood right now have never known a world without an omni-present high-speed connection to anything and everything. They’ve never tasted the ancient peace of living offline.

It didn’t feel remarkable at the time. The boredom, the freedom to let your mind wander without constant digital input. Scanning magazine covers and candy bars in the grocery store checkout line. Waiting for your favorite television program to come on the air. It all just felt normal. And now that peace is gone.

It is tempting to give into the nostalgia.

I often catch myself wishing for that earlier version of the internet—the one I could more easily log off, go outside, wonder, and be bored, and despite technically being less connected, felt more real and satisfying. It is tempting to give in to the nostalgia. But here’s the deal: those times can’t come back. And maybe they shouldn’t. «The world is too connected, too efficient, too transparent now.» There’s no going back. No matter how much we long for the days of intermittent connection, the ship is too large to steer.

But beyond that, here’s the question that scares me most: What if the peace and slower pace of life I remember from my childhood never really existed in the first place? As we age, we look on the past with rose-colored glasses. The past seems simpler now, but it was fraught with its own anxiety. Got stranded on the side of the road in the middle of nowhere? Hope that a car passes by, is willing to stop, and is not a serial killer. We had to rely on the details that the news media felt were important, unable to do our own research. If you had an unusual hobby, it was more difficult to find others who shared your interests.

Counting on the kindness of strangers.

And if I’m being truly honest, I never really lived that slower pace. It’s a false memory of a time before my own. The internet grew up with me, and I grew up with the internet. I was relatively young when the internet became ubiquitous, and I was still in college when I got my first smartphone.

There’s also some good in that connection. It pays my salary, yes—but there are tangible benefits to having the wealth of the knowledge of humanity at your fingertips. I don’t have to brave the busy aisles of a grocery store and wonder if they’ll have what I need in stock. I can order replacement parts for my home appliances and have them arrive in the mail within a day or two, at half the cost of the manufacturer’s part. I can purchase a book and be reading it on my tablet within seconds.

I don’t think what I really miss is the alien sounds of the dial-up modem or waiting for my computer to boot up as the spinning hard drive skipped across bad sectors. I think what I miss is the boundaries and constraints that came naturally with the limitations of the technology. Boundaries between the real and digital world, and boundaries between contexts. It all collapsed into one– «work, family, news, entertainment, faith, and politics, all on the same screen in the same feed.» The boundaries gave life rhythm and meaning, intentionality and purpose, and now daily life has been reshaped at its foundation.

We have to take control of the technology so it does not take control of us.

Where do I go from here? Grieving the loss of what we once had is the first step to moving on. We have to understand the new foundation for living in the digital world, and define the way we will use that technology rather than letting it define us. We have to take control of the technology so it does not take control of us.

We need to establish rhythms of life that bring us in and out of the online world. Weekly events like church, clubs, workout groups, or neighborhood barbecues can pull us out of our digital haze and foster the offline connection we crave.

Maybe it’s OK to keep your world small. Maybe it’s OK to ignore the news headlines and political ranting on social media. Maybe it’s better to let friendships drift apart when you move away rather than try to keep up with it all.

As for me? Being mindful of my screen time and using apps that nudge me in the right direction¹ has palpably increased my satisfaction and enjoyment of both my family time and computing time. I’m going to put down my phone and make the world a little smaller.

I'm going to put down my phone and make the world a little smaller.

Currently, I’m using Be Present. It has a huge advantage over iOS’s built-in ScreenTime controls, and for me, it strikes the right balance of restriction and flexibility. In the last month, I’ve reduced my smartphone usage by over 50%. If you’re looking for something to encourage healthy screen time habits, check it out! This is how we reclaim peace and sanity as culture. ↩︎

Grace in Gratitude

2020-11-24T09:00:00-06:00

I can’t tell you how many times I’ve received a gift and meant to say thank you, whether in person, over the phone, or with a note, and completely forgotten. By not making a priority of letting them know their gift was appreciated, I am being ungrateful. This is a sin that can easily slip into habit without notice.

Look at the lepers who were healed by Jesus on His way past Samaria in Luke 17:11-19. Ten men received miraculous healing, yet only one returned to give thanks and praise to the One who healed him.

[The ten lepers said,] “Jesus, Master, have mercy on us.” When he saw them he said to them, “Go and show yourselves to the priests.” And as they went they were cleansed. Then one of them, when he saw that he was healed, turned back, praising God with a loud voice; and he fell on his face at Jesus’ feet, giving him thanks. Now he was a Samaritan. Then Jesus answered, “Were not ten cleansed? Where are the nine? Was no one found to return and give praise to God except this foreigner?” And he said to him, “Rise and go your way; your faith has made you well.”

Your faith has made you well. I think the interesting thing here is that statement here is true of all ten men, not just the one who returned. They were each given a gift, yet only one was marked by true gratitude. I like to imagine that the other ten were indeed thankful, but they were so excited to be healed that they ran off to the priests to get a clean bill of health, ran home to hug their families, and carried on with their lives. It’s like the kid on Christmas morning who’s so excited about his gift that he tears into it and begins playing with it, but he’s so preoccupied with the gift that he forgets to show gratitude.

Jesus continues to show grace to the ungrateful, but the one who returns the praise back to Him has a richer understanding of the grace he received. This Thanksgiving I’m most grateful for the grace God has given me in every area of life, even (especially?) the areas where He’s growing me.

Getting Unstuck

2020-11-23T09:00:00-06:00

I want to write, more but I’m not sure what to write about. My blog is all over the place and lacks a singular focus, because my interests are varied and lack a singular focus. Some ideas I’ve had over the years:

Christianity, both theology and the practical outworkings of that theology
Programming
Music or musical instruments and musicianship
Apple products
Tasting delicious food and trying new restaurants and recipes
Traveling together with my family

However, if you’ll notice, nothing’s been posted on this blog since January of 2018—nearly three years! I’ve discovered I like the idea of writing more than I acutally enjoy writing. There is a certain romance to the idea of being able to sit down and write a thousand—or five thousand—words that eloquently string together a point that simultaneously resonates with and challenges my audience. There is a reality check than needs to happne here. In high school and college, I had difficulty writing more than a few thousand words on topics I’d thoroughly reasearched and was passionate about. I’m not sure how I expect myself to be able to wirte that on topics in which I have much less knowledge and expertise.

As an effort to unstick myself, I’m going to be OK with writing things that never get published. I’m going to be OK with writing things that only consist of a few hundred or even a few dozen words. The idea is that the more I do it, the more I make a habit out of putting my thoughts to pen and paper (or pixels and bits), the more I will feel comfortable expressing my ideas, interests, and expertise in written word.

I want to say, “expect many more new blog posts from me in 2021.” But I know how that has gone in the past, with interest waning with the moon and passion for writing fizzling out. And that’s okay, too. I sit behind a drumset every few weeks to remind myself how difficult it is and how I’m more in love with the idea of playing drums than I am with putting in the hard work and effort to learn to play them well. Perhaps this will be a catalyst for writing more, but it may also be a “sitting behind the drumset” moment.

Book Review: You Need a Budget by Jesse Mecham

2018-01-03T09:00:00-06:00

I reviewed You Need a Budget by Jesse Mecham on Amazon:

While I am a huge fan of the YNAB product, and have been using it actively for several years, I half expected this book to be a thinly veiled advertisement to purchase his tool and convert to a paying customer. Mecham only mentions YNAB in passing as part of the real life examples, and puts a link to it in the appendix.

This is not a book about how to use YNAB. This is a book to convince you first that the budgeted life is a better life, and secondly that a budget is something YOU can do.

Mecham’s Four Rules are presented in plain English using words that anyone can understand. These four rules really get to the heart of why You Need A Budget. Other financial advice will tell you what and how, YNAB is focused on the why (with a little bit of how).

Even though I was already familiar with the Four Rules, reading this book gave me a renewed vision and helped me realize I was looking at my budget as a “what,” but I needed to examine the “why”.

Adam Asks: Parity, Blocks, and Enumerators. Oh My!

2017-03-10T09:00:00-06:00

Adam is a friend of mine learning to program. Every once in a while, he’ll send me a bit of code to look over to learn how he might be able to do it better. The latest one was an exercise involving splitting an array based on the parity of the number, that is to say whether the number is even or odd.

Can you check this over for me?

Method that takes an array as argument, counts and displays the odds and evens from the array. It works as expected—I’m just looking for best practices.

Here is Adam’s original code:

def numbereven(numarr)
  neven = 0
  nodd = 0
  narray = []
  narrayo = []

  numarr.each do |n|
    if n%2 == 0
      neven += 1
      narray.push(n)
    else
      nodd += 1
      narrayo.push(n)
    end
  end

  puts "There are #{neven} even numbers and #{nodd} odd numbers in this array."
  print "Evens: ", narray, "\n"
  print "Odds: ", narrayo, "\n"
end

my_array = [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 100, 1001]

numbereven(my_array)

I may have scared him off with the sheer volume of my reply. I wrote a bit how I would write it, another way to write it, and why I wouldn’t write it that way.

How I Would Write It

def parity_counts(numbers)
  evens = numbers.select { |number| number.even? }
  odds  = numbers - evens

  puts "There are #{evens.size} even numbers and #{odds.size} " +
       "odd numbers in this array."
  puts "Evens: #{evens}"
  puts "Odds:  #{odds}"
end

An array knows its size, so there is no need to keep track of it separately.
Ruby integers have handy even? and odd? methods that return true or false. Under the hood, that method is doing exactly what you did: n % 2 == 0.¹
Reducing if/else branches and loops helps the reader understand what is happening.
Yours might have better performance, especially as input size approaches infinity.
Your print statements ending in \n are fine, but you’ll more likely see people using puts with string interpolation (#{}) when they want to print a string ending in a new line.
In a real-world application, I would probably not put the screen output as part of the computation logic. You’d probably want to return a hash or other data structure and put the presentation logic somewhere else.

An Alternate Implementation

There is more than one way to skin a cat. Here’s an alternate implementation:

def parity_counts(ns)
  ns.reduce([[],[]]){|(e,o),n|n.odd? ? o<<n : e<<n;[e,o]}
end

This will return an array where the first element is an array of even numbers, and the second is an array of odd numbers. But there is very little about this solution that is best practice. Clarity should be preferred to brevity:

Variable names should reflect what is stored in them. Single letter names and even some abbreviations don’t make it immediately clear as to what it is to be used for.
reduce is a concept that is not easy to grasp. I use it sparingly when it really makes the most sense, rather than just because it makes me feel smart.
When you have more than one statement inside an iterator, it’s good style to use do/end instead of curly braces, and break up the statement into multiple lines. There are varying schools of thought on this:
- Some people use curly braces when they are using the value returned from the iterator (e.g. give me the parent contact email for each student), and do/end when they are mutating state (e.g. take this list of students and calculate each one’s grade)
- Other people say to always use curly braces for single line and use do/end for multiple lines. I follow this pattern; I think it looks nicer.

A Better Version of the Alternate Implementation

So, given the concepts above, using reduce, I would write it this way:

def parity_counts(numbers)
  numbers.reduce([[], []]) do |(evens, odds), number|
    number.odd? ? odds << number : evens << number
    [evens, odds]
  end
end

The [[], []] sets up the data structure that we are going to pass to the first iteration to store the values in. Reduce takes the return value of the block and passes as the first argument to the next iteration. By putting parens around (even, odd), we can “unwrap” the array into separate arguments. It’s kind of a confusing concept to wrap your mind around, but once you get it, it’s incredibly handy for writing iterators.

Takeaways

In conclusion, a couple takeaways: First, Ruby has a huge standard library, along with helper methods to do common things. It’s usually a good idea to check out the documentation for a given class to see if there is already a method to do what you’re trying to do (even? and odd? from above, for example).

Second, with Ruby, it’s possible to write code that can almost be read as a sentence. Variable names and function names following this pattern make your code easier to understand and, in my opinion, much more fun to write.

UPDATE: How about `Enumerable#partition?`

There’s almost always a better way. Turns out, Ruby already has a method for partitioning an enumerable. In this implementation, we ask each element in the array if it is even. If so, it goes into the first partition. Otherwise, it goes into the second.

numbers.partition(&:even?)
#=> [[2, 4, 6], [1, 3, 5]]

This is the shorthand syntax (using Symbol#to_proc—a post for another day, perhaps) that would be equivalent to writing

numbers.partition {|number| number.even? }

So to put this back in to solve the original problem:

def parity_counts(numbers)
  evens, odds = numbers.partition(&:even?)

  puts "There are #{evens.size} even numbers and #{odds.size} " +
       "odd numbers in this array."
  puts "Evens: #{evens}"
  puts "Odds:  #{odds}"
end

The source for this function is in C, but here it is if you want to take a look at it. ↩︎

I Lost My Wedding Ring

2017-03-03T09:00:00-06:00

I have a wedding ring in the Pacific Ocean somewhere.

No, it wasn’t some dramatic gesture. In fact, the story is not really that interesting. We were in Puerto Vallarta for my brother’s wedding a few years ago, and my wife and I decided to go for an evening swim. I left my wedding ring on because I didn’t think it would fall off. It fell off. I never found it.

The little cove in the Pacific where my ring forever rests.

When I still had that original wedding ring, I was constantly afraid of losing it or not knowing where it was. Losing the wedding ring was liberating, in a way, because now I place more weight on the meaning of the symbol on my finger than the actual object itself.

The modern practice of men wearing a wedding band only really dates back about a hundred years, and mostly due to increased advertising from the jewelry industry.

All that aside, I wear the band for two main reasons. First, to remind myself that I made a commitment to my wife. It’s heavy, almost 18 grams. I notice it when it’s there, and I notice it when it’s not there. Whether or or off, I’m reminded of my commitment. And second, to proclaim my undying and unconditional love for my wife to the rest of the world.

Divorce has never been an option, in my mind. I can’t imagine doing life without her. I only have one ring that I’ve replaced the wedding band with, but if I had to replace it again, it wouldn’t matter. Because it’s not the ring itself that’s important, but the constant reminder of my commitment to her.

Programming as an Image-Bearer

2017-03-02T09:00:00-06:00

Adam Murray, in an interview with The Gospel Coalition:

As an image-bearer of God, how does your work reflect some aspect of God’s work?

At the hackathon last year, a theology professor gave a talk where he described coding as an extension of Genesis 1–2. First, there was nothing; then, God spoke and made this awesome universe. Similarly, coding is turning ideas—essentially nothing—into reality through computer language. It’s a cool reflection of the image of God in humanity.

That you’re building something from basically nothing has long been my favorite thing about programming. I’ve never thought about how it ties in with Genesis in God’s act of creating the universe ex nihilo. Every time we start a new project or feature, we have the opportunity to bring glory to the First Creator.

A Little "Special" Goes a Long Way

2015-03-04T09:00:00-06:00

This post was originally published on the Oven Bits Journal, under the titles, “Your Success is Our Success”, and “Being Agile Leads to Results”. It has been reproduced here for posterity.

Every client will tell you that they want two things: be on time, and stay within their budget. At the heart of every project, every client also wants a third thing–results. We like to take it one step further and say that every client wants us to be on time, stay within their budget, deliver results, and be special every step of the way.

Discovering Results

Before we even start work on a project, we ask, “why?” and dig deep. Why are we building this? Why would someone use it? What is the desired result? We want to learn what makes our clients and their customers tick. We outline key metrics during this time that we’ll use later on to validate our assumptions. If we don’t know how our clients define success for the project, we won’t know if we’ve succeeded in meeting their expectations.

Far too often, a project will set off on the wrong foot if the team doesn’t have a clear vision. The tricky part of any engagement is knowing when to stand your ground versus being flexible to what the client desires. The key is knowing how to test and validate your assumptions. With a disciplined build, measure, learn process you’ll be able to validate any assumptions that the team has developed. Basing six-figure decisions on intuition and gut feel is just plain foolish. If it can be measured and tested, it should be. As one of our clients so brilliantly put it, “My opinion does not matter.” After we all had a good chuckle, he was clear to reiterate, “No really, it doesn’t matter. All that matters is what our consumers say.”

Beware: Vanity Metrics

Finding the right metrics to track is no simple task. Product development teams—including ourselves—can easily get distracted by vanity metrics. These are the big numbers that sound impressive and make you feel confident you are on the right track. 100,000 downloads on the App Store. 20,000 signups. How do these impressive and seemingly benign statistics secretly undermine what you are trying to accomplish?

In The Lean Startup, Eric Ries writes, “Vanity metrics can allow [you] to form false conclusions and live in [your] own private reality.” Ultimately, vanity metrics distract us from—or worse, blind us to—the results we’re trying to achieve.

What metrics point us toward the results our clients want? Ones that provide a clear path for action. Sure, 100,000 downloads is great. But how long has your app been available? How many of those who downloaded the app continue to use it on a regular basis? Are people purchasing the in-app purchases? Do you know why? We stay ahead of the game by asking these questions early and often.

Instead of focusing on the vanity metrics, take a look at time in app, repeat usage, and metrics that can’t be propped up from a massive, one-time advertising spend. That’s when things will get real for you.

A Dash of Special Sauce

Being on time and on budget will get you a passing grade. Asking the right questions and having the right answers will get you a B. In order to truly succeed, we can’t simply meet expectations. We must exceed them. Chris Bowler hits the nail on the head: “Your goal as a business should be to make your customer look as good as possible.”

This can come in many forms. Often it’s something extra thrown in with a deliverable, whether it’s a planned feature ready to go ahead of schedule or a special interaction animation. Sometimes it’s a batch of cookies or cupcakes. Sometimes it’s just a note of encouragement or gratitude written to the client.

Our success rides on the shoulders of the success of our clients and customers. Only when we meet and exceed their expectations can we say we’ve done our job.

On Being Friendly

2015-01-06T09:00:00-06:00

John Saddington, on his blog:

Friendly? Ah, that’s curious to me because I have not always been the nicest nor the friendliest person on the internet. The truth is that I once saw the internet as an opportunity to just shout my own opinion at people instead of seeing it as an opportunity to close the gap between myself and many other individuals. . . . After being treated poorly myself I decided that I didn’t want to do that to others.

So, I suppose, I became a much more friendlier netizen.

It never hurts to be a nice person. People will take notice. And, of course—as we all know—Mean People Fail.

Update: Fixed the broken link to John’s blog which was previously on the Desk site.

Telegram Reaches 1 Billion Daily Messages

2014-12-18T09:00:00-06:00

Not sure how I missed the whole Telegram thing back in February when WhatsApp went down. While nothing compared to WhatsApp’s 600 million active users, Telegram’s 50 million active users each sending 20 messages a day with no money spent on advertising is nothing to shake a stick at.

What intrigues me the most is that it is set up as a non-commercial organization promising to never show ads or sell your data, along with the ability to send “secret chats” with end-to-end encryption–only the two devices in question can read the message.

It’s a cool initiative; I’m curious how this all is going to shake out as Facebook continues to pour more money into Messenger and WhatsApp and vie for market share.

And please—feel free to Telegram me any time.

Google Web Fonts Typographic Project

2014-12-17T09:00:00-06:00

There are over 650 Google Fonts available for free. But, pairing typefaces isn’t easy and many of those fonts don’t work for typical websites. Part of the 25x52 initiative, this collaborative, ongoing project offers inspiration for using Google’s font library.

Passages from Æsop’s Fables set beautifully in web fonts available from Google fonts. There are some great choices available if you can find them and use them properly.

The curse of compressing reality

2014-12-16T09:00:00-06:00

Noah Lorang on Signal v. Noise:

Any creative endeavor is highly non-linear, but the sharing of it almost always skips a lot of the actual work that goes into it. That’s ok; a clear progression makes for a good story that’s easy to tell. But don’t judge your reality against someone else’s compressed work. It’s ok if it takes you a day to make a cutting board like one that someone made in six minutes on YouTube; the truth is it probably took them a day too.

It’s easy to look at all the awesome stuff on the internet and get depressed that you can’t do something well in a day or a weekend. Spending a couple hours doing parts of a project that are not glamorous is OK and to be expected.

O Come Let Us Adore Him

2014-12-13T09:00:00-06:00

John Piper:

And so the “us” of “O come let us adore him” is the “us” of heaven and earth, singing together. According to Revelation 9:16, there are at least 200 million angels at God’s bidding. Surely at least some of them are assigned to give heed to us as we sing, and do what we say: “Sing choirs of angels, sing in exultation!” If we could only hear! Heaven and earth married in singing to Christ.

The choir of men and angels join together in song to worship the creator of the universe, who came to rescue and restore the human race from our own brokenness.

I love this dissection of such a beautiful and rich hymn. Christmas is about so much more than just a cute little baby Jesus in the manger.

Being Careful About What We Ingest

2014-12-10T09:00:00-06:00

Romans 8:5-6 (ESV):

For those who live according to the flesh set their minds on the things of the flesh, but those who live according to the Spirit set their minds on the things of the Spirit. For to set the mind on the flesh is death, but to set the mind on the Spirit is life and peace.

Many people are very careful about what they eat. We are careful to make sure we eat our fruits and vegetables. We are careful about how much fat and how many carbohydrates we eat. In the same way, we should be careful about what we eat, but on a spiritual plane. With what kinds of things are we saturating our minds? What kinds of junk food are we allowing into our homes and minds via television, books, and the internet?

Paul urges us in his letter to the Colossians to be mindful of our mindset. “If then you have been raised with Christ, … Set your minds on things that are above, not on things that are on earth. For you have died, and your life is hidden with Christ in God.” ¹

Have we been raised with Christ? Yes! So what are we to do? Set our minds on the the things of God. Not that we shouldn’t be concerned with day-to-day affairs, but that our mindset would be one that reflects God’s glory. I don’t want to make lists of movies, books, or websites that we can or can’t watch, read, or visit. Our motive is not to increase morality. It is not to try to appease God’s anger at our sin by saying “I don’t watch R-rated movies.” We don’t gain any favor with God by keeping lists of rules. The Pharisees tried that and Jesus spoke most vehemently against their ways. Our goal is to reflect to others the glory and grace of our Lord to those who have not seen or known him.

The only way to reflect God’s glory is to saturate our minds with the things that do. “Finally, brothers, whatever is true, … honorable, … just, … pure, … lovely, … commendable, [excellent, or praiseworthy], think about these things.” ² Is that movie we want to go see really excellent and praiseworthy? How about that latest bestseller we see on the shelves at the bookstore? Is it pure and lovely? The blogs we like to read—are they true and honorable or do they defame others with lies? We shouldn’t shelter ourselves from these things, ³ but we should always be on our guard.

Jekyll, S3, and CloudFront

2014-12-06T09:00:00-06:00

GitHub pages is awesome. Like, really awesome. But I wanted to try something new that would give me a little more flexibility (e.g. the ability to use Jekyll plugins). So I started looking around for the best way to deploy my static site to S3.

Boy, did I ever tumble down a rabbit hole. It seems like everyone’s got their own idea of how it should work, each with their custom one-off scripts that I would hate to debug, and not fully supporting what I was trying to accomplish. That’s when I remembered seeing jekyll-s3 at some point in history, so I thought I would at least check it out to see if development was still active. Well, as it turns out, development has indeed continued, just on a different project called s3_website. s3_website can deploy any set of files to Amazon S3, but Jekyll and Nanoc are supported right out of the box with no additional configuration.

s3_website is a mature and powerful tool. Not only does it deploy your site to S3, but it has a host of other features. If you set up CloudFront to distribute your website from S3, s3_website will handle all the cache invalidation commands so you don’t have to worry about it. It also supports dotenv out of the box, so you can commit your configuration file without sensitive access keys.

The only downside is now I have to be at my computer to make a change to my website, but let’s be honest here—how often is that situation going to come up? Considering my last post was written nearly two years ago, not often.

Returning Multiple Formats with Rails Dynamic Error Pages

2012-12-13T09:00:00-06:00

Yesterday my coworker Eric asked me to build out some dynamic error pages for the upcoming version of one of our apps. He suggested a certain pattern that we used in a previous application. Being the adventurous soul that I am, I wanted to make sure his way was the best way. After some deep ~~soul~~ google searching, I found out that Eric’s method was considered best practice. It’s described in some detail on José Valim’s blog post, but I’ll break it down for you here:

First, we need a controller for our error pages. Start out with something simple, like this:

# app/controllers/errors_controller.rb
class ErrorsController < ApplicationController
  layout 'error' # only if you want a separate layout for your errors

  # 404 Not Found
  def not_found
  end

  # 422 Unprocessable Entity
  def unprocessable
  end

  # 500 Internal Server Error
  def internal_server
  end
end

Now let’s add some routes:

YourApp::Application.routes.draw do
  match '/404', to: 'errors#not_found'
  match '/422', to: 'errors#unprocessable'
  match '/500', to: 'errors#internal_server'
end

With Rails 3.2, error pages have been extracted to a Rack Middleware. Fortunately for us, so are the application routes. We can to tell rails to use our routes app for the error pages in config/application.rb:

# config/application.rb
module YourApp
  class Application < Rails::Application
    # ...

    config.exceptions_app = self.routes
  end
end

And that’s it! Add some page templates in app/views/errors/ and you should be good to go.

To test it, we need to turn off the development mode error pages. Set consider_all_requests_local to false in config/environments/development.rb:

# config/environments/development.rb
YourApp::Application.configure do
  # ...

  config.consider_all_requests_local = false

  # ...
 end

You’ll want to set that back to true before committing your code, or else you won’t receive any helpful feedback while developing your application.

All of the above steps were fairly simple, but figuring out how to determine which format to return is a little complex. env['REQUEST_PATH'] contains the error path (e.g. “/404”) and no format information is getting passed to the errors controller, so the standard Rails respond_to stuff is not going to work here. We can grab the original request path via env['ORIGINAL_FULLPATH']. I added a few private methods to errors controller to help out. The first two help me figure out what format the request came as. I want to return a JSON response if the request starts with ‘/api’ or ends in ‘.json’. The third will render our template in the format we want.

class ErrorsController
  private

  def api_request?
    env['ORIGINAL_FULLPATH'] =~ /^\/api/
  end

  def json_request?
    env['ORIGINAL_FULLPATH'] =~ /\.json$/
  end

  def render_error(error)
    if api_request? or json_request?
      render "#{error}.json.jbuilder"
    else
      render "#{error}.html.haml"
    end
  end
end

Now we just need to update the actions in our controller:

class ErrorsController
  def not_found
    render_error "not_found"
  end

  def unprocessable
    render_error "unprocessable"
  end

  def internal_server
    render_error "internal_server"
  end
end

And we’re done! Make sure you add your json templates to app/views/errors.

$ curl http://localhost:3000/api/v1/nothing-here
{ error: "not_found" }
$ curl http://localhost:3000/nothing-here.json
{ error: "not_found" }
$ curl http://localhost:3000/nothing-here
<h1>404 Not Found</h1>

Thank you, Dvorak!

2011-11-11T09:00:00-06:00

Dvorak. Not exactly a common name.

There are a few famous folks with that name. Antonín Dvořák composed the New World Symphony during a visit to the United States in 1893 (but I can’t hum a single theme from it). There’s also Dusty Dvorak, a gold-medal-winning U.S. volleyball player in the 1984 Summer Olympics (I wasn’t born yet). Some people might think of Ann Dvorak, the American actress from the early 20th century (I haven’t seen any of her movies). If you’re into technology like I am, you probably have seen John C. Dvorak’s columns in PC Magazine or InfoWorld (I tend to disagree with his opinions).

The Dvorak that I am most thankful for is August Dvorak, the educational psychologist and professor at the University of Wisconsin. You probably haven’t heard much about him. He and his brother-in-law, William Dealey, created the Dvorak Simplified Keyboard layout. What’s the Dvorak Simplified Keyboard layout, you ask? I’ll tell you. But first, a quick history lesson:

It was the late 1860s. Christopher Sholes, Carlos Glidden, and Samuel W. Soule were designing and building the first commercially-successful typewriter, the Remington No. 1. As they were building it, they noticed that when two keys near each other were pressed in rapid succession, the typewriter had a tendency to jam due to the mechanical levers that would swing up to print the letter on the page. This was a problem. They couldn’t just tell people to be more careful, so they decided to rearrange the keys so that there would be fewer jams. The keys were spread out so that common letter sequences would not be near each other. The typist could then press these two keys more rapidly with fewer jams. This was the birth of the QWERTY layout. Despite initial criticisms, typewriters were a huge success.

Back to our friend August who, while advising a student’s master’s thesis on typing errors, became interested in the subject. By the 1930s, typewriters were much more advanced, and their propensity to jam had been greatly reduced. He and his brother-in-law meticulously studied every aspect of typing, from the physiology of the hand to the most commonly used letters and letter sequences in the English language. The end result: a keyboard layout that was able to to be learned, on average, in one-third of the time that it took a typist to learn QWERTY, and consistently produced faster typists who made fewer errors.

The US Dvorak keyboard layout. (Image: Wikipedia).

The effort and attention to detail on the design of this layout is insane:

70% of all keystrokes are on the home row (vs. 32% for QWERTY)
Most words will alternate back and forth nicely between your hands, because all vowels are typed with your left hand.
The least common characters are placed on the bottom row

Fast-forward to the digital age. We still type on keyboards, and we are no longer bound by the lever action required to put characters on the page, yet we are still using a keyboard layout that was designed for typewriter efficiency, not typist efficiency. Barbara Blackburn, the world’s fastest typist, uses the Dvorak layout. Other proponents include Steve Wozniak of Apple fame and Matt Mullenweg, creator of WordPress.

Before the Dvorak Simplified Keyboard layout, I was unable to teach myself to touch type. Forever bound to my lovingly-named “3-finger claw” stance, I would type three letters, then backspace two, type three letters, backspace two, and so on. By the time I finished a paragraph, my whole arm hurt. Then Dvorak found me. I was forever changed. A fresh start. Within a month I was back up to my original typing speed of 35 WPM. Within a year I doubled that. Thanks, Dvorak! (Take that, QWERTY!)

Less.app

2011-05-04T09:00:00-06:00

Less.app:

{Less} extends CSS with variables, nested rules, operators and more. If you’re still building websites without it, you’re an idiot. This app makes it dead simple to use {Less} by automatically compiling *.less files into standard CSS.

~~Using it to precompile the styles for this very site.~~

Update: NOPE.

Mom’s Secret Sauce for Success

2011-05-03T09:00:00-06:00

I owe every accomplishment I have ever achieved to three things my mom instilled in me from a young age. She was relentless, frustrating at times, but looking back I know it was out of a deep love for me and my future well-being.

Always do your best

Most of the time my grades were good. But I never got in trouble for bad grades, unless they were a result of not trying. Every time I brought home a paper with a bad grade, she’d ask me, “Could you have gotten a better grade?” Most of the time, the answer was yes and I knew I had been lazy. If I really had done my best, my mom praised me for the effort I put in, and encouraged me to prepare for next time. As long as I did my best, she was satisfied.

Now, in my professional life, a job isn’t done until I’ve done my best. Even when something appears to be done, I dig in and refine, refine, refine until I can honestly say I’ve done my best work.

Never be satisfied with the status quo

I was a rockstar speller when I was in grade school. I’d look at a word, write it out once or twice, and then it would be in my head forever (or at least until the next spelling test). My mom caught on to my abilities, and wouldn’t let me stop at the lists of 5-letter words my teachers gave me.

To this day, I try to push myself into uncomfortable territory. The only way to better yourself is to step outside your comfort zone, and now I have twenty-some years of practice thanks to my mom.

Be a life-long learner

Our family vacations were primarily about spending time together, but they were always educational. We wouldn’t just stop and look at the sights to see, we would read every sign and soak up all the information we could get. Even our destinations were educational. I learned all about wildlife, caves, the Cold War, and cutting-edge scientific research on our vacations.

Every once in a while, when I feel like I’m just drifting through life, I’ll go learn something new and it gives me just the pick-me-up I need.

Thanks, mom, for always encouraging to do my best, to not be satisfied with the status quo, and to be a life-long learner. Happy mother’s day!

Gumbo

2011-04-30T09:00:00-06:00

My wife and I made “Quick & Easy Gumbo” for an event at our church this evening. I know it’s not “real” gumbo, but let me tell ya, it’s pretty tasty.

We modified the recipe in a couple ways, and learned a couple things in the process about how we should have adjusted for those changes.

First, we used hot breakfast sausage. The flavor on this stuff is amazing, but I think it must have been a little leaner than the sausage the recipe was expecting. I would recommend adding a tablespoon of oil to the your pan before putting the sausage in so you don’t burn it.

Second, and I suppose this applies to recipes in general, chop up everything ahead of time because you won’t be able to stir continuously and chop up all your veggies at the same time–at least not in 5 minutes.

Third, I added about twice as much creole seasoning as the recipe called for. I tend to like my food salty and spicy, so the recipe might suit your taste, but I would suggest 1½ times the amount listed in the recipe.

My cooking adventures will continue!

UPDATE: Louisiana lady says, “Good job on the Gumbo!” I’d say that’s a success :)

Head in the Clouds

2011-02-17T09:00:00-06:00

Valentine’s day was a few days ago, and since I’m still in the mood, I am going to declare my love for the cloud. I love that at Balcom Agency, we host all of our websites in the cloud. What advantage does this have, both for us as developers and for our clients?

Backups – Nothing eases the mind of a web developer more than knowing there are recent backups available in case of a disaster. Our hosting providers make scheduling automatic backups dead simple.
Launching a website – I love launching a website into the cloud. Instead of ordering a dedicated server from a hosting company that takes a couple days to be built, I can, with a few clicks, create a new virtual server with just the specs I need to run the website.
Server replication – When we need to make a change to a website, sometimes we don’t want to mess with the version that’s currently running live. I can create an exact duplicate server in a matter of minutes, and deploy the updates to that server. The client can see and approve the changes without running the risk of having unwanted side effects on the live website. This takes a lot of stress off us as developers, because, believe it or not, we can make mistakes too. We can do as many test runs as we need until we feel comfortable and confident that we’ll make the changes swiftly and professionally.

The cloud is not the solution for every problem (as Microsoft has suggested), but we sure get a lot more done around here because of it.

Using the Heroku Gem Inside Your App

2010-08-12T09:00:00-06:00

Brett asked this question on the Heroku mailing list:

I want to let users purchase custom domain names to access my app on Heroku. When the purchase transaction takes place, I’d like to add the custom domain to my app in real time. I thought of using the Heroku gem/command line tool from within the app, analogous to the command-line expression: heroku domains:add www.example.com. Is this possible?

It really got me thinking about exploring all that the Heroku API has to offer.

I thought I remembered seeing someone using the Heroku gem inside their code before, but I couldn’t remember where, and I couldn’t seem to find any examples in the Heroku docs. Good! Time to get my hands dirty with some raw source code.

I opened up the Heroku client source code and it’s actually designed to be used inside your code. How about that. The first few lines after all the requires describe perfectly how to use it:

# A Ruby class to call the Heroku REST API.  You might use this if you want to
# manage your Heroku apps from within a Ruby program, such as Capistrano.
#
# Example:
#
#   require 'heroku'
#   heroku = Heroku::Client.new('me@example.com', 'mypass')
#   heroku.create('myapp')

As you scroll down through the functions, you find that you can do anything in your source code that you could do on the command line. To answer the original question, if you wanted to use this command-line task:

heroku domains:add www.example.com

You would simply put these lines in your source code:

require 'heroku'
heroku = Heroku::Client.new('me@example.com','mypass')
heroku.add_domain('myapp','www.example.com')

The add_domain function is found on line #120 of client.rb.

UPDATE: Someone replied to the thread, pointing out that you should keep things such as your Heroku credentials out of your code and in an environment variable. Heroku already stores your app name by default in ENV["APP_NAME"]. So, to make the above code a little DRYer and safe to distribute. At the terminal:

heroku config:add HEROKU_EMAIL=me@example.com HEROKU_PASS=mypass

And then in our code:

require 'heroku'
heroku = Heroku::Client.new(ENV["HEROKU_EMAIL"],ENV["HEROKU_PASS"])
heroku.add_domain(ENV["APP_NAME"], 'www.example.com')

Now we won’t be distributing our Heroku credentials with our code.

UPDATE #2: Someone asked me if storing Heroku credentials as config variables would be secure. Ben Hartney had the same question and contacted Heroku support. Here is their response:

Storing this information in an environment variable is safe. We also store your database connection info in environment variables, so we consider it secure.

Shapeshifting “Classes”

2010-08-05T09:00:00-06:00

I’ve been using Ruby full-time+ for the last few weeks, working on projects at work and then personal projects when I get home. There are a few things I’ve learned about Ruby that surprised me. Most of these were discovered when, going out on a limb, I tried something completely random to see if it worked–and it did!

On one of my projects, when initializing the application, I needed to be able to choose one class (we’ll call it FujiApple) if certain variable was set, and a different class (call it Gala) if it was not. So I tried the following:

if @certain_var.nil?
  AppleType = GalaApple
else
  AppleType = FujiApple
end

Now, keep in mind, AppleType is not a class that I have defined anywhere in my code, yet I could use it just the same as the class that I set it to!

What’s really going on here?

It took me a while to realize that I’m not actually creating a new class here (and hence the reason I put “classes” in quotes). What’s actually happening is that AppleType is just a plain old constant. However, because that constant is pointing to a class, you can use the same as if its own class.

If my FujiApple class looks like this:

class FujiApple
  def eat
    "NOM NOM NOM"
  end
end

I can do this:

> AppleType = FujiApple
  => FujiApple
> myApple = AppleType.new
  => #<FujiApple:0x1011b31c0>
> myApple.eat
  => "NOM NOM NOM"

Should I use it?

I’m not quite sure how good this practice actually is, because if we need to change the AppleType somewhere in the middle of the application run, then we are changing a constant, which is a no-no. However, if used properly, and only set during the initialization phase of the app (i.e. once), I believe it is a handy shortcut to having to check for the existence of that variable every time you want to use the class. Feel free to correct me on this. There is probably a better way to do it, such as using inheritance, but I found it fascinating that a solution like this would even work.