✍ Evan Travers

Committing with the last command as the description

2026-03-26T08:33:00-05:00

I realized recently that I really often end up committing a simple jj desc to reflect that I've just run a linter, updated a package… things like:

bundle update
nix flake update
npm format

You get the idea.

I had opencode whip up a simple fish command that mimicks what ⌥-p already does: takes the current prompt line and wraps it in <prompt> %| less -R. If the current prompt is empty, it uses the previous command. This is exactly what I wanted!

function jj_desc_wrap --description="Wrap previous token or history command in jj desc -m """
    set -l tokens (commandline -pc)
    set -l token $tokens[-1]
    if test -z "$token"
        set token (history --max=1 | head -1)
    end
    if test -z "$token"
        commandline -i "jj desc -m \"\" "
        commandline -C 12
    else
        commandline -r -- "jj desc -m \"$token\""
        commandline -C 12
    end
    commandline -f repaint
end

I have this bound to ⌥-o for now, because I think there's no conflicts.

Already this has been such a pleasant little UX change to my daily work.

AI in the City of God

2026-03-09T20:11:00-05:00

Theological evaluation begins and ends with the sovereignty of God. A technology is more than a tool, more than a technique, more than a telos. It is all three, and all three must be asked together. The maker is an image-bearer under a mandate. The method is stewardship under a standard. The end is the glory of God and the flourishing of all He made.

Starting from Augustine's twin cities serving two masters, Randy starts to build a framework for technology.

Every tool serves a calling, every technique serves a standard, and every telos either honors God or rivals Him.

I really like this framing, and I'm very excited to see what Mr. Caldejon writes next. (Bonus points for the Bavinck reference!)

A Theological Understanding of Technology - Tool, Technique, and Telos - www.decalogue.ai

2025

2025-12-30T12:08:00-06:00

This was a pretty wild year all 'round. Birth of a child, caught in a layoff, freelancing, working in a startup.

Spiritual Life

I've been keeping track of prayer requests this year. As of this writing I have recorded sixty-one specific answers to prayer this year. Praise the Lord for his provision!

I added kneeling prayer once a day to my habits… targeting the time I'm usually most tired and prone to irritability. I think that's helped. My personal quiet time has been very fractured and infrequent. I've been a little more consistent at having family worship at the table.

That being said, I have felt the Lord's hand and protection through the big waves. In so many ways this year has been a dark mirror of 2020: political unrest, job change, pregnancy, much discussion and stress… but rather than being pulled away from the Lord and pulled apart in marriage, we have been strengthened and upheld.

Relationships

Because of chaos and baby, we didn't take too many trips this year… just three wild trips to the beach.

We've had a lot of folks in our home for board games and dinner, and that's been a big joy that we want to do more of next year.

Another big answer to prayer is that we are both enjoying our kids more than ever. They haven't been easier to deal with, I have just been granted more grace to enjoy them.

Health

Equal victories and defeats here.

FitBod continues to be my workout buddy.

After a short weight loss before baby arrived, I've shot up to 180lb and stayed there since June. Some of that was post-birth weight gain, but some of that has been some re-building of new muscle. I bought my first t-shirt in size L this year, and it's because of my shoulders not because of my gut. 💪

I tapered off Zyrtec, and have been six months or so without my daily dose. Doesn't seem to have affected my sinuses at all.

My skin stuff and gut owing to stress has continued to be a problem, but not as bad as previous years. I think that the smartwatch's input on my sleep has been a good lag metric to help me stay healthy.

My Morton's Neuroma hasn't really improved in a year of PT. I'd say I'm about 70% faithful to do the physical therapy, but I hoped to be farther along. I am starting to consider surgery… mobility is a big deal. I was really sad that for the first time in five years I was unable to run the 8k race near my house. I really have to get that fixed.

Creating

I only published 20 blog posts, less than half of last year.

My writing habit got severely curtailed by both the birth of a child, loss of a job, and pursuit of a new one. I still have written a fair amount (I've created 1100+ files in my wiki!) it just hasn't made it's way out to the blog.

Scrolling through the list of wiki notes… there's a lot of thinking about work, about personal growth, about people that I'm praying for, and of course, AI.

(I have a couple larger essays stuck in my system again… maybe I'll get a chance to polish them up over the break, or at least find a format where I'm ok with publishing them imperfect.)

I did spend some time updating the blog itself… adding slashpages and updating some old ones. The /blogroll and /uses get updated relatively often, and I usually update /now when I switch books.

I've continued to double down on nix as a strategy and am loving it. It was a huge blessing to have that tool in my arsenal when I wound up consulting as a software engineer for ~6 months. The fact I could update my computer without fear of destroying a project's configuration… wonderful.

While I use AI a good bit for work, I still try to write my prose and code by hand. No good reason. Just sharpening a tool, and hopefully my mind.

Sketchnoting

I continue to sketchnote weekly sermon entries. I also did sketchnotes for the dev track of a local tech conference. Looking back at this year I realized that I've relied a bit too much on speechbubbles, and I'd like to do more diagrams and visual storytelling.

I've also started (very recently) a weekly sketching habit. I've not been utterly faithful to it, but it's been nice to sit and just see for thirty minutes. I'm super thrilled that my kids sit with me and show me their results.

Reading

This year I changed my lead indicator of reading to "books I have started" to try and allow myself to drop a book that "gets stuck." By this metric, I started 26 books, finishing 21.

Some highlights:

Polostan by Neal Stephenson: In a world of change, spies navigate the rising power of communism and the new science of the atomic age.
Teach Them Diligently by Lou Priolo: A look at raising children through the light of 2 Timothy 4:2.
Consider The Lilies by Jonny Ardavanis: Extinguishing anxiety by looking into the overwhelming goodness of our Father God.
Piranesi by Susanna Clarke: Wandering a strange house, an optimistic journal ponders the most important questions.
How to Lead Your Family by Joel Beeke: Exhorting men to pattern our lives after Christ in his roles as prophet, priest, and king in our homes.

My new Boox Palma has really helped… I really love KOReader.

I've listened to less podcasts over the year… in part because of realizing how vulnerable my mind is to words. Some standouts from my subscriptions:

Gospelbound has been a great source of books for my queue.
Dwarkesh has kept me apprised of AI thought, although I find myself disagreeing with his positions more over the year. (I find his comparisons of baby's learning styles to LLM training sequences utterly ridiculous, and I fundamentally disagree with his concept of what is knowledge.) I do really admire his curiosity and tenacity.
I don't listen to every episode of Presbycast, but I do listen to every episode of their Fathers and Brothers collection.

I've been continuing to use Refocus to block all social media, shopping, and distractions for much of the year. That has contributed to the growth of my /blogroll.

Work

Ironically, I had set this year's theme as "Look for Work," and boy did that hit with a vengeance.

The original meaning for me was partly about finding my next job, but also about just avoiding laziness: to have a mindset that always looks for the next thing to do… but when the layoff struck it collapsed down to finding a paycheck.

Even before the layoff, I've been beset with the eternal question of "what do I want to do when I grow up?" UX or programming? Management or IC? How does AI change any of this?

I'm still working on that. 🤷

2026

I'm hoping to spend some time with my wife planning the next year… at the moment the time has been coming too fast and furious so I haven't set a good Theme for 2026.

One thing is for sure, I know that the Lord will display his infinite faithfulness and love even more fully in this year than the last. Happy new year!

Finding Gene Cernan's Missing Moon Camera

2025-12-06T16:19:00-06:00

When you’re going to the moon, you’re assigned a camera with a 60mm lens that gets strapped to your chest to document samples, experiments, and the lunar terrain. Both astronauts get their own, clearly labeled with a sticker on the side: “CDR” for commander and “LMP” for lunar module pilot, but they weren’t mutually exclusive, often being swapped throughout the mission based on what film they were shooting with.

I love websites dedicated to a single purpose… a lifetime of attention to detail. This is such a treasure hunt. Also the stickers from their store are awesome.

Finding Gene Cernan's Missing Moon Camera — Space Camera Co. - www.spacecamera.co

Hollyland Lark M2S Lavalier

2025-10-06T13:35:00-05:00

I've been using a small USB-C lavalier microphone to enhance my daily audio calls. I'm a fan.

Since I deleted my microphone and light off my desk to just use an iPhone, I've wondered if there was a small microphone system that could slightly upgrade my setup. I have seen the small fuzzy mics held by YouTube personalities, but they were just a little too expensive.

After seeing the tags on David's private post I've been trialing the Hollyland Lark M2S microphone. I've been using it pretty much daily since April.

The biggest advantage of doing this is higher incoming audio quality coming from my Airpods Pro. There are better audio-specific reviews out there on youtube, and the Hollyland Lark M2S does perform very well, but the sheer advantage of getting more bits when someone has a bad connection or a hard audio environment… I'm finding myself reaching to plug in the USB-C dongle for the Larks as soon as I've asked someone to repeat themselves.

My only complaint was that the USB-C dongle blocks one of my precious ports, but it at least sports a charging pass-thru, so I haven't suffered too much.

The small USB-C charging case also fits perfectly in my office in a bag, which is of course a requirement. 🙌

Updated my blogroll

2025-09-11T07:27:00-05:00

My friend Scott was asking me for some healthy links to help replace doomscrolling, and I've been sending him a curated list of joyful RSS-friendly links, and I realized I should just put that somewhere.

Previously I was generating my blogroll semi-regularly from a tag in my Feeds.opml, but I kind of missed the ability to introduce the source and point out some favorites, so now the /blogroll is written by hand from a YML file and updated when I find stuff to send.

Just because a source is listed doesn't mean I agree with every jot and tiddle, just that I have either found joy or interesting thoughts from something there. Most of them are also independent creators who are living the indieweb life, and some are personal friends. I'll keep tweaking it.

Jujutsu's Whole Deal

2025-08-22T07:18:38-05:00

Part of jj's whole deal is that it collapses many Git concepts (stashes, staging, fixups, in-progress rebases, conflicts) into a single unified model of working with history, which then lets you use the same tools to do all of those things.

👏👏👏

Extremely good explanation. I've been programming for a living for the past few months and using jj anywhere that doesn't use git-lfs and just loving it.

Tech Notes: Understanding Jujutsu bookmarks - neugierig.org

Code is only half the System

2025-07-30T07:22:27-05:00

Chris's talk has been out there for quite a bit, but I'm clearing out my Instapaper and rediscovered it.

He argues that our rigor and formal methods must extend beyond the code and into the human systems that the code models:

Software is both the artifact that produces the program and the running program in the real world. If we miss either of those, we will end up in a really confused place. And by the same token, if we want to make our software good, we have to tackle both parts.

He then discusses how wisdom and discernment (which he calls mētis) is critical when working in the complex systems we are often called upon to model and improve… and how in the pursuit of "high engineering" (my phrase) we flatten a messy complex world in our code, and therefore have overly simplistic and incorrect models of the real world and therefore… bad software.

Fantastic talk. It's still sinking into my brain.

Seeing Like a Programmer (LambdaConf 2024) — Sympolymathesy, by Chris Krycho - v5.chriskrycho.com

Be There When They Look Up

2025-07-19T14:11:00-05:00

I love listening to people talk passionately about their life's work, and I had such a treat listening to Tyler Cowen's interview with David Robertson.

The section that I found most interesting had David speaking on the non-verbal interactions between orchestra and conductor.

One such gem was this: be there waiting when your team looks up.

The funny thing is, every single individual in an orchestra looks up at a different time. It’s totally personal. […] One of the challenges for a conductor is, as quickly as possible with a group you don’t know, to try and actually memorize when everybody looks up…

What a joyful and insightful thought. How often I assume all the other players "look up" when I would look up. This concept has really challenged me in parenting the past few weeks. As a parent if I get the desired behavior from my kids, my first thought is to turn from them to the next problem to solve. If I take the conductor's wisdom, then I want to be there waiting when they look up.

For those of us who aren't sure, David has a confident assurance as to at least one moment where we should be waiting to communicate: right after the action.

But the thing that no one will tell you, and that the players themselves don’t often realize, is that instinctively, and I think subconsciously, almost every player looks up after they’ve finished playing something. I think it’s to just check in to see, “Am I in the right place?”

That is a personal moment, where the orchestra is thinking about what’s coming. They’re not thinking about what’s behind because you’re constantly in the present and the future, but that’s your moment to send a message to somebody which can be, “Oh wow, that was incredible.” That sense of “Maestro came to my house” is really one of the things that I think helps build this kind of communication and trust between the players.

Even though I am relatively illiterate when it comes to modern classical music, David is full of useful insights on performance, teamwork, and leadership. Enjoy the whole engaging interview below.

David Robertson on Conducting, Pierre Boulez, and Musical Interpretation (Ep.248) - conversationswithtyler.com

Sketchnotes: Sloss Tech 2025

2025-06-27T19:48:00-05:00

I just finished up two days of Sloss Tech, mostly attending the Developer and Engineer track at the McWane Center. The BASE crew did a fantastic job facilitating the event, and as someone who entered the tech industry at Barcamp events long ago, I really enjoyed it.

There was a strong emphasis on (surprise!) generative AI. While my previous opinions have not changed greatly, I felt both excited to try new things and daunted by the organizational and cultural changes that will come with these new tools.

Here are the sketchnotes, along with a small takeaway that resonated with me.

Thursday

TypeScript: Mother Tongue of the AI Apocalypse

David Lormor (x.com/davidlormor)

If you want the AI to "speak" your favorite language, write good community docs and blogs.

Debugging Our Vibe Code

Mikhail Kozorovitskiy (x.com/MKOperator)

If you don't own it, don't ship it.

Same Threats – New Vectors: OWASP Top 10 for LLMs

David Hawthorne (davidhawthorne.com)

OWASP principles apply to LLMs. Treat the LLM's output as untrustworthy as a junior-dev.

Build Yourself Like You Build Systems

Alandis Seals (linkedin.com/in/alandisseals)

Be focused, and find healthy nos.

How AI is Changing Software Development Forever

Will Blackburn (linkedin.com/in/willblackburn)

Working with AI will be many developers first experience with delegation.

AI-Driven Engineering For Teams

Bill Abel (linkedin.com/in/billabel)

LLMs are sensitive to syntax and grammar nuances more than logic.

Play to Win: Engineering Joyful Tools That Compound Over Time

Praise Daramola (linkedin.com/in/praisedamarola)

In a talk near to my automation-loving heart, be a toolmaker!

Panel: New In Tech; The Struggle Is Real

Moderator: Elizabeth Anderson (linkedin.com)
Panelists: Mary Hendon Yelverton, Wendy Porter, Elizabeth Okunbor

Find the helpers.

Friday

Societal Thresholds

Zack Kass (zackkass.com)

Societal thresholds have been moved historical by distraction and entertainment.

Building Brand Loyalty, Customer Obsession & Memorable Experiences

Johnny Cupcakes (johnnycupcakes.com)

What are my sprinkles?

Bridging the Distance: Leading Effective, Distributed Dev Teams

Debra Hays (concertidc.com)

In sprinting to answer "how", we can forget to ask "why?"

Good Testing is Bad News

Sergei Gapanovich (linkedin.com/in/sgapanovich)

Love on your QA wizards. They are more than "uh, it's broken."

Back To The Future: What Year Is Your Software Org In?

Steve Hallman (linkedin.com/in/stevehallman/)

Sometimes it's good to just see a full menu.

Network Automation at Scale with Kubernetes and Crossplane

Jeremy Sanders (linkedin.com/in/jeremybhm)

Bespoke automation is tech debt.

What I Learned About Cybersecurity From Auditing Data Centers

Kim Cote (linkedin)

Maintain every layer as if it's your only protection.

From Microservice to Monolith and Back: How to Migrate React

Jared O’Neal (linkedin.com/in/therealoneal)

Do you have the org structure to maintain your architecture pattern?

Panel: Tech Roles in the Era of AI

Moderator: Keyona Meeks (linkedin.com/in/keyona-meeks)
Panelists: Paula Guevara, Raj Vedula, Ananya Thakur,

Nobody knows what's going to happen, but there will be humans. Soft skills matter.

For those who want to keep a copy of my notes from their talk, you can download a PDF here.

Thank you to Nibuild who sent me, it was real treat. 🙏

Achievement Unlocked

2025-06-20T17:55:00-05:00

This is an RSS-only post. It's a secret! Read more about RSS Club.

While listening to Derek Sivers on Focused I was struck by two things: how much I owe to Sivers for some indieweb accomplishments and generally celebrating plaintext… and how jealous I am of his incredible sive.rs domain name.

And now trave.rs is mine!

I've wanted it for quite a long time, but up to this point there's been a high price. The Serbian TLD used to require a local business license, but sometime while I wasn't paying attention that changed.

I'm definitely going to setup some emails for myself and my family using that domain. What I'm wrestling with now is whether I should make it this website's main TLD. It's been evantravers.com for thirteen years. Is trave.rs rad enough to warrant a change? I'd certainly have redirects to prevent linkrot, but I dunno. I'll sleep on it. 🤷

What do y'all think, should I make the big change?!

Hearing is Becoming

2025-06-09T11:10:00-05:00

Whose voice speaks to you the most?

It should be the people closest to us, the ones we say we love the most.

But it's not.

Historically the voices that built our worldview, exposing us to new ideas and cementing patterns of thought, were the voices physically nearest to us.¹

While this physical limitation has been slowly eroding since the printing press, the shift in influence has accelerated dramatically since we put iPods in our pockets.

For the first time in human history, two entire generations of people have grown to adulthood where voices outside our immediate community are the absolute primary force in forging our minds.

(The back half of my life is inside that window.)

I've taken a stab at mapping my life into seasons marked by the "voice most heard." I'm not talking about emotional impact, I'm not talking about person who means the most to me, I'm talking about raw audio hours. It looks something like this:

I am ashamed of this.

It's not that these particular external voices are bad. I'm very grateful for what I've learned from them. I'm ashamed of the sheer hours, the sheer a-mused² exposure that this (admittedly hasty and unscientific!) exercise reveals.

This simplistic map of my listening hours reveals far more about me than I care to admit.

When I look over the map, I see the ghosts of past selves. Versions of me with different goals, algorithmic filter bubbles³, even different accents.

The first time I became aware of this force, my parents asked me why I was talking strangely. I realized that I had subconsciously acquired Merlin Mann's humor… and his vocal ticks… through endless hours of Macbreak Weekly and other podcasts.

Regardless of who I claim to be my hero, my influences, my preachers… what I've learned is that whoever I allow to speak the most has the most influence on my life.

In a Girardian sense… these voices mediate my desires and mold me into what I aspire to be. This can be a good thing, and this can be a bad thing. Terrifyingly, this is a very invisible thing. If the hours spent listening don't have a huge emotional impact, we can be quietly shaped without even feeling the water boil around us.

It’s said that we are the cross section of our five closest friends… I’d rephrase that to say our five most listened to humans. I challenge you to take a similar exercise: working backwards from today, whose voice have you heard the most in each year? Not the most "impactful," we are looking for sheer count of hours.

This natural hierarchy mirrors the catholic doctrine of Subsidiarity: parents, family, village, state, world. ↩
Amusing Ourselves to Death by Neil Postman reminds us that "amuse" literally means "no think." ↩
My definition, created after watching the Social Dilemma ↩

A system rewards you with more of the content you engage with will self-reinforce until it eliminates other perspectives.

An Algorithmic Filter Bubble is incentivized to bring you content you engage with… not content that is morally good. The Algorithm doesn't care if the engagement is sadness, envy, jealousy, lust, or rage. Our fallen human nature triggers strongest on our basest emotional and instinctual desires.

Looking For New Opportunities

2025-06-06T09:26:00-05:00

I've been affected by a layoff at my previous employer, and I'm looking for new opportunities.

I'm still processing the suddent deletion of half my brain, but I am confident that it's the right moment and I'm excited about what will come next. I'm contemplating what skillset in which I wish to invest in this new stage of my career: do I stay in UX/Product, or re-visit my software engineering roots?

At this point, I want to work on meaningful products with nice people. 🤷 If you've been reading this for any length of time, you probably have a good idea of what drives me… probably better than the portfolio and resume that I ought to be updating right now.

If you have advice for handling this season, resume tips and tricks, or potential opportunities, I'd love to hear from you. Please shoot me an email.

Wrestling with Jujutsu

2025-05-03T20:56:00-05:00

I've been using git for about fifteen years. It's ingrained in my muscle memory. Why would I try anything else?

Jujutsu is a new VCS built on-top of git. It's worth looking into if you…

craft careful commits to form a good story because you know spelunking is hard.
manage long-running dev branches, -rerere rebases, or octopus merges.
are curious about what might come after git.
just like neat things 🤷
are a rustacean. 🦀

I'm not yet a rustacean, but all the others have been part of my life for more than a decade (except octopus merges. I've only had to do that once or twice!)

While I'm reeling from deleting some of the aliases that formed that git muscle memory (and learning new things while I do!) I've been learning about Jujutsu. Jujutsu first came to my attention when my git mentor Tom Miller raved about it¹. It came back when Chris Krycho posted his amazing writeup and also emailed me that it could solve some of my stranger git problems.

I played with it a year ago, and I immediately hated what it did to my local git repo:

…avers.com ! via  v3.2.1 via  impure (nix-shell-env) 󱗆 wtno | Working on the article took 13s
❯ git log --all --graph --oneline
* 86395da2 Working on the article
| * 5f73e7af
|/
| * 46a1cdba
|/
| * 50a93636
|/
* 0e9cb805 (HEAD, origin/draft) Drafting Wrestling with Jujutsu
| * cb863a74
|/
| * 3dee836c
|/
| * 7544ddc8
|/
* c0d593aa (origin/master, origin/HEAD, master) Remove lfs from actions
* c87a0994 unlfs
| * 2da3e478
|/
| * 2c58fe5f
|/
* ba8fec6b Remove lfs
* f9beb3fc Disable watcher so I don't have to type that again
| * 27e3fab0
|/
| * 2488ea80
|/
* a100e216 Typos

The reason for all these anonymous commits is jj's amazing jj undo. Every operation can be completely undone, and you can do even more with jj op.

Here's what jj log displays:

@  wtnowmoz evantravers@gmail.com 2025-05-04 17:32:28 86395da2
│  Working on the article
◆  zorywspm evantravers@gmail.com 2025-05-04 13:11:22 draft@origin git_head() 0e9cb805
│  Drafting Wrestling with Jujutsu
◆  wzonpurk evantravers@gmail.com 2025-05-04 09:05:35 master c0d593aa
│  Remove lfs from actions
~

jj log's default behavior is to only show you currently mutable working trees… which are basically various heads you have available. Everything else is considered immutable. You can rewrite it easily, but it's hidden from you to prevent you causing conflicts with other remote branches.

After some prompting I've been easing back into it and am really enjoying it this time. Playing with jujutsu feels very similar to playing with Helix: the muscle memory changes make "easy" actions frustrating, but the new mental model makes "impossible" tasks super easy.

All that to say, I'm having a grand time.

There are amazing resources out there for anyone interested in learning more, but here's a few tips that helped me as a novice wrestler:

as I said before, jj log by default only shows you "mutable" changes, which if you have no outstanding branches is basically just your current git staging². These changes are essentially your current "workbench," so as you work with the tool it feels more natural and git's "full history" feels heavy-handed, but it's confusing at first.
jj describe is available at point in your development workflow. You can jj describe your intention at the start of a change, or jj describe what you actually wrote at the end. Very forgiving, especially with jj split letting you take a patch and move it to a new change³.

…

I'm enjoying the learning process and using it daily on my blog and dotfiles the past week or so. As an individual software craftsperson, I like the freedom to explore the problem and not feel the pressure "oh I can't handle that in this commit, I'll need to handle it in another commit" because I can just jj split.

If I was doing release management, I'd really like its revision/merge management for handling complex rebases. I think it'd be really nice for some of the problems I used to face running long development branches.

Check it out!

References

Honestly, Tom raved first about mercurial, then pijul, and then brought jujutsu to my attention saying "this has all the good things from those other ones you didn't try." ↩
If you want to see all history you need to use jj log -r '::', or jj log -r 'all()'. There's a very powerful revision selection syntax that is very readable/scriptable but I haven't dived into that yet. ↩
While working on this article, I finally realized while reading the middleman code that I could hardcode in the argument for --watcher-disable and never have to type that again. I used jj split to move that change to a different revision than the one where I'm working on this blog post, and bobs your uncle! ↩

Jumping to Projects in Tmux

2025-04-30T11:27:00-05:00

It's nice to quickly set up your workspace. When coding, I am using sesh for this in tmux.

Years ago, I worked in advertising. In those days, it was common to have six to ten projects "open" at once—two in active development but others in a constant stream of small post-launch tweaks.

To keep things organized, I had a little bash script called tspace. I'm fairly certain Tom wrote it, but it used fasd to navigate to a directory, run tspace, and be set up.

In jobs following, I either worked on only one repo, or none at all. I haven't needed robust tmux session management for years, but the number of my personal repos has crept up to where I found myself frustrated at how often I ran the following commands:

Make new session
Navigate to highly used folder
Name session after folder

I had started using zoxide for "smart cd" (fasd is no longer in development) but I still wanted to close the gap.

I found two solutions: tmux-session-wizard is a one-click solution that adds almost no bulk to your config. Dead simple.

I'm presently trying sesh. Sesh is wildly configurable… but I'm using none of it. Sesh has the extra power of multiple interfaces (even a Raycast plugin!) and the ability to run commands on workspace launch… but I don't need that at the moment. direnv currently sets up the environment in the folder.

For many of you, this is old hat… but for anyone who hasn't set up a script like this: it's really pleasant.

Reorganizing My Nix Dotfiles

2025-04-17T15:54:00-05:00

I rewrote my dotfiles config again! 🤷

I had been feeling some friction: my folder/file organization smelled bad. There was a bunch of ../.././config… style paths that felt gross. I had files in folders by themselves like nvim/default.nix that made it hard to file find and was weird. Most importantly, I didn't really feel like I understood how it actually worked.

@mitchellh is very smart.

I found github:mitchellh/nixos-config I spent a fair amount of time reading through his config… pretty much assuming that the author of ghostty made all decisions intentionally and wisely. (It was a good assumption! He's been writing nix for long time.) His config is simple and clear. I did learn quite a bit, and accepted a lot of his decisions. I don't think I was using the home-manager settings for global and user packages before, I think that's speeding up builds.

Friendship ended with `git reset --soft` Now `git commit --fixup` is my best friend

While rewriting my WIP pile to be intelligent commits using my trusty git reset <sha> --soft I discovered a wonderful git tool I didn't know about: git commit --fixup=<sha> and git rebase --autosquash <sha>^1 lets you take the current staged change and effectively git commit --amend anywhere back in history. Y'all. The number of times I've git reset --soft HEAD^3 just to add something to previous commit #2? Oh my goodness. 🙌

vibe-coding works… inexplicably... until it doesn't

I tried a lot of vibe-coding via claude.ai. I used a snippet I found on stackoverflow of tail -n +1 **/*.nix | pbcopy to throw every nix file into a chat, and that produced much more effective results than my previous attempts of pasting isolated snippets. I got successfully running code… that was unreadable, not concise, and downright wrong in some places. I mostly used AI to get the code compiling, then while untangling the syntax jungle I would read the docs and find the right way to do it.

After vibing and copying from @mitchellh, I finally felt I can read and understand my nix config except for three dastardly lines. Originally, @mitchellh was using:

home-manager.users.${user} = import userHMConfig {
    isWSL = isWSL;
    inputs = inputs;
};

But that threw errors with my darwin-nix config… I imagine because I'm using more features in mine? Unsure.

Claude's vibe-coding suggested conflicting solutions.To fix nix-darwin not having inputs, claude suggested this.

home-manager.extraSpecialArgs = { inherit inputs; };

That worked for darwin, but broke nixos. Nixos's home-manager complained about not seeing lib, so Claude suggested this crazy thing:

home-manager.users.${user} = { pkgs, lib, ... }: import userHMConfig {
    inherit pkgs lib inputs isWSL;
};

(I'm having a hard time reading that… a function that takes pkgs, lib, etc., and imports my home-manager file, passing an attribute set that inherits pkgs, lib, inputs, isWSL from the surrounding scope? Is that needed?)

We vibing. 🤷

Voltroning the two solutions together worked, but I don't know why. It smelled bad. Eventually I read home-manager's documentation, understood that it only wants parameters passed through extraSpecialArgs and wrote:

home-manager.users.${user} = import userHMConfig;
home-manager.extraSpecialArgs = { inherit inputs; };

Vibe coding works… yes? Was it mostest correct? No.

How about speed? I ran three runs on each commit, trying to make sure I filled any caching systems and minimized network costs.

Before (git checkout c06716e, changing wsl.defaultUser to avoid exploding my system again):

dotfiles on  HEAD (c06716e) ! took 7s
❯ time sudo nixos-rebuild switch --flake ./
warning: Git tree '/home/evantravers/src/github.com/evantravers/dotfiles' is dirty
building the system configuration...
warning: Git tree '/home/evantravers/src/github.com/evantravers/dotfiles' is dirty
activating the configuration...
setting up /run/booted-system...
setting up /etc...
setting up /bin...
setting up /bin/login...
setting up /sbin/init shim...
reloading user units for root...
reloading user units for evantravers...
restarting sysinit-reactivation.target
Done. The new configuration is /nix/store/3rasc0zq4vbkhy5iajs150ihlbd4mlwh-nixos-system-nixos-25.05.20250403.2bfc080

________________________________________________________
Executed in    7.21 secs      fish           external
   usr time    1.28 millis  188.00 micros    1.09 millis
   sys time    6.80 millis  188.00 micros    6.61 millis

After:

dotfiles
❯ time sudo nixos-rebuild switch --flake ./#wsl
building the system configuration...
activating the configuration...
setting up /run/booted-system...
setting up /etc...
setting up /bin...
setting up /bin/login...
setting up /sbin/init shim...
reloading user units for root...
reloading user units for evantravers...
restarting sysinit-reactivation.target

________________________________________________________
Executed in    1.73 secs      fish           external
   usr time    4.30 millis  149.00 micros    4.15 millis
   sys time    0.02 millis   19.00 micros    0.00 millis

Not bad! I haven't done any A/B testing, but I'm willing to bet most of that is the home-manager.useGlobalPkgs = true and home-manager.useUserPackages = true settings my previous setup was ignoring. The speed is great, but the real win is learning more about the nix language and how it all fits together... if only by fixing vibed code.

I rather think much of my code career going forward is going to have "improving vibecode" as a theme.

Random Stuff:

If you change your wsl.defaultUser on the system that you are developing on, and run that nixos-rebuild switch, you are going to have A Bad Time(TM). I had some flashbacks to trying Ubuntu for the first time when I was a kid. I wanted that compiz wobbly windows stuff so bad and jacked up my graphics settings so many times. Probably the reason that I'm willing to put up with Nix now.
I discovered that my wezterm -> tmux setup on WSL was reliant on some .terminfo file I'd downloaded long ago. While reading through the wezterm nix code, I discovered that wezterm helpfully passes it's terminfo as a passthru so you can just use pkgs.wezterm.terminfo. I'm using that on wsl because ghostty isn't on windows yet.

The Lestrade Effect

2025-04-14T11:09:00-05:00

Just as Sherlock concludes his insightful first analysis of the crime scene, the hapless Inspector Lestrade comes puffing up, right on schedule, but too late. Sherlock turns away disdainfully, moving on to the next problem. Lestrade now faces spending the next hours re-discovering what took Sherlock a few minutes to discern.¹

This is all very funny and plays to Sherlock's vanity, but Lestrade has an actual job to do. Lestrade is not here for the joy of the sport… he actually has to collar the criminal in the conclusion.

I want to be Sherlock, but I often find myself playing the role of Lestrade. Like Lestrade I straddle the double diamond² of problem definition (evaluating the crime scene) and execution (arresting the murderer.) If your work domain connects definition and execution you are liable to become a Lestrade.

The Lestrade Effect is when someone must re-investigate a situation because they were not present at the first experience.

Whether in the gaslit streets of Victorian London or in a remote office, handoffs are costly. I see the Lestrade Effect often in emergency or experimental product design: When UX is not in the room in for early customer conversations where user needs are discovered then stakeholders (often Product and Marketing) can hand off User Stories to the UX team. The UX practitioners now must play Lestrade: repeating the same questions, making the same discoveries: who is the customer, what is the value of this decision, why did we decide this, etc.

If your friendly neighborhood UX practitioner is asking you questions that you (frustratingly!) have already asked and answered… it is possible that you have inadvertently cast them as Lestrade and not included them in the first analysis.

Be insightful as Sherlock, but be considerate like Watson. Include everyone in the room when the problem is being discovered… or risk leaving colleagues trailing behind.

You can find Lestrades all over any fiction with a eccentric and brilliant problem solver: House MD, Bones, Fringe… pick your poison. Pretty much any media where one character has a Eureka Moment and runs off to save the day without enlightening their colleagues (thereby continuing the suspense for the audience) has potential for Lestrades. However… continuing suspense is not helpful for non-fictional collaborative work. Don't model that. ↩
, originally posted in Opportunity Solution Trees ↩

Portable Standing Desk

2025-03-20T12:16:00-05:00

I've wanted to use my keyboards outside for a little bit. The seed of the idea was struggling to work with my mobile office in a "workspace" at an airbnb with only a couple inches of countertop at a very uncomfy height.

What if I could just bring my desk with me?

I experimented with just buying a laptop/projector tray and sticking it on a lightstand I had lying around. Very wobbly.

I eventually found the Intention Design Tripod Table. While there are other options¹, the Intention Design offering is a pretty safe bet for those who want to stand outside and work on their computer.

I'm pretty sure that the tripod Intention Uses is 99% identical to this smallrig option on amazon. If you are mechanically inclined, you could install a tripod socket in a butcher block or cutting board and build your own for about 50% the cost.

Intention offers a keyboard tray, but I opted to build a weird little arm thing using SmallRig parts:

SmallRig Super Clamp to attach to the tripod body
NiceyRig 2.5" rod for a little bit of standoff
2x SmallRig Universal Magic Arms with Small Ballhead creating articulation between 2.5" rod and Cheeseplate, and articulation between Cheeseplate and Magsafe disks
SmallRig Cheese Plate mounting to space out the split keyboard sides
SmallRig 1/4" to 1/4" adapters for the rods
Minifocus Magsafe Adapter to mount to my existing Magsafe Enabled keyboards

Mounting the keyboard at a tented neutral arm position does create a little bounce in the laptop, but not much more than a normal-ish desk. I've thought about putting some hanging weights to counter this motion, but haven't played with it yet.

I honestly haven't figured out a good pointing device solution here yet. I will likely buy a cheap trackpad and put a magsafe disk on the back. I'd like to tripod mount my Bitra Trackball, but as they seem to have stopped making them, I don't want to risk destroying it. For now I'm using a combo of Homerow.app and ZMK's pointing devices, I already have a lot of practice just using the keyboard.. I can't really design on it, but it's a great focus writing setup.

Other things I found: ↩

Formwork and Minimum Viable Artifact

2025-02-25T10:21:00-06:00

The permanent foundations for a bridge are molded by impermanent forms. Plywood and other flexible formwork creates shapes in which the concrete hardens. No wise worker over-optimizes their forms: Formwork is created for the pour then discarded.

The formwork and the concrete pour both require skills and craft. Both are connected to the final bridge.

therpf.com¹ is a community obsessed with the production design of objects inhabiting the worlds of film. Screen-used Lightsabers, Voight-Kampff tests, Tricorders and more are studied, screengrabbed, and discussed. Screen-used props are divided into two categories: incredibly detailed "hero props" for closeups, and rubberized stunt props for action sequences.

The hero prop sets up the world, the stunt version sells the story.

Brad Frost has talked about Minimum Viable Artifact for a while: Minimum Viable Artifacts are the least amount of work required to communicate.

The plywood formwork is a Minimum Viable Artifact. So is the stunt prop. The hero prop has far more detail, but just enough to tell the story. The lightsaber hilt, made of microphone mounts and greebles, can't actually cut through metal… but we believe that Luke is a jedi.

A Minimum Viable Artifact carries enough detail and craft to serve the outcome: a pour for a bridge, or a close-up of the hero's armament… and no more². While there is a joy to craft and skill, it should be in service of a goal.

Otherwise we spend our life on expensive artisanal plywood forms and build no bridges.

TheRPF is one of my favorite places on the whole internet. I could spend hours learning from them. While I am not a maker, I have bought a few small replicas off the amazing folk on the forum. One of my favorite internet moments was when a forum-user discovered that the original hero prop for the DL-44 Han Solo Blaster was used in a movie from a decade before, and that they all had to rebuild their exact replicas after discovering it was based on the flat-bottom barreled artillery Mauser. As a lover of detail and craft… there are endless incredible threads. ↩
The desire to expend effort doing something we are good at rather than moving the needle on our goals is a form of The Resistance from The War of Art by Steven Pressfield. ↩

Raycast Focus

2025-02-13T06:29:00-06:00

Last month Jesse texted me:

Jesse

Have you tried Raycast Focus yet? Pretty neat.

I have not yet

Sheesh

That's my headspace idea, only maybe better

After invoking the mode Raycast Focus provides a lovely little timer with your stated goal and uses Screentime to block apps and websites in every browser.

Look how elegant that is!

It's almost exactly what I built as Hammerspoon Headspace. only missing the Toggl timer integration and the ability to save preset "focus" presets. I had been working on my version of this, but their implementation is so good.

I think it's possible that with their arguments system you might be able to pass presets to the Focus system, and that would let me integrated it into my shortcut-based Headspaces. I just haven't decoded their JSON yet.

As it is, if you are the kind of person who was interested in Headspace but really only wanted one focus mode for your deep work: this is an amazing free tool that you really should try.

Boox Palma Review

2025-02-11T11:34:00-06:00

I wanted a calm device to reduce my screentime. A "calm device" for me means minimal notifications, no hyper stimulant applications, and even no browser. Simply reading and writing.

While I prefer Olivetree for study, I have been using Youversion with my small group to do a read-through-in-a-year plan. Being able to scroll with the physical hardware buttons on a device with no other notifications has been helpful for keeping up.

There have been a wealth of new e-ink devices¹ in the past year. While I would like to replace my distracting iPhone outright, I use my phone for my webcam. As the new Boox Palma 2 has just released, I bought a used refurbished Palma on ebay for just over the price of a Kindle.

The Palma works ok out of the box but it only shines after nerding out on configuration. There is a whole custom subsystem just for handling how monochrome e-ink renders the underlying colors. Some apps display such that you can't see the scroll bar on Android 14, so they are very hard to configure. It is not an easy software package to love.

But I do.

Here's what I did to set up my Palma:

Niagara Launcher

For maximum minimalism: bigger sans-serif font, minimal icons, hiding status and notifications bar.

Gboard

The default keyboard is good, Gboard is much better: swipe to type, great dictation support. I am really liking dictating notes while I read physical books.

Obsidian

Let's be honest, I just wanted a kindle that ran Obsidian.

I've got the light mode theme of Minimal Theme set to their e-ink setting, and it works well. I do wish that I could set the theme or theme settings per device easily.

Kindle

The Kindle.app just works. The only wrinkle is that unlike all the other apps, you need to set the hardware buttons to be volume buttons, and use Kindle's settings to turn the page with the volume keys.

Olive Tree

I have set Olive Tree to scroll the page using the physical buttons. There's an in-app setting to keep the device from sleeping while up front, which means I never have to do the "stay awake" tap during the whole sermon.

Bible by Youversion

I don't love the social network aspects, but it's the best for reading plans, and I'm enjoying going through a reading plan with my small group.

KOReader

I don't have much content to read here, but there's a system that allows it to sync to a calibre library over wifi and that's nifty.

Instapaper

I switched back² to the first read later app I ever loved, and it's still great. I abandoned it only because I believed I wanted unlimited API access.

The device is comfy to hold. Not too heavy. The screen is clear enough, I tend to run it 100% warm and with the lowest backlight. I get a couple days battery life out of it. Maybe a week.

The official folio case is not amazing. The magnetic clasp is right in the way of the volume buttons that I use for "turn page," and once a day the cover jams the button and flips madly through the book like a runaway train. I actually like that it's too big for a pocket - this forces me to put it down. It certainly does not hurt that I can prop it up for writerdeck roles.

Isn't that a cute little typewriter? In use here, the Bancouver40 by Sporewoh. I was hoping to have built an Endgame by now, but it hasn't arrived yet.

Since buying the device, I've already finished two digital books and 70% of a physical book, so the experiment is a success. Turns out having a reading-specific device is increasing my reading, imagine that.

I have had my eye on: ↩
- Daylight Computer. The most interesting "tablet as main computer" device.
- Minimal Phone. Android as E-ink, physical keyboard.
- Mudita Kompakt. Modified Android for Calmness.
- Supernote. Maybe the most interesting art-tablet option.
Instapaper -> Readwise -> Reeder -> Omnivore -> Reeder -> Instapaper. ↩

Automatic Webmentions

2025-01-24T22:41:00-06:00

This is an RSS-only post. It's a secret! Read more about RSS Club.

Thanks to Matt's post, I think I've finally got webmentions automated, which is lovely. Thanks Matt!

This blog's comments are powered by Webmentions - steele.blue

Beware and Be Encouraged

2025-01-14T10:07:00-06:00

While researching a trip, I came across a devious strategy: Thieves sometimes put up a sign in english that reads "Beware of Pickpockets," then they wait. English-speaking tourists read the sign and invariably place their hand on their most expensive possession. Having found the most valuable target, the thieves make the snatch.

When something precious to us is threatened we impulsively check on it. Some scriptures trigger that reflex on our most precious possession: our salvation.

Hebrews 6:4-8 triggers this reaction in me especially the words "who [...] have tasted the heavenly gift, and have shared in the Holy Spirit, [...] and then have fallen away." (Hebrews 6:4-8, ESV, emphasis mine)

This passage summons disquieting emotions. However, that instinctual "have I still got it?" is but a confirmation of the preciousness of the salvation which we already have been given. The tourist pats his pocket because there is something precious within.

If we had nothing to treasure, if we were still dead in our sins… we would not have the instinct to check and protect. Be encouraged.

Setting Up Obsidian Tasks as a Things.app User

2025-01-06T14:02:00-06:00

For the past month, and through the last quarterly review and planning of 2025, I've been using Obsidian Tasks instead of my trusty Things.app, and I am very happy.

Obsidian Tasks lets me manage my tasks in the context of my work. It lets me fully customize how and where tasks are stored and presented, allowing for flexible and focused workflows.

It's also very nerdy.

I was hard pressed to choose a successor to Things. For me it came down to taskwarrior, Todoist, Bullet Journal, and Obsidian Tasks.

I chose Obsidian Tasks for a few reasons:

I do all my writing and thinking, journaling and rituals in Obsidian. I found myself constantly linking back and forth from Things to Obsidian, or wrestling with where to store the notes and context.
I have become project-list focused during my workday. I wanted a kanban to manage lots of projects, and I can easily do that in Obsidian.

There are downsides I'm willing to accept:

Slower performance. It's honestly surprising how fast it is considering what it is doing, but sluggish compared to Things.app. I do notice it especially in search.
It's yet another task format, when there is already orgmode, todo.txt, and taskpaper. I believe I understand the Dataview-driven decision, but I do wish it used a known format for attributes.
I don't think I'll be able to easily put time blocked tasks on my calendar like I can with Reminders or Todoist. Since I tend to assign a project to a day, this is ok.
Push Notifications with Obsidian are hard or impossible. I don't rely on as many notifications and nudges¹ for project work. For general scheduling, I have my Daily Note.

A quick overview of what I've done since I began thinking about it a year ago:

I investigated Obsidian Tasks, choosing plaintext over emoji, and seeing if it had the ability to do what I wanted.
I removed old instances of - [ ] from my markdown repo to "clean it up".
Six months ago, I disabled the quick capture shortcut from my phone and started just using Drafts for all capture.
3 months ago I developed a Things -> Obsidian Tasks export, first in Shortcuts and then thankfully in ruby.
I hacked on Obsidian Tasks queries to mirror the screens in Things.app that ran my life.
Two month ago, I moved the Nudges to Reminders, and finally the Things and Obsidian experience converged.
One month ago, I ran the export for the last time, changed my keyboard shortcuts for capture to a new custom shortcut, and since then I've been 100% using Obsidian for tasks.
I've continued to write some new views, merge my ritual documents with their datasources, and discover new powerful workflows.

Configuration and Import

I tried writing this in Shortcuts, but I ended up with the structure I had written for orgmode and taskwarrior and just used JXA. Much easier.

There are a lot of different ways you could configure Obsidian Tasks, but I made a couple executive decisions:

I prefer readable plaintext. I chose the Dataview format rather than the default Emoji format.
I imported files that corresponded to the Areas and Projects lists in Things into projects folder.
I wrote conversions to let tags import well.

I also spent some time going through my wiki repo, cleaning up wherever - [ ] appeared. I chose to represent every - [ ] checkbox as a task, and not use the - [ ] #task form. I like the simplicity.

While I was trialing the import I developed Obsidian "views" that corresponded to the screens in Things. I kept tweaking them until they consistently matched their corresponding Things.app views.

Obsidian Templates

Obsidian Tasks lets you create custom blocks using either a SQL-style language or javascript. Assuming you've made similar configuration decisions in your repo, here are the Things.app-style queries that drive my current Task Management system, although it's still evolving:

Inbox

(Literally just a file named Inbox. 🤷)

Today

I used to visit Today daily to see what I had assigned for that day.

> [!warning]+ Overdue
> ```tasks
> not done
> ((starts before tomorrow) AND (has start date)) OR (due before today)
> ```

```tasks
not done
((scheduled on today) OR (scheduled before today)) OR ((starts before tomorrow) AND (has start date)) OR (due today)
```

I am finding that I use Today less as I have pushed some of the actions I once performed on it into Daily Plan.

Tomorrow

```tasks
not done
scheduled before tomorrow
```

Workday Todos

> [!error]- High Energy
> ```tasks
> not done
> tags include High
> filter by function task.file.property('tags').includes('#Areas/Work')
> filter by function !task.file.hasProperty('status') || task.file.property('status') == "In Progress"
> ```

> [!tip]- Low Energy
> ```tasks
> not done
> tags include Low
> filter by function task.file.property('tags').includes('#Areas/Work')
> filter by function !task.file.hasProperty('status') || task.file.property('status') == "In Progress"
> ```

> [!attention]- Long Estimates
> ```tasks
> not done
> tags include 4h
> filter by function task.file.property('tags').includes('#Areas/Work')
> filter by function !task.file.hasProperty('status') || task.file.property('status') == "In Progress"
> ```

> [!todo]- Quick Wins
> ```tasks
> not done
> tags include 30m
> filter by function task.file.property('tags').includes('#Areas/Work')
> ```
```tasks
not done
filter by function task.file.property('tags').includes('#Areas/Work')
filter by function !task.file.hasProperty('status') || task.file.property('status') == "In Progress"
group by function task.filename
path does not include Wishlist
```

Upcoming

```tasks
not done
has scheduled date
group by function task.scheduled.format("MMM Y")
sort by scheduled date
```

Deadlines

```tasks
not done
has due date
group by function task.due.format("MMMM")
sort by due date
```

Anytime

```tasks
not done
no scheduled date
no due date
filter by function task.file.property('status') == "In Progress"
folder does not include booknotes
group by path
```

Someday

```tasks
not done
group by path
tag include Someday
```

Quick Capture

While now I need to maintain yet-another-crufty-shortcut, it works well enough and thanks to my hammerspoon layer, it uses the same shortcut.

While I thought I had weaned myself from using Things.app'a quick capture, I quickly missed having that power. So I whipped up a custom Shortcut and bound it to the same keys. It's not as good… but in some ways it's better.

I used Claude.ai to generate some of the AppleScript blocks, then I fixed them in Script Editor.

Firefox is the bane of my Quick Capture existence. I may try running Hammerspoon Lua eventually.

Obsidian Kanban to Obsidian Projects

All a "project" is a note with the tag `#project`. So I can have lists of tasks, a home repair, or an upcoming essay or blog post.

I started² using Obsidian Projects because it lets me populate the board's based on a tag. Think of it as a pre-made fancy Dataview view, it's not a real file. This causes some inconsistencies in the system, but it serves my needs well. I do wish it was just plaintext, but I don't see how.

I have mine set to be columns based a status property. When I drag between columns, it updates the property.

I like Kanban as a way to visualize the intensity of my commitments. I missed that in Things.app, and it's part of what drove me to try this experiment.³

Neat Things I Didn't Expect

Dedicated "Views"

I built a custom view for workday tasks:

> [!error]- High Energy
> ```tasks
> not done
> tags include High
> filter by function task.file.property('tags').includes('#Areas/Work')
> filter by function !task.file.hasProperty('status') || task.file.property('status') == "In Progress"
> ```

> [!tip]- Low Energy
> ```tasks
> not done
> tags include Low
> filter by function task.file.property('tags').includes('#Areas/Work')
> filter by function !task.file.hasProperty('status') || task.file.property('status') == "In Progress"
> ```

> [!attention]- Long Estimates
> ```tasks
> not done
> tags include 4h
> filter by function task.file.property('tags').includes('#Areas/Work')
> filter by function !task.file.hasProperty('status') || task.file.property('status') == "In Progress"
> ```

> [!todo]- Quick Wins
> ```tasks
> not done
> tags include 30m
> filter by function task.file.property('tags').includes('#Areas/Work')
> ```

```tasks
not done
filter by function task.file.property('tags').includes('#Areas/Work')
filter by function !task.file.hasProperty('status') || task.file.property('status') == "In Progress"
group by function task.filename
path does not include Wishlist
```

This only shows tasks for In Progress projects because of the line task.file.property('status') == "In Progress". If I drag the parent Project to "Done" or "Blocked," child tasks disappear like magic. I love this.

In Things.app, the closest thing I had to dedicated Views was to save a URL that contained tags I wanted to search for: Work and High Focus. Obsidian Tasks lets me dataview query all kinds of interesting and custom views.

One neat one (inspired by bookworm) is "action items from books I've read":

All this lovely information with only…

```tasks
not done
path includes booknotes
group by path
```

Another is being able to show child projects of an "Area" style project:

```tasks
TABLE status, due
FROM #Areas/Personal AND #project
WHERE status != "Done"
```

Obsidian's tooling and plugins allows for even more creative custom UIs… I could embed a set of views into a Canvas for a productivity dashboard if I wanted to!

Reducing Ritual Friction

I've been using this same power in my daily rituals.

Here, the question of "protect the essential" tasks of the day is made easier by filtering today's list to just `#Focus/High` and tasks with elevated priority.

I've been abusing Obisidan's callouts system to hide the extra filters so it's not overwhelming.

My rituals are typically prompts and questions like "what projects have we scheduled for this month" or "what questions should you ask your spouse?" Before, I'd read the prompt and then go find the answer in the system by searching projects by scheduled date or tasks tagged "Wife." With Obsidian Tasks, I can embed those custom views right into the ritual.

- ❓ Any questions to answer?

> [!question]- Questions for Wife
> ```tasks
> not done
> tags includes Family
> group by path
> ```

This has been a game-changer… much less app-switching when I'm trying to just plan the week with my wife.

Power of Plaintext

Because Obsidian Tasks is all plaintext… there's just so much that's possible:

You can search for todos natively in Obsidian using task:<string> or even task-todo:<string> for uncompleted todos.
I can filter tasks to only projects marked "In Progress" in my kanban: filter by function !task.file.hasProperty('status') || task.file.property('status') == "In Progress"
Any note can become a project on the Kanban… this blog post is a project right now.
I can leave # FIXME style tasks on any writing project anywhere in my notes, and they show up in my task lists. This post had - [ ] tasks all over it while I worked on it.
I love being able to combine documentation and todos. I now have documents for my vehicles containing both a log of repairs, and tasks for what to do next.

Some useful plugins that I've found helpful:

TextTransporter let's me push a line to another file. It's not as good as orgmode's refile action, but it's close. Super handy for processing the Inbox into project lists. I have some of the TT actions under Leader Key plugin under ⌘+.
Outliner has some shortcuts for moving around bullet points that makes reordering a task nice
Omnisearch makes searching much faster.

…

Obsidian Tasks is definitely not for everybody, but it is for me. Who else might it be for?

spiritual folks who want to include quotes among their task system to remember a higher calling.
writers who spend a lot of times deep in a document, and want their tasks to link to a place in that document.
people who stare longingly at orgmode and want something approachable and mobile friendly.

I'm delighted. I have a system that is free and flexible, on all my devices. It lives among my writing and my work, and can be summoned to appear magically among my rituals documents for minimal friction. I can make custom and detailed overviews. And it's still growing.

I talked a little about how I define nudges in the last post. ↩
I had been using Obsidian Kanban to manage writing projects for the past few years. I love that it's plaintext under the hood, and you can easily script it or modify it. ↩
Yes, I know Todoist.app has this feature now… but then I'd still be bouncing back and forth between context and tasks, and a lot of this is reducing that friction. ↩

Aposiopesis Aggression

2025-01-03T20:45:00-06:00

This is an RSS-only post. It's a secret! Read more about RSS Club.

Recently someone at work said "people around here use the ellipsis a lot." (I smiled, and nodded, knowing that that almost certainly meant me.)

Then they added… "and to me an ellipsis always means stress and anger."

I was horrified.

I've used an ellipsis… (steps away to count using rg and vim) 1,317 times on my blog. Have I been threatening y'all?!

I honestly don't know what to do. I have always loved the ellipsis as a pause, to represent my internal self-conscious search for the right words.¹ I believe that I picked this up from dialog in novels I loved². (I'll probably dig into my bookshelf to see which ones used the device.) It was there that I picked up the belief that … meant a gap in racing thoughts, a self-conscious and humble admission that the right words do not come easy to me.

The wikipedia article calls using an ellipsis at the end of a sentence as Aposiopesis as "is the use of an ellipsis to trail off into silence." It's associated with strong emotion like fury or modesty.

For me it's mostly been modesty.

Now I find that my intended humility is read as… dismissive. It's very associated with boomers and older.³

For those to whom I have miscommunicated since I discovered that ⌥-; sends the … character: please forgive me. It was done unintentionally.

I'm probably going to attempt to erase it from my daily work text communication. I'm not going to stop using it in my personal writing. It's become part of my voice that I like… and it'd be hard to erase that.

So…

Some sources seem to support my thinking, like this article from editorsmanual.com which calls them suspension points: ↩

The three dots that make up an ellipsis are also called suspension points. In informal writing, these dots can indicate indecision, an incomplete thought, or a pause. The ellipsis then acts as a punctuation mark that indicates a pause.
I'm still looking through my old books to see which ones may have had this effect on me. Close at hand, The View From Saturday by E.L. Konigsburg has this pattern. ↩
In my mind, I've always read and spoken as if I didn't fit in my generation so I suppose that fits. ↩
- https://www.purewow.com/tech/why-do-boomers-use-ellipses
- https://theoutline.com/post/3333/why-do-old-people-text-like-this-an-investigation

✍ Evan Travers

Committing with the last command as the description

AI in the City of God

2025

Spiritual Life

Relationships

Health

Creating

Sketchnoting

Reading

Work

2026

Finding Gene Cernan's Missing Moon Camera

Hollyland Lark M2S Lavalier

Updated my blogroll

Jujutsu's Whole Deal

Code is only half the System

Be There When They Look Up

Sketchnotes: Sloss Tech 2025

Thursday

Friday

Societal Thresholds

Building Brand Loyalty, Customer Obsession & Memorable Experiences

Achievement Unlocked

Hearing is Becoming

Looking For New Opportunities

Wrestling with Jujutsu

…

References

Jumping to Projects in Tmux

Reorganizing My Nix Dotfiles

@mitchellh is very smart.

Friendship ended with git reset --soft Now git commit --fixup is my best friend

vibe-coding works… inexplicably... until it doesn't

The Lestrade Effect

Portable Standing Desk

Formwork and Minimum Viable Artifact

Raycast Focus

Boox Palma Review

Kindle

Automatic Webmentions

Beware and Be Encouraged

Setting Up Obsidian Tasks as a Things.app User

Configuration and Import

Obsidian Templates

Inbox

Today

Tomorrow

Workday Todos

Upcoming

Deadlines

Anytime

Someday

Quick Capture

Obsidian Kanban to Obsidian Projects

Neat Things I Didn't Expect

Dedicated "Views"

Reducing Ritual Friction

Power of Plaintext

…

Aposiopesis Aggression

Friendship ended with `git reset --soft` Now `git commit --fixup` is my best friend