Michael Crenshaw, programmer

I'm a full-stack developer for Crutchfield Corp. I strongly believe in writing beautiful code, developing accessible solutions, and giving back to the coding world. The easiest way to reach me is @m4c9416 on Twitter.

Jumping to conclusions

Jumping out of the dark and shouting "BOO!" at a friend is fun. Why? Because they start, maybe scream, generally act scared. Then of course, they realize that instead of a menacing ghost it's a good friend, and you both enjoy a good laugh.

Humans jump to conclusions, and they do it quickly.

If you're a fan of evolutionary biology, you might believe the tendency to jump to conclusions emerged out of a higher survival rate for those who did. If your brain says "I think I see a bear," its time is better spent calculating escape routes than contemplating, "Well what if it's just a funny-looking tree?"

If you believe god(s) made the human mind pretty much the way it is, it's reasonable to suspect she/he/it/they designed the brain to think fast and therefore live longer and more happily.

As web user experience designers, we tend to think of "experiences" as activities such as "switching tabs" or "clicking the play button."

But some experiences are moments rather than activities. Your mind sees an image and, by nature, must draw a conclusion.

If the image presented by a user interface is likely to be interpreted in a negative way ("This interface is ugly/confusing/irritating."), it doesn't matter if 500 milliseconds later everything looks great. Your interface yelled "BOO!", your user is scared, and the rest of the experience is contaminated by that negative feeling.

In some later post I'll argue that progressive JPEGs (images that initially load low-quality but then improve) are perceived with more negative emotions than baseline JPEGs (images that load top-to-bottom) because of the conclusions to which the user jumps in the first few milliseconds of the loading process.

For now I'll simply assert: user experience must be designed with an eye toward moments (even ones that last mere milliseconds) because of the human need to jump to conclusions.

Reviving Keryx

Keryx is an offline package manager for Debian-based operating systems.

On my previous blog, I wrote "Reviving Keryx" announcing my intention to refactor and bring the project back to life.

But then life happened. I haven't pulled up the Keryx source code in a long time.

Keryx provided a huge service for the Linux community, impacting people around the world. I wish I were in a position to continue development.

If you've landed on this page and need something like Keryx in your life, please hit me up @m4c9416 on Twitter and let me know. If there's enough continued interest, I might carve out enough time to update the code base.

Developing from my phone

I installed an FTP client and text editor on my phone. Writing HTML is a bit unwieldy with the mobile keyboard, but it works.

Once I'm back to my keyboard, I'll attempt to make individual "pages" out of my blog posts, using the history API.

I'm thinking all requests will have to be served index.html via server-side rewrites and particular posts shown according to the URL by JavaScript. I'm still not sure how to handle 404 errors.

Update: I'm back at my desktop and wouldn't want to develop from a mobile device for extended periods of time.

Service worker struggles

I like Chrome's Lighthouse tool. So much so that I set up a service worker for this site, just so I could earn 100 scores across the board.

I used a boilerplate service worker. But didn't really know how it works.

So now the worker is caching index.html, and I'm not sure how to bust the cache. I incremended the precache version, but no dice.

Oh well, I'm learning.

Update: the boilerplate has a runtime cache which caches everything. Busting the precache won't update index.html. So I removed the runtime cache logic, and now the page updates when I increment the precache version.

New website, day 2

I broke down and added styles. Sorry it's so monotone. I'm not a designer, and I can sense when I'm out of my depth.

I also put this site on a free Cloudflare account. It's easy, so why not.

There's something therapeutic about writing plain old HTML.

New website

I got an email today that my site had malware. It was a Wordpress site that I didn't really maintain, so I guess I shouldn't be surprised.

So I said "forget this" and deleted my blog.

I've been meaning to rewrite it for a long time. But each time I started, I'd think "I should really version control this. And then automatically deploy on a commit. But to a staging environment first, then to production. Should the staging environment be behind a login? Also I'll need to bundle styles and scripts and minify everything and serve it from a CDN."

I'm exhausted just from typing it. So I'm writing up this page in gedit, and I'll upload it as index.html in FileZilla. Why? Because those tools are the ones I have handy. Maybe this site will some day be more complex. But for now? It's easy.