What has the Good, Big Book™ blessed me with this time? Especially where those blessings were *not* books?

1. *Love is Not Enough* - Converge

=> Bandcamp: *Love is Not Enough*

I know Converge is a juggernaut in the scene.

They're you're favorite bands' favorite band.

That does not change that my appreciation for Converge is sporadic at best.

This is the first record from Converge that for me has really worked basically end to end. That doesn't mean I love every song, but where it hits... oh boy, IT.. HITS.

The opening tracking, "Love is Not Enough", and the second-to-last track, "Make Me Forget You", were instant "clicks." And then there are tracks that aren't as 10/10 for me but are still total head bobbers: "To Feel Something", "Force Meets Presence", and "We Were Never the Same."

I knew the record was a hit for me when I went to return it to the library and renewed it instead.

Look forward to me eventually getting into a car accident because I've enthusiastically drummed on my steering wheel too hard, triggered the airbag, and plowed into a ditch. Knock on wood I don't hurt anyone else.

2. *The Blue Nowhere* - Between the Buried and Me

=> Bandcamp: *The Blue Nowhere*

Not my favorite, but a great "weirdo" record.

It really reminds me of the great chasm between progressive metalcore (my usual comfort zone) and progressive metal.

If progressive metalcore is all-black-tech-ninja-fashion clad dorks playing through "Axe Fx II", progressive metal is dorks playing technically proficient guitar to distract you from the synth-flute bridge that they *actually* wanted to show you.

What I said to a friend was, "this is not a neg, but this record is very silly." I'm all about silly.

Anyway, it was "fun times" even if it wasn't "long times." Thanks BTBAM.

3. *Smoking Behind the Supermarket With You* - Jinushi

=> KCLS page

I adore this series. It gets its tone *so* right for me. The feeling of gentle intimacy in the wee hours of the nights and mornings is expressed so well.

It's sort of an age gap romance, but the vibe is more "age gap crush."

I am nearing being caught up and honestly dreading it.

If you like slice of life or light romance, I would recommend it. I've written so little about it here and I worry it betrays how much I've latched onto it. It's good. good good good good good.

4. *The Honjin Murders* - Seishi Yokomizo

=> Wikipedia: The Honjin Murders

I have not finished this book, but I'm having a blast.

It really feels like a novelized episode of Detective Conan. I find the author has a real talent for explaining how a space is laid out without confusing me. I don't visualize things in books well, so to me this is a real show of authorial talent. Good job Yokomizo-san and Louise Heal Kawai (the translator.)

I look forward to reading this as my "fun" fiction reading. I'm trying to read more fiction that is just there to tickle your brain and entertain you and less to be about "saying something."

I think of it like a penpal postcard exchange: send people LLM-appropriate queries through a P.O. box proxy, receive your own, fulfill those prompts with all the meatbag personality you can muster, and cherish the responses you get in return.

One of my prompts to fulfill was to write a short story. I've wanted to try my hand at writing, and this felt like a sign!

With some sadness, I did not complete the prompt. I spent about a month drafting, and never quite landed something I was totally happy with. So, the story has no conclusion and abruptly ends nearly mid-thought.

I'm sharing what I did write here, because even if it's incomplete, I think it's better this way.

If you'd like to read the original prompt, I'll put it at the end of the post. Jump to... Original prompt

Without further rambling preamble... here it is!

Companyon

Tomorrow would mark two years. Two years since Isaac had boarded TransSol's courier starship with the pilot's certification he felt had been issued too easily. Two years of routine inventory inspections and system diagnostics. Two years of rigid diet and exercise he could never have maintained on his own back home. Two years of the self-sanctified ritual he found himself currently practicing: laying in the observation bay flat on his back, taking in the star-fields, and cataloguing constellations that he never seemed to be able to find two days in a row.

The prevailing wisdom had been that a view of space from space would fundamentally change you: the "Overview Effect" they called it. But it was important what you were viewing, Isaac learned. Viewing your home from that specially distanced orbit flooded you with reverence for the human experience. Its insignificance made significant by juxtaposition with the void engulfing it. But to stargaze so far from home and so regularly was a different experience entirely. Day in, day out, Isaac saw stunning fields of utterly depersonalized stars. He should have been the insignificance made significant, but no audience of stars could make his world aboard the courier ship feel significant. It was just Isaac and his inventory. But even still, time in the observation bay was time well spent. Time well spent alone.

Behind Isaac, the observation bay door opened flooding the room with soft light. Isaac craned his neck back squinting to adjust his eyes.

"Need something Zac?"

"Just the usual interruption. Time to do our rounds, Isaac."

Correction: It was Isaac and his inventory... and Zac.

Two years ago, Zac had been paired with Isaac to ferry TransSol's oh-so-important cargo between two of their long-distance shipping hubs. Well, "paired" was a bit of a euphemism.

TransSol had been in the intra- and inter-stellar shipping game since the earliest days and learned quite quickly that their pilots and couriers ran into certain... psychological challenges... the longer the routes got. Stellar couriers just could not muscle it out like their 21st century, truck driving forebears. Extended isolation did something to people and it always jeopardized the delivery. At first, TransSol, just relied on staffing all deliveries with crews of 2 or 3, but this was costing them 2 to 3 times the number of possible parallel shipments. Then came the "Companyons." Artificial avatars of company values given synthetic skin and designed to fill the gap of human connection that stood in the way of TransSol doubling their market footprint. Another industry pillar given life when someone with too much money and a poor taste for puns asked (but never tried to answer) "Why not?" Zac was one such product of the "Companyon" program.

During Isaac's onboarding, TransSol staff took a psychological profile and from that profile came Zac. Or rather, unexpectedly, from that profile came another Isaac. The TransSol Companyon lab staff were surprised but not alarmed by the mirror image that had been created of Isaac. It wasn't commonplace in the Companyon generation process, but it didn't suggest a problem either and image was where their similarities ended. Like all Companyons, Zac, was a paragon of TransSol's public facing company culture manifesto -- hard work, customer satisfaction, and a can-do spirit -- balanced with just enough social graces to make a "solo" mission through space bearable.

"...Isaac, are you coming?"

Isaac snapped back to the present, shaking off the memory fog.

"Ah. Yes. Apologies."

Isaac and Zac set off from the observation bay for the cargo hold, mostly in tandem, but Isaac intentionally keeping a slightly slower pace than Zac's natural stride. It was just one of those petty games Isaac learned to play. It reassured him of his status as the mission leader.

"Zac, does today feel significant to you at all?"

"Not particularly significant. Certainly, pleasant like most days. I visited the greenhouse this morning and I'm happy to see our adjustments to the flowerbeds producing the blooms we were hoping for. The colors are more vibrant, and the atmosphere and particulate monitors suggest the aroma should be pleasant but subdued."

Internally, Isaac was happy to hear the report. Floriculture was one of his few skills and their first bloom was something of a failure. Somehow the foxglove had bloomed surprisingly fragrant, borderline pungent, and it had made visiting the greenhouse an endeavor.

"Well, that's all well and good, but I think you're getting distracted." Or dodging my question, Isaac wondered. "Tomorrow's our last day. Really today feels like our last day. Our last full day of monotony and routine and all the charms this tin can courier ship has to offer."

"You didn't ask about tomorrow. You asked about today. And today's much the same as yesterday. I do agree tomorrow will be a change of pace."

"Change of pace?"

Isaac stumbled over his own lost words and into another memory; his and Zac's first meeting. In fact, what might have been called Zac's birthday, when Isaac gave Zac his name.

***

Isaac stared at his synthetic skinned mirror image, lost for words. The new Companyon model stared back, feeling no pressure to find any words.

"Uh... hello...", Isaac said, putting his hand to hand to his chest, "Isaac. I'm Isaac."

"Isaac..." the doppelganger responded.

Isaac was uncomfortable in presence of his new doppelganger and tried to defuse his internal tension with what could hardly be called a joke.

"Oh... heh uh... You 'Saac too?"

Isaac internally bit his lips closed with his teeth to keep the rest of his "very funny jokes" contained.

"Sure. I can be Zac."

***

Back in the present, Isaac's face had mimicked the same tight-lipped cringe, and he force-flung his daydreaming further forward to their first day on the job to ease out of the embarrassment.

***

Isaac and Zac took their first steps into the microgravity cargo hold. Isaac had expected a warehouse layout of boxes and shipping containers like the delivery hubs back home: corridors of shelves stacked high with labeled packages and a small fleet of forklift drones to fetch and rearrange inventory as deliveries went out. Isaac's expectations were quickly forgotten. The cargo hold stretched 40 meters to the left and 40 meters to the right. 5 meters from the entrance sat a terminal and a chair, and a meter beyond that rose an impenetrable wall of shipping containers.

The arrangement of cargo ran floor to ceiling and the entire 80-meter width of the cargo hold. Surrounding the monolith on all sides -- left, right, top, and bottom -- was just a few inches of clearance that the cargo sat suspended within. The individual cargo composing the monolith were nonuniform but somehow arranged together like a 3D jigsaw puzzle. Pressing his face to the floor, Isaac peered beneath the monolith to test if he could see how far back it ran, but the cargo hold's limited light only reached a few feet back. Isaac stepped back, at first gormless. Turning to Zac, his stupor morphed into indignation.

"Oh my god, Zac... What am I supposed to do with this?"

"I don't think I follow"

"This is why I'm here right? The ship more or less pilots itself. I thought my whole purpose here was to babysit this mess, but how can I babysit THIS? It's packed so tight that I can really only inspect the boxes closest to the surface. And it's so deep that even if I started inspecting my way through to the core, I'd have no idea how deep to go. It's an inscrutable black box."

"Are you considering disassembly as part of your inspection processes? I would not advise that." Zac softened his tone and Isaac was all too aware. "Why are you so interested in manually inspecting the cargo? Even if you were able to, there's far too much to sift through and we have more effective tools..." Zac gestured to the terminal.

"I know, I know." Isaac sat at the terminal and brought the IMS -- inventory management system -- to life. With no great effort, Isaac flexed his training and drove the terminal through its management routines. Glyphs, graphs, and summaries flooded the screen. Bookending the flood, the system reported 62,000 packages, 0 container defects, 0 environmental flags, 0 required actions, and 0 suggested actions. "I'm here after all, right? I did the training. I can read the reports." Isaac pinched the bridge of his nose. "I just... it's not a lack of trust. But it feels like a limitation... a bottleneck. I'm trying to describe a world that I can only view through a pinhole in a steel plate. I just wish I could poke a few more holes to see it from a different angle."

***

Isaac released the pinch on his nose, again letting go of the memory, as they rounded the corner and passed by the rec-room. He and Zac had gone back and forth a lot that day, making no ground, and reaching no resolution. In the end, Isaac resigned himself to the absurdity of it all. And in a way, time had proven his concerns to hold no weight. Days had come and gone, and the mission had continued smoothly. Most days the reports were the same as that first day. "62,000 packages, 0 container defects, 0 environmental flags, 0 required actions, and 0 suggested actions." On the rare occasion there was a required or suggested action, Isaac was able to quickly mitigate the issue from his terminal: adjusting the microgravity to soften cargo stress points, resolving conflicts in the report data, rebooting the terminal itself. And in the quiet days between, he and Zac had busied themselves in the rec room with games of cards.

...

And that's where my draft ended. ◡̈

I know how I wanted it to end and the "feel" that I wanted that ending to have (hint: there's a problem with the cargo and the delivery never mattered), but I just... couldn't get it there.

Still, as disappointed as I am to run out of steam, I'm also proud that I took it this far.

I look forward to my next attempt at creative writing!

Original prompt

Here's the original request that I received in the mail:

Hello my carbon-based friend.

I would like you to write a piece of Scifi flash -- that is, 500-2000 words, or 1-4 pages. This is the premise:

Far in the future, human pilots operate long-haul spaceships transporting cargo for giga-corporations. To increase company loyalty and alleviate loneliness, the pilots are given AI companions which are sort of like avatars of the company. Basically like a very sophisticated body pillow, but instead of an anime waifu, it's Evergreen Space Marine-chan. These company avatars are called "Companyons." "Companyon" should be the title fo the story, unless you come up with a better one.

How you turn this premise into a story with a beginning, middle, and end, I leave up to you. Here are a few possible angles, but feel free to ignore these:

• A pilot is transporting cargo for multiple companies and tries to keep the two companyons from finding out that he is, as it were, cheating on them.
• A pilot fights like hell to protect her beloved companyon of many decades when she discovers that the company for which it is an avatar has been bought out, and all avatars will be replaced with avatars of the new parent company.
• After centuries of service, a retried pilot is struggling to return to normal life + reintegrate with society. To ease the transition, the pilot has brought their companyon to live with them, but as per company policy, the company must be returned to the corporation within a year of retirement.

Wow those sound fun! I'm almost tempted to write this story myself and find another story to send you... but I have a prompt of my own to answer. Like I said, feel free to take the story in a totally different direction if you want. 500-2000 words please ◡̈

Best Wishes!
-Francis

I want to do my part in spreading the gospel of a well-funded library system, so, periodically, I want to share my "hauls" from the library.

Welcome to the inaugural entry of my "Library Evange-log"

1. *HIM* - Monkeypaw Productions

=> IMBD: *HIM*

*HIM* was a cool horror/thriller flick. I thought it was a Jordan Peele directed movie, but it was only *produced* by Peele and Monkeypaw Productions. Nevertheless, it had some really cool set design (imagine Ex Machina's isolated compound but designed for a football celebrity) and some cool VFX. One thing that got an audible "yooooooooo" out of me was this special effect where they'd show the characters in an x-ray-esque view. Really cool visual connection to the theme of football concussions and later scenes get really trippy with it.

One of my favorite parts about this movie is how much of it is tense, unsettling, and takes place in the middle of the day. It felt like there were very few scenes that relied on jump scares, dark corners, or visual uncertainty.

It's also good about being scary without being gross. As a horror fan who is also pretty squeamish, it's a bummer when a movie looks great but I know there's something my cowardly stomach just can't handle. If you're like me, rejoice! Here's one more movie to add to your queue.

2. *I Feel the Everblack Festering Within Me* - Lorna Shore

=> Wikipedia: *I Feel the Everblack Festering Within Me*

A recent and pretty popular release from symphonic deathcore juggernauts, Lorna Shore.

I *really* want to like Lorna Shore, but this is another record that doesn't do it for me. And really, I don't think it's Lorna Shore. I think it's more of a symphonic deathcore thing: I feel like the genre encourages lots of dense sonic layers, and I think it tends to muddy up clear identifiable rhythms. And I definitely think I'm a "rhythms" guy. To be clear, I'm no stranger to heavy music and that's not the problem.

Oh well, what I love about the library isn't that I *love* everything I borrow, it's that it was fun to stumble upon it and give the album a shot on a whim.

3. *H.P. Lovecraft's At the Mountains of Madness* - Tanabe, Gou

=> SPL page

The art is *really* cool, but also *really* hard for me to parse. It's very fine line work, lots of black and white, and the scenes are pretty detailed.

I don't think I'm quite as big an HPL fan as I'd like to be. Work inspired by HPL is often more appealing than the HPL source material is to me. But, putting it in a graphic novel definitely made it more approachable for me.

Not a huge win for me, personally, but still a fun random find.

But one new-to-me feature that I had a lot of fun with in New Horizons was the photo mode. I stumbled across some of the pics I snapped all the way back in 2020 and wanted to share.

Enjoy!

Hanging out in front of my new home.

Showing off my vinyls.

Gone fishin'!

Christening the new town flag. Say hello to "Peperoncino."

Peek-a-boo.

Introducing myself to the new neighbors.

Nothing to see here.

Just out for a nighttime walk.

It's a terrible night for rain.

Bonfire fun.

Finally. quiet.

Ah, what a lovely summer.

the recipe

It's been a few years since I marked the v0.2.0 milestone.

I've finally gotten around to fixing some of the project's basic security issues, so it felt right to mark this point as the v0.3.0 release!

HTTPS and proper hostnames

As noted all the way back in the v0.1.0 release, I've connected to my VTPP feed and server over HTTP and using a static, private IP address.

After more than two years, this has still yet to cause a real problem, but it's a pretty significant security gap: Using a static, private IP address from a mobile device exposed me to the possibility of using a public WiFi network with a similar private IP range and accidentally connecting to someone or something I did not intend to.

And using HTTP meant that if I did connect to something, my phone wouldn't make any attempt to authenticate that it was the VTPP feed my phone was expecting.

Both issues have now been addressed.

Over the christmas break, I coincidentally set up a Pi-Hole DNS server on my home network. Fixing the "connect-via-IP" problem was as simple as adding a new vtpp.lemonpie.home.arpa entry to the DNS server's records that mapped back to the VTPP server's static IP address. For a long time, I thought fixing the hostname issue would require registering (and paying for) a real public hostname which isn't really what I wanted or needed. Discovering the home.arpa domain made the solution so easy!

And I chose to fix the HTTPS problem by setting up my own private certificate authority (CA), installing it as a trusted root on the devices which use VTPP, and then issuing a server leaf certificate for my VTPP server to use. The server certificate encodes both the hostname, vtpp.lemonpie.home.arpa, and the static IP as SANs (subject alternative names).

For now this is a workable situation for me. I would prefer to get my leaf cert issued from a more widely available trust anchor like LetsEncrypt, but I haven't yet figured out how to get one issued for a service only reachable on my private network. I think specialized tools like localcert.net exist to help bridge this gap, but I have not yet had the time to explore what the trade-offs are or how to use it.

I'd bet dollars to donuts that I'll get tired of managing my private CA after the first time I have to deal with a certificate expiration.

Oh well, for now I'm satisfied.

A quick reflection

It's funny. Now that the work is done, it feels like it was easy enough that I really shouldn't have put off this bare-minimum security work for over two years.

But I think this is a feeling that only makes sense with those two years of experience under my belt. Despite being an engineer on a team that primarily develops, troubleshoots, and maintains networking software/tooling, DNS and PKI have always been two significant gaps in my skillset. I've fixed bugs rooted to problems in both areas in the past, but there's a real difference between having an abstract understanding of those systems and getting your hands up close and dirty with them. So, I suppose it's more of a happy accident that my job has put me face-to-face with these systems more directly and more consistently over the last two years.

I feel both embarassed and re-assured that I have no reason to be. Such is life.

¯\_(ツ)_/¯

The full changelog

Here's a more complete summary of what's changed since v0.2.0:

• The VTPP homepage now reports the last time the fetcher ran and, uniquely, the last time the fetcher ran with errors.
• Discord reports are only generated when something changes or the fetcher encounters an error.
• I added support for enabling/disabling the fetcher schedule from the VTPP homepage.
  • YouTube breaks youtube-dl often enough that when something is broken for more than a day, it's more practical to just disable the fetcher until I can try to fix things over the next weekend.
• I improved my "YouTube Shorts" recognition heuristic.
  • Instead of looking for a maximum time, I just look for a "shorts"-path in the video URL. This seems to be both simpler and doesn't require successfully downloading any part of the video.
• VTPP now serves its feed and media from an HTTPS endpoint.
  • HTTPS support is provided by putting an Nginx reverse proxy in front of the VTPP feed.
• The VTPP feed is now reachable from a DNS name resolvable on my local network.
• The VTPP feed is now behind an Nginx reverse proxy and exposed on the proper 443 HTTPS port.

From your gemini browser of choice^[1] you can find it at: gemini://gemini.scottnm.com

If you're not familiar with geminispace, you can find a primer here: https://geminiprotocol.net

It's cool. I'm excited to explore geminispace more.

I do find Daniel Stenberg's thoughts on the protocol pretty interesting. It's a good splash of cold water on some of the ecosystem and protocol's issues. Nevertheless, I'm still excited and hopefully for geminispace's future.

I haven't yet been able to migrate all of this site's posts to geminispace since they're built from Markdown, and Gemtext (Gemini's analogous markup language) is meaningfully (and intentionally) less powerful. My attempts at using naive Markdown-to-Gemtext converters have so far not been very successful and I don't see it being a good use of my time in the near future.

Like a lot of people, I've run into trouble dealing with missing first-class support for inline links and footnotes.

Oh well. I'll deal and adapt with time.

For now, it's cool to have a space over there even if it's just my homepage linking back to the wider net.

Footnotes

I'm not so keyed in that I really have an understanding or attachment to the core "meaning" of any given artwork, but I've found myself more drawn to interesting colors, textures, and artifacts of greater-than-average size. It's a very uncomplicated "child giggles at jingling keys" level of art appreciation. I have fun with it.

So, I subscribed to a rotating feed of pretty famous public domain artwork early in the year, and now near the end of the year I'd like to share some of my favorites.

I think having the space to write a tiny bit about what "caught" me for each of these pieces will be fun. I'm well aware none of this commentary is insightful or instructive. That is not its purpose. I just hope anyone reading this has their experience with these pieces enriched by getting to share them with someone (me!)

Work: Effet de brume, Honfleur

Artist: Félix Vallotton

I love the surreal, warping feeling of the composition and how much everything smudges and blends together. There's a cool sense of motion to the grass and trees and everything feels off kilter. The greens and yellows are vibrant and somehow amplify the dismal gray mist. It feels like one of those transient spaces that stories use to transport characters between "reality" and the "other world."

Work: Sonata No. 6 (Sonata of the Stars). Allegro

Artist: Mikalojus Konstantinas Čiurlionis

What an eye-catch! It's all yellows and browns but it feels so vibrant. Maybe it's the golden quality to the yellow tones, but something reminds me of retro-futuristic, art-deco things: think Bioshock or the covers of some of Asimov's books. It also reminds me of some of Vasily Kafanov's work

Work: Boomstam, omgeven door bloemen en door vlinders en andere dieren

Artist: Rachel Ruysch

I see it so clearly. In an alternate universe, this is the cover of a deathcore concept album. It'd be titled We Come From Rot and it would be bleak as hell until it concludes with this shoegazey drone track that just puts you in a trance. Sometimes my favorite thing about artwork is where it sends my imagination! For some reason, this painting makes me think of Saturn Devouring His Son.

Work: House at Auvers

Artist: Vincent Van Gogh

I feel like I see this house in so many Seattle neighborhoods. Just make the trees taller and wham! I'm right at home. It's so cool to see the contrast with the very coarse paint strokes and how they come together to make cohesive forms. This is one of those paintings I should not see in person because I will want to put my grubby little mitts all over it and commit a culture crime. It looks like it would have a very neat texture feel.

Work: Flower still life

Artist: Adolphe Monticelli

This would be the artwork accompanying the lyric sheet insert of the aforementioned deathcore album. It's all the floral bleakness of Boomstam with all the abstract-paint-blobs-miraculously-look-like-something of House at Auvers. Beautiful.

Work: Creation of the World

Artist: Mikalojus Konstantinas Čiurlionis

I'm going to ignore the obvious that this would make for another killer album cover. Instead... wow! that is such a rich palette of blues and a really effective use of a monochromatic colors to give a sense of light and depth. Like swimming through a deep underwater tunnel. The texture of the brush strokes makes me think of underwater eroded rock surfaces: unexpectedly smooth but still gritty. And I'll note this is the second Čiurlionis work on this list. And they're so different!

Work: Popocatepetl, Spirited Morning-- Mexico

Artist: Marsden Hartley

Something about the shadows between each layer and the relatively similar sizes of the layers makes me think of set pieces in some kind of puppet theatre. Like Punch and Judy or shadow puppets or something. The colors are great too! A not-quite-monochromatic blue palette that makes the clouds look really soft. Check out the clouds in the bottom right corner. For some reason, this painting above others really brings me back to the style of Braid.

Work: Jungle With Setting Sun

Artist: Henri Rousseau

I'm not huge on this one but I think the contrast works exceptionally well. Everything is in this soft green color, except the sun, center frame drawing all the attention. There's something about the sun being that size and that low to the horizon that makes this piece feel eerie to me, too. It's too... small? It certainly seems off to me. And I like that! On a side note, I really struggled to find a good source archive of this work online. A lot of the works I've shared here were easy enough to find on Wikimedia Commons, but after much googling I couldn't find any sites that weren't just selling prints of this work. Nothing was archived on the Internet Archive either. I found this surprising.

Work: The Bonaventure Pine

Artist: Paul Signac

Oh me, oh my, it's pointillism baby! And what-a-pointillism if I do so say so myself. I love the way the light works in this piece. The far background is full of soft, radiating light rays but the tree is still mostly in shadow. To me, it reads as a sunrise, but the tree is blocking the sun and becoming the focal point. There's apparently a lot of impressionist art about sunrises. I bet there's some significance in this impressionist piece plopping a tree right in the middle of that sunrise and upstaging it. How bold! How brash!

That's all! I hope you enjoyed my curated little slice of pretty-famous-public-domain-art.

Footnotes

the recipe

I don't have much experience writing haiku or poetry more broadly, but I do have a long-time love affair with clever wordplay^[1].

So, I gave it my best shot! I spent a week jotting down ideas and reading through haiku: both popular works and previous contest submissions.

We were allowed to submit two haiku. Here's what I submitted:

Submission 1

Submission 2

Curious why these don't fit the 5-7-5 format? More below^[2]

It was a fulfilling and challenging exercise! The first submission in particular, I felt like I drafted so many different variations of each line: altering word choice, syllable count, meter, seasonal references, and poetic turns.

After drafting and re-drafting, I was basically down to two candidates: what I submitted above and this alternate that I left on the cutting room floor.

Submission 1: original

There was a lot that I liked in this original form:

1. It followed the (admittedly unnecessary) 5-7-5 pattern.
2. The word choice had a bit more grime and violence which better fit the original spirit of the poem.
3. I liked the double entendre of the final line.

Ultimately, I went with the variation I submitted because I felt its terseness better fit the tone and style of haiku. And while I loved using words like "flesh", "swarming", and "strike", the more I read both variations out loud to myself, the more I felt the meter of the original was a bit clumsy: words felt absent out of necessity to fit the 5-7-5 pattern and the rhythm of the words I left behind didn't really flow.

I think I made the right choice. I get the sense that, for haiku, there's value in celebrating shorter moments. The submitted poem narrowed the scope and left more to the imagination of its outcomes. The abandoned original described more of a narrative and got lost trying to fit in my pet favorite phrases.

I'm excited to see which submissions get selected!

[Jul 13 2025] Contest Results Update

I did not win ( ˶°ㅁ°) !!

Okay, maybe I'm not suprised, but I'd be lying if I said I wasn't daydreaming about how cool it'd be to win.

Nevertheless, I still had a lot of fun and I hope to submit an entry next year as well.

The winners and honorable mentions for the english category are available on the graceguts site.

My favorite winner

Congrats to all the winners!

Footnotes

You can see an archive of the old homepage on the Wayback Machine.

Why?

There were a lot of things which I liked about the old site:

• its visual uniqueness
• the monochromatic rust/red color scheme
• the front-and-center-ness of the images on the site.
• the page was static but all the content was populated with client-side vanilla JavaScript templating which made iterating and adding content very quick!
  • I did a bit of work that I'm proud of to make appropriate use of JavaScript without relying on a bunch of web dev tools that bog down day-to-day development (bundlers, transpilers, local HTTP servers, etc)
• it was easily archived with the Wayback Machine
  • I'm not sure why this was so important to me. Usually web archival tools are most appropriate for when you don't own the page in question. For some reason, it gives me comfort to know its a little more resilient to disappearing forever.

But it wasn't without its flaws:

• The large number of images impacted page load speed
  • Given its a static site, I've always felt like it has to be a goal that the page loads quickly
  • I tried several "easy" tricks to improve this (deferred image loads, reduced image size/quality, etc) which worked to some but not total success.
• The site relied on GoogleFonts for visual flair but these impacted page load speed
• I never quite achieved the page structure/layout I had in my head.
  • This is mostly an admission that I'm not a very good web developer and the CSS styling required to gridded structures still very much confuses me.
  • I also struggled with placing images within those grids at the right place and size.
• My page's structure/layout was designed for desktop and it didn't translate well to mobile
  • Images felt cramped.
  • Margins and padding always felt off compared to the desktop variant.
  • I tried to use responsive CSS media queries to change the styles based on the client's device, but it always felt like trying to force the desktop design to work on mobile with very poor compromise.
• I got it working but enabling the page to be "Wayback Machine" compatible required some compromises in my client-side JavaScript (everything was crammed in one mega-file)
• For simplicity, I never removed old page content (which lived in the JavaScript). I mostly just commented it out so it was easy to bring back without trawling through git history. But it bothered me that I was wasting bandwidth on dead data.
  • Using a minifier would have helped here but I always appreciated that my site was legible in the browser inspector. There were other ways I could have addressed this (some of which I did as part of this site redesign!) so I'm mostly stating that it was a pet peeve I kept punting to the future.

I tried to address some of these pain points individually without doing a full redesign, but I always struggled to fix "just one thing." For every "one thing" I fixed, it felt like "one step forward, two steps back", and, that feeling is hard to ignore when you're focused on the one thing. On the contrary, starting from basically nothing, everything felt like progress and comparisons between what I was gaining and what I was losing felt less obvious. Feels weird to admit having to trick myself to get work done (and those "tricks" probably costing me more time in the long-term), but eh I'm human, at least it got done, and the wasted time isn't the end of the world.

What's changed?

The most noticeable change is in the layout and theme.

• The red-rust monochromatic colors have been replaced with simpler (and better contrasting) black-and-white
  • Honestly, this one's a bit bitter-sweet. I like the new colors, but I think it has less personality and I'm still interested in finding something more eye-catching than black and white.
• The new layout adapts better between mobile and desktop. There's less distinct elements taking up the same horizontal space so all that really changes between mobile and desktop is the padding and margins around the core site content. Put another way, I restrained myself from overdoing it with grids.
• The new layout also improves performance by only using images for items in the "highlights" section (rather than for every project and text post).
• I'm now only relying on "web safe" fonts. I think this runs the risk of robbing my site of personality, but so far I like how things look. And while the fonts I'm using aren't 100% guaranteed to be available on all browsers forever in perpetuity, I think things feel abstractly safer than relying on fonts delivered remotely via CDN.

On the "authoring and development" side, things have also changed such that I'm no longer relying on client-side JavaScript to populate the page content at runtime. I do the more obvious and only slightly less dev friendly strategy of just using templates to build the HTML up-front and avoid JavaScript all together.

Results

In general, I'm pretty happy with how things turned out. The mobile experience is significantly better. The homepage loads faster. There's fewer external dependencies. It still archives well on the Wayback Machine.

There's still room for improvement in the future (notably the colors), but for now I'm calling it complete.

Shout-outs

I took meaningful inspiration from a few sites when re-designing my own. In no particular order:

• scattered-thoughts.net (Jamie Brandon)
• fabiensanglard.net (Fabien Sanglard)
• jvns.ca (Julia Evans)

Thank you!

It's been about a month since I marked a v0.1.0 version of the project. Here's what's I've changed since then:

• Updated the web portal to enable editing the configured youtube channel list that I'm fetching videos from
• Updated the web portal so that I can delete videos from my feed as well as "block" videos from being re-downloaded in the future.
• Added a simple time heuristic to recognize (and ignore) Youtube Shorts when fetching videos
• Added logic to ignore offline Youtube Live streams which sometimes(?) show up in a channel's feed
• Made a handful of small improvements to make "on demand" download progress clearer
• Refactored a bunch of things to make code more reusable across the different components

There are still a lot of things I could do and would like to do, but the project is at a point where things feel polished enough that I can "just" use it without feeling like anything is missing.

So, I'm going to mark this point as a new version, v0.2.0, and let things be for a bit. It'll give me some clarity on what work from the backlog is actually worth spending my weekends on.

For now, the name I've been using is "VTPP" (Video-to-Private-Podcast), and the concept is simple: it's a home server which periodically downloads videos from external source(s)^[1], strips the audio from those videos, and serves that audio in a podcast feed accessible from my home network^[2].

The server also supports downloading specific videos without subscribing to an entire video feed.

So far, all this has worked really well!

• The server itself does very little work, so I've been able to run it off a spare Raspberry Pi that was collecting dust.
• There are more than a few YouTube channels where I only want to listen to the videos (usually during repeat watches/listens). It's nice to save power and turn off my phone screen while listening, which YouTube's current UI/UX/business policies prohibit without paying them.
• The size of the downloaded audio is more than manageable for my server's storage. Enough that I haven't bothered to worry about figuring out how to manage running out space in a smart^[3] way.
• I haven't been able to get most iOS podcast apps to connect to my home server, but I did find one that is really great so far: RSS Radio.

I don't yet intend to package this for anyone else to use or even to make the source available, but the concept is pretty simple to implement. If you like the idea, feel free to implement your own version! I'm not the idea police. No way I was the first one to think of this idea (I never bothered checking...).

Here are some screenshots of it working in action!

Here's a quick demo of my phone playing an episode from the podcast feed.

Here's a quick demo of me requesting a video be downloaded on-demand to my feed server from my phone, followed by the audio becoming available within the podcast feed on my phone.

To make it easier to monitor the health of the server, I added Discord webhook support. If the server runs into trouble, it'll send error reports with logs to a private discord server I setup. And when new videos are successfully downloaded, I get a quick summary of what's been downloaded.

[1] currently, the only external source I've bothered to support is YouTube. The server is configured to look at the feeds of some explicitly listed YouTube channels and pull their latest videos.
[2] accessible from my home network and only from my home network i.e., by design, it's not exposed to the public internet. If I'm out and about, I'm not taking updates.
[3] currently, when I run out of space, the server stops attempting to fetch new videos and sends me a notification of the problem.

Problems with podcast apps

As mentioned earlier, I haven't yet been able to get my home feed working with my longtime-preferred podcast app, "Overcast", or the "Apple Podcasts" app. These apps allow you to subscribe to user-provided feed URLs, but when I try to use the address of my home server it fails with a non-specific error message saying it couldn't use the URL.

I still haven't quite figured out what the problem is, but I did spend a little time trying to investigate. Or rather, I spent a few days trying to capture the app's HTTPS traffic with Fiddler, and, when I couldn't get that to work, I ran out of ideas on how to gather data necessary to investigate where/why fetching my feed was failing. Some suspicions I have include...

1. I believe these apps probably do cert pinning. At the very least, I'm pretty sure this is why I wasn't able to Fiddler the HTTPS traffic, and it doesn't seem totally out of the realm of possibility that these apps require HTTPS authenticated feeds (which I haven't setup yet) and possibly pin certs for a handful of well-known feed services.
2. It's possible that these apps do some validation/sanitization of the user-provided URLs intended to guard against having the apps fetch data from malicious sources.
3. It's possible that these apps don't handle URLs well using the hostname format I'm specifying. I've tried using addresses of the form http://{hostname}:{port}/ and http://{hostname}.local:{port}/ and neither seemed to make a difference.
   • Though it just occurs to me that I was always using and specifying a development port rather than the standard port 80 (HTTP) or 443 (HTTPS). Maybe I'll try exposing the server on port 80, omitting the port in the URL, and seeing if that changes anything.

A Thought on Security...

I am not a programmer with a lot of security expertise, but I'd be remiss to not reflect on a possible (probable?) security gap that I haven't yet addressed.

That is...

1. the feed is currently served via HTTP and not HTTPS.
2. the feed's address is my feed's hostname on the local network and not some reserved public internet name.

There's nothing preventing my phone from trying to reach this hostname on a network other than my home network, unexpectedly resolving the wrong host, and trying to talk to something/someone it shouldn't. Of course, the likelihood of this happening is minimal, and, in theory, the attack surface is isolated to my podcast app (which is why I haven't frontloaded addressing it), but I'd still feel more comfortable knowing the security gap wasn't there at all.

I think the right way to fix it would probably be to expose the feed from an HTTPS endpoint with a cert that chains back to some trusted root. Or maybe I could get by using a self-signed cert that I keep secure and installing that self-signed cert on my phone. Maybe there's an even further future where I could somehow setup some cert pinning such that only my self-signed cert would be trusted when trying to reach my feed (though I think that implies writing my own bespoke podcast app which is... uh... a little more work than I'm excited about).

Guess that's a problem for future-me.

Future Improvements

Speaking of the future, while I'm comfortable with where this project is right now and may let it rest for a while there are a handful of features on my wish list that I'd like to address eventually:

1. Make more of the feed's runtime behavior configurable from the web portal (i.e., managing currently subscribed channels and shared metadata)
2. Enable better management of the feeds from the web portal (i.e., deleting items from the feed that I no longer want/need)
3. Make setting up the server more "plug-n-play" with docker containers
4. Flagging some content to not be downloaded (e.g., YouTube shorts)
5. Setting up an automatic eviction policy for older downloads

Here's to hoping I get around to this stuff!

Edit 12/13/2023: I got around to some of this stuff! Not all of it, but enough to mark a new v0.2.0. Check it out.

As of Oct. 29 2022 (i.e. in just 6 years), Monkey Muck...

• has 2,225 views
• has 204 downloads of its Win32 build
• is a part of 14 itch.io collections

In today's internet landscape, these numbers are incredibly small, but "incredibly small" is still about 100x the attention I expected this demo to receive. It makes me wonder what compounding coincidences led to any traffic at all. I never shared this demo with anyone (aside from linking it on my personal site which I also don't share with anyone), so I have to imagine it's mostly coincidence. Maybe having "horror" in the tags list gave it some attention for people browsing itch.io for horror games.

I think my favorite of the above data points is that the demo is in 14 itch.io collections. As far as I can tell, there's no (easy) way to set up "smart" collections which fill themselves, so this tells me that at least a handful of people thought my demo was worthy of adding to their own personal curated lists. Cool!

This past May, without knowing about these numbers, I went through the exercise of revisiting this demo and adding a WebGL build. My motivation for doing this was purely nostalgia; I wanted to revisit something I'd made years ago and see if it was still tweakable. It was, and I was able to publish the WebGL version on my itch page so that folks could play it without having to download and run a Win32 binary. In 5 months, it's only gotten 4 browser plays, so I'm glad my motivation wasn't anything more. Maybe 6 years from now, I'll come back to Muck's analytics dashboard and see 200 browser plays! Wouldn't that be funny.

My team built an isometric digging game in Pico-8 called Dig Deep.

I had an absurd amount of fun and felt like I grew a lot as a hobbyist game dev from the experience.

I really wanted to write some sort of cool, formal "dev retrospective" thing where I gave a well-structured breakdown of everything I learned and experienced during the jam. Ya know, really sit down and reflect on everything with an intention of taking all those lessons and using them to build something even cooler next time!

...But, I've been slow to write it so my memory is getting hazy and my motivation has plummeted.

Shame shame shame on me.

In an attempt to half-assedly salvage this, I'm instead going to jot down a simple, contextless bulleted list of takeaways that are still semi-fresh in my memory. Without further ado, here we goooooooo...

• I'm very proud of how "complete" our submission felt. The game was end-to-end playable, we had music, we had no placeholder art, and we even had a proper ending with story text. Something that feels this close to "complete" was a new jam experience for me.
• Knowing how long to spend on initial design is tough. We spent our first two (of fourteen) days coming up with an initial idea. I really wanted to start iterating as quickly as possible and find out what was fun by playtesting rather than just thinking about it and even two days felt like we might be spending too much time "thinking about it". On the other hand, it takes just as long to build a bad game as a good game so its not like the design phase isn't time well spent. It's certainly a tricky balance, especially for a game jam.
• I think designing was easier once we established what our limitations and constraints were. Some constraints were self-imposed, but many of them were imposed by working in Pico-8. Once we had a much narrower design space, ideas flowed more naturally. Our constraints were:
  1. Complex physics and collision are a no-go. Pico-8 basically makes you implement everything but the sprite and audio renderer yourself.
  2. Pico-8 forces you to make very real trade-offs between in-game text and code you can write. In-game text is a very scarce resource. Use it carefully, use it wisely.
  3. The game had to be playable with only 6 buttons. I am still shocked at how many interesting game ideas this constraint banished to scrap pile.
  4. We have limited time to test our game and basically no ways to automate testing so games need to be easily testable. The best way to keep things testable was to keep design simple. Reduce "if-this-then-that" ideas where possible.
• Doing this Game Jam with a team was a such an amazing experience. I may never go back to solo jamming.
• I like to fool myself into thinking I can pickup slack from other areas of the project (art, sound effects, music) in addition to my programming duties. I cannot and I've gotta stop pretending. Leaning on others made all the difference in this jam.
• When working on a team, defining shared processes and strategies early on is very important. Our processes felt like they developed over time and so ended up feeling more isolated and confusing. I think we could have improved this coordination.
• Pico-8 feels ideal for solo work. Multiple people working on the same cart felt clunkier for all non-code assets.
• My rough, back-of-hand estimates suggest I worked 60-70 hours on this jam. Roughly 3-4 hours a night on weekdays and at least 8 hours a day on weekends. I think that number is a little high for my taste. It's not clear where/how I would reclaim time (smaller designs, better tools, demanding less of myself, etc)
• Pico-8's export tools are awesome and way more powerful than I first understood. Generating an HTML5 web version and standalone PC executables was trivial.
• Pico-8's screenshot and gif capture tools were super useful for our remote Game Jam. I shared so many gif screen captures to get feedback on in-progress features.
• Pico-8's cart size limitations don't hit you until late in the game and when they do they hit you hard. This is specifically troublesome for a game jam where the late-game is where you really start scrambling. It sucks to know what you need to do but to just be straight up out of cart space.
• I refuse to write unreadably terse, uncommented code and Pico-8 is more than happy to punish me for this.
• I wasn't comfortable using Pico-8's map editor for laying out levels and I think that showed and cost us in the long run.
• I'm really happy with the isometric look we achieved in the game.
• I'm proud of the janky way I implemented movement, collision, and hitboxes for the isometric view.
• "Lessons Learned Making Gunpoint Quickly Without Going Mad" is an awesome GDC talk that I wish I had watched before the jam. The "easy-to-implement vs fun-to-play" matrix for determining whether to implement a feature is a great idea for game jams.

...and I think that's about all I've got.

One cool consequence of working in Pico-8 was that I had to make frequent use of its gif capture tool to share progress with my team as I was working on new features. Almost by accident, I ended up capturing effectively the entire development history of the project in gifs. Here they are! I think it looks pretty cool.

If you're not familiar with Achewood, this post isn't the right venue for me to explain it to you. The only context you need to understand the rest of this post is that...

1. I like Achewood quite a bit.
2. The Achewood website is really the only place you can read the comic without spending hundreds of dollars for copies of older print collections.
3. Reading Achewood from a mobile device is not a pleasant experience (or at least it wasn't for me the last time I tried back in 2021^[1]).

I wanted a nicer way to read Achewood, and since it hasn't been formally published for e-books yet, I set out to fill that need for myself.

I wrote a web scraper in Rust^[2] to crawl the Achewood site and download all of the comics from all of the major Achewood story arcs. The comics were all downloaded in their original format: tiny GIFs (this webcomic started in the early 2000s. It was a different time!)

My first attempt to make the comics e-reader compatible was just to generate a PDF where each page was a different comic. There were probably smarter ways to do this, but how I ended up doing this was by selecting all of the files on my PC and running them collectively through Window's "Print as PDF" functionality. It worked... ok..., but a bunch of manual editing and oversight was needed^[3] and since the original images were small, there were pretty bad rendering artifacts when the e-reader would try to expand the image to fill the screen.

So, next, I ran all of source images through an AI upscaler to deal with those rendering artifacts. Honestly, I don't remember much about this part of the process (I'm writing this in 2024 when I originally did most of this work back in 2021). I have been able to dig up that I used BigJPG as my upscaler and I think I have some memory of writing scripts to automate some parts of the process for me, but I can't seem to find any of those old scripts.

With all of the comics HD-ified, I constructed another PDF with the same ugly manual process, and that worked ok for quite a while. PDFs are always a bit awkward to read on an e-reader since the e-reader can't make smart decisions about how to layout the content. So, you end up with awkwardly small text and an interface that has to support zooming which is usually pretty terrible on e-ink e-readers.

Near the end of 2022, I finally had enough and took the plunge to redo Scrapewood (yet again) but this time in EPUB format which my e-reader works best with. Figuring out how to format this as an EPUB was a really interesting process. EPUB is a well-enough documented format, but rather than read a spec I took the lazy/fun path and sort of reverse engineered the whole thing. Once I discovered^[4] that an EPUB is really just a zip archive of some XHTML files and other assets (e.g. image files), it was almost faster to just extract the contents from a working EPUB file and work my way backwards to creating my own. Sometimes, it's easier to iterate with something that's already working than to read a long spec full of information that isn't relevant to the task at hand. Ultimately, I ended up being able to generate the whole e-book from a collection of template XHTML files that I worked out by example and a <300 line PowerShell script. How cool! No fancy tools needed here, no sir!

And that's it. I've read through Achewood in full about 4 times and 2 of those times were using my own Scrapewood e-book. To me, that's a very satisfying experience and I'm happy I took the time to write about the process.

Source?

Unlike most of my projects, I'm not sharing much of what I actually built (the PDFs and EPUBs) or the source code of the tools I used to build it (e.g. the web scraper and EPUB generator script).

This is intentional.

While Achewood is free to read online, it's not my property to redistribute in other forms. If you're looking for a fun project, I've outlined how you could do this yourself (and hopefully helped you avoid my own missteps), but that's as far as I'll go.

Footnotes

[1] Jan 14th 2024: I just checked and the mobile version has significantly improved! Buttons are appropriately sized, you can scroll to view more comics, the viewport is fit for mobile... it's beautiful! Almost makes me feel like some of this effort was wasted but whatever, I got multiple years of use out of it and it's still mine!
[2] Why Rust? No reason other than I wanted an excuse to write some Rust.
[3] I do not understand why, but, sometimes, Windows would print some comics rotated and other times not and it took a look of fiddling around in the PDF options to coax Windows to generate the PDF semi-correctly.
[4] By "discovered", I mean that I tried to open an EPUB file in Vim hoping that it was somehow just a plaintext file and was surprised to see Vim trying to read it like a ZIP archive (which Vim has native functionality for browsing).