Billy's Blog

Stay In Your Car

Sun, 21 Dec 2025 18:00:00 +0000

I went on a walk today to return a library book. I knew they would be closed, so I looked for the book return drop-off. The only option was to walk into the drive-through. A convenience just for drivers, so they could stay in their cars and not have to step onto the sidewalk.

I grabbed a coffee and took the long way home. I breathed in the polluted air and observed the lifeless streets full of rapid traffic. I passed by the Starbucks with a line of cars stretched around the corner idling. I passed the McDonald’s where I watch dozens of drivers pick up their breakfast every cold weekday morning while I stand at the bus stop. They can’t be bothered to speak to the workers face to face. They order through a speaker box.

“Stay in your car”

They were in their car before ever leaving their home. The double garage door opens and their machine emerges into the world so that they don’t have to.

“Stay in your car”

Finally, I passed the new oil change garage on the corner. Bigger than the brand name, bigger than descriptions of sales or offers, was a giant sign.

“Stay in your car”

What I've Been Reading Lately

Sun, 14 Sep 2025 15:00:00 +0000

Updated daily from my StoryGraph and Pinboard profiles.

On The Web

Last Updated: 2026-03-16 10:57:34

Books

Last Updated: 2026-03-16 10:57:15

The Hundred Years' War on Palestine by Rashid Khalidi
Paved Paradise: How Parking Explains the World by Henry Grabar
When Driving Is Not an Option: Steering Away from Car Dependency by Anna Zivarts, Dani Simons
The Color of Law: A Forgotten History of How Our Government Segregated America by Richard Rothstein
The Sirens' Call: How Attention Became the World's Most Endangered Resource by Chris Hayes
The AI Con by Emily M. Bender, Alex Hanna
How Infrastructure Works: Inside the Systems That Shape Our World by Deb Chachra
How Fascism Works: The Politics of Us and Them by Jason Stanley
Children of Dune by Frank Herbert
Walkable City: How Downtown Can Save America, One Step at a Time by Jeff Speck
The Jagged Orbit by John Brunner
Dune Messiah by Frank Herbert
Bowling Alone: The Collapse and Revival of American Community by Robert D. Putnam
Blood in the Machine: The Origins of the Rebellion Against Big Tech by Brian Merchant
A Philosophy of Walking by Frédéric Gros, Clifford Harper, John Howe

Sharing my daily Rocket League ranking.

Fri, 08 Aug 2025 15:00:00 +0000

I recently added a page to this website that shows my latest Rocket League Rank. When I first had the idea, I was worried it might be a significant project. In the end, it was less than 100 lines of Python and a simple GitHub workflow. Here’s how I did it.

	1v1 Duel	2v2 Doubles	3v3 Standard
Rank:	Diamond I	Champion II	Diamond III
Division:	Division I	Division II	Division IV
MMR:	818	1231	1065

Last Updated: 2026-03-16 10:57:08

Ranks pulled from rlstats.net.

I was initially inspired by this post explaining how to use cascading GitHub workflows to update a static GitHub Pages site. I’m not quite interesting enough to have a live feed of all of the movies/books/shows I’ve consumed, but I do play a lot of Rocket League in my free time. The first problem I had to solve was where to pull my rank data from, and then how I wanted the workflow to pull that data into my repo. I considered following the article’s lead and have the rank data live in a Google Sheet that would serve as an intermediary data source between a cron job that would pull the rank data and the workflow that would insert it into my website. Instead, I decided to just dump a simple JSON file into my websites _data directory which is overwritten on each run of the cron job with the current rank data. In the future, I may extend my solution to keep a historical log of my rank progression or other game statistics in a proper data store, but I decided that just displaying my current rank would be good enough for a proof of concept.

The source of truth for Rocket League rank data is controlled by Psyonix (or now, Epic Games). There is a semi-public API to pull rank data for any specific user which is how you can look yours up on sites like RL Tracker and RLStats.net, but you need specific permission from Psyonix to get an API key. After reading a few dejected conclusions to Reddit threads of other curious programmers, I knew my chances of have having direct API access were slim to none. This meant I would have to go through the existing rank tracking websites, either through an intermediate API or by scraping my profile off of their site. At this point I reached out to the owner of rlstats.net since it appeared to be just one guy providing the service. He confirmed my fears that they don’t give out API access any more, and part of his agreement to continue using it was to not extend access through his own API. He was gracious enough however to give tacit endorsement of me scraping his site by saying he wouldn’t try to stop me as long as I didn’t send any heavy traffic. All I wanted was at most a couple of GET calls per day, so I was in luck!

I had never written a web scraper before, but turns out it’s really easy. After a few minutes of reading the BeautifulSoup documentation I had a few lines of python that would pull the data I needed. The hardest part was just sifting through the HTML structure to find the elements I was interested in. I’ve never been much of a frontend coder and usually touching HTML or JavaScript or CSS at all is enough to ruin my day, but again, this was easy! Then I just dumped the ranks to a JSON file, wrote my own static web page that is populated with the contents of the JSON file, and began applying my learnings on GitHub workflows. The workflow runs each night, triggering my scraper script which pushes the new rank data which in turn triggers the re-build of my static Jekyll site that is then published through GitHub Pages. Easy!

Now that my mind has been opened to the possibility of sharing more than just extended versions of Bluesky rants on this site, I hope to fill it out with more stuff that I find interesting.

COBOL, Social Security, and the epoch myth.

Sun, 16 Feb 2025 15:00:00 +0000

Elon Musk is a liar, but that doesn’t mean you need to lie too.

First, a disclaimer. I am not a COBOL expert, nor do I have any experience with old IBM mainframe systems. In my 8 years as a Software Engineer, the oldest thing I worked on was a codebase from the 90s written in Ada. That is pretty old in software age, but the COBOL ecosystem is still comparatively ancient. If any actual COBOL experts read this, I would happily take your feedback.

Background

While scrolling bluesky I saw a post with this screenshot making fun of Elon Musk and his people for not knowing how to code.

At first glance this explanation seems plausible. I have used timestamps based on the Unix Epoch on almost a daily basis in my previous work, and it would make sense for a language created before the Unix Epoch to simply use an earlier epoch date. But then I stopped and thought about it for 5 seconds.

In the end, as sure as I am that Elon is not a good coder, I am also sure that the people spreading this explanation do not know what they’re talking about either. After seeing this same misinformation pop up repeatedly on several social media platforms receiving tens of thousands of likes and reposts, I was fed up, people were wrong on the internet so I had to step in. (Update: Since I wrote this, several publications have reported on this unverified conjecture citing a Wired article which relies entirely on anonymous social media users for its conclusions on COBOL date types.)

ISO 8601

First of all, the ISO 8601 standard does not use an epoch! ISO 8601 is a format for representing a date and/or time as a string of characters. In other words, it does NOT represent itself as a count of units of time after some chosen starting point (aka. an epoch). Its dates are just a textual representation of a day in the Gregorian calendar (e.g. today is ‘2025-02-16’).

In 2004 the ISO 8601 standard was updated to include a “reference date” of the day of the Metre Convention, May 20th 1875. This is certainly the source of the confusion here, but it also does not change the format of an ISO 8601 date whatsoever. If you read the actual standard (before it was updated again in 2019 to remove this date) it’s just a single sentence:

The Gregorian calendar has a reference point that assigns 20th May 1875 to the calendar day that the “Convention du Mètre” was signed in Paris.

Why was that sentence added and then removed? Maybe the authors felt this was necessary to shut down some remaining Julian calendar fans who were referring to that same day as 7th May 1875. I don’t know. It doesn’t matter. There’s still no epoch.

COBOL

But what about the representation in COBOL? Maybe it uses the day of the Metre Convention as a default? Except it doesn’t. Yes, I subjected myself to learning the basics of a programming language that was designed three and a half decades before I was born. But you see, someone was wrong on the internet.

The one thing OP gets correct is that COBOL does not have a built-in date or time type. It does however have some built-in functions that will help you use ISO 8601 dates, and will even convert them for you to an integer that counts from an epoch. But once again, that epoch is not in 1875. The epoch is configurable using COBOL’s INTDATE option, but the only two choices are January 1st, 1601, or October 15th, 1582. This means that in theory a 0 value should default in COBOL to 1601, 424 years ago, not 150. Except even that is not right. COBOL indexes from 1 and not from 0. So really a 0 value is invalid, and just returns a 0000-00-00 representation. But don’t take my word for it, here is some COBOL I wrote, and its output:

PROCEDURE DIVISION.
INITIALIZE WS-DATE-DATA. 
DISPLAY 'Default initialized ISO 8601 basic format:'
DISPLAY WS-DATE-DATA.
DISPLAY 'System defined current date:'
DISPLAY FUNCTION CURRENT-DATE.
INITIALIZE WS-INTEGER-DATE. 
DISPLAY 'Default initialized number:'
DISPLAY WS-INTEGER-DATE.
DISPLAY 'ISO date of the default initialized number:'
DISPLAY FUNCTION DATE-OF-INTEGER (WS-INTEGER-DATE).
DISPLAY 'ISO date on day one (1) of the epoch: '
DISPLAY FUNCTION DATE-OF-INTEGER (01).
DISPLAY 'Integer of current date (days since 1601 epoch):'
MOVE FUNCTION CURRENT-DATE (1:8) TO WS-TODAY.
COMPUTE WS-INTEGER-DATE = FUNCTION INTEGER-OF-DATE (WS-TODAY). 
DISPLAY WS-INTEGER-DATE.
STOP RUN. 

Output:

Default initialized ISO 8601 basic format:
0000000000000000
System defined current date:
2025021604532153+0000
Default initialized number:
00000000
ISO date of the default initialized number:
00000000
ISO date on day one (1) of the epoch: 
16010101
Integer of current date (days since 1601 epoch):
00154910

The point is, there is no way to get COBOL to give you a date in 1875 without first being supplied a representation of a date in 1875. It is not the default. It is not the epoch.

Conclusion

Okay fine, but what if the engineers at Social Security made 1875 the default in their system or database? I mean yeah, sure, they could choose to do that. Maybe the guy that created the original database schema was just a really big fan of the Treaty of Metre.

CREATE TABLE SSA_RECIPIENTS
    ( SSN_NUMBER INTEGER NOT NULL PRIMARY KEY,
      FIRST_NAME VARCHAR(40),
      LAST_NAME  VARCHAR(40), 
      # WOOOOO I LOVE INTERNATIONAL STANDARDS OF WEIGHT AND MEASUREMENT!!
      BIRTHDAY   DATE WITH DEFAULT '1875-05-20') 

If you’re using a DB2 database and you omit the value after “WITH DEFAULT”, the default would automatically be CURRENT-DATE. Not zero, not the epoch, not 1875.

But now we’re purely speculating. In the end there’s just no evidence backing any of the claims made in that original tweet. We don’t even know if Elon really saw records with birthdays in 1875, and yet we’re all making up a false reality where his words would make some sense. I can’t speak to what’s in the Social Security database, maybe there really are some people who are listed as 150 years old, but I doubt it. We don’t need to be spending our time making sense of fascist misinformation that’s designed to destroy our country’s foundational institutions. We should be outraged that an unelected oligarch has access to people’s personal data in the first place.

Even if it is fun to poke fun at Elon online, he is known to fall for all sorts of obviously false facts about fraud or otherwise. His minions that feed him information about the vast and complex federal government that he’s taken control of are mostly young, unexperienced, and untrustworthy. You shouldn’t take anything he says at face value, nor random people on social media. I hope you learned something new today, I know I did.

Notes:

If I were to poke holes in this, here’s where I’d start:

I did not personally test out the DB2 default values for DATE types, but the documentation I referenced seems clear enough to me.
I did not try out every previous version of COBOL compiler to make sure the demonstrated behavior was unchanged.

For reference, here is the Working Storage Section for the program I wrote above.

WORKING-STORAGE SECTION.
01 WS-INTEGER-DATE   PIC 9(8).
01 WS-TODAY          PIC 9(8).
01 WS-DATE-DATA.
   05  WS-DATE.
       10  WS-YEAR         PIC 9(04).
       10  WS-MONTH        PIC 9(02).
       10  WS-DAY          PIC 9(02).
   05  WS-TIME.
       10  WS-HOURS        PIC 9(02).
       10  WS-MINUTE       PIC 9(02).
       10  WS-SECOND       PIC 9(02).
       10  WS-MILLISECONDS PIC 9(02).

My first post.

Sun, 26 Jan 2025 18:33:46 +0000

I am in the process of removing myself from most corporate social media as I embrace what made the internet great before all of these walled gardens were created: The freedom to share information and media across a global public network via a humble website. I plan to write more on this topic and my other interests and share them here.

Check back soon.