Happy Halloween 2025!

 

The end of comically bad fiction

Sadly, the Bulwer-Lytton fiction contest has finally come to a close. Over 40 years of comically bad opening lines to imaginary novels still reside at the website, though.

https://www.bulwer-lytton.com/about

The original, penned by Lord Byron Bulwer-Lytton himself in 1830:

“It was a dark and stormy night; the rain fell in torrents—except at occasional intervals, when it was checked by a violent gust of wind which swept up the streets (for it is in London that our scene lies), rattling along the housetops, and fiercely agitating the scanty flame of the lamps that struggled against the darkness."

"Virtual particles" are not real

Oh, sometimes things just piss me off. Take “virtual particles” from physics. Please. Eliminate them. They are neither necessary nor helpful.

“Virtual” means “non-existent” or “fake”. They are not particles.

Here’s a Feynman diagram for two electrons repelling each other. Take a look at it. It contains no less than four virtual particles.

sourced here

1. The electron on the left, sensing the approach of the electron on the right, emits a virtual photon, the recoil of so doing forces it to change trajectory.

   1. How did it know?

   2. How did it get the trajectory exactly correct?

   3. Emitting a photon makes the electron on the left lose energy and change spin.

2. The highly energetic virtual photon splits into a virtual electron-positron pair.

   1. Why?

   2. The virtual electron should be repelled by the real electrons, and attracted to its counterpart. This mostly cancels out to an attraction to its counterpart.

   3. The virtual positron should be attracted to both the real electrons and its counterpart. This mostly cancels out to an attraction to its counterpart.

   4. The real electrons should be attracted to the virtual positron, and repelled by the virtual electron. This mostly cancels out.

3. The virtual electron-positron pair recombine, releasing a highly energetic virtual photon.

   1. The annihilation happens at the right place and time to conserve momentum and send the new virtual photon in exactly the right direction.

   2. The creation-annihilation takes time. How did the sending electron know this random reaction would take place, and adjust the trajectory of its virtual photon accordingly, before it happened by pure chance?

4. The virtual photon continues on just the exactly correct trajectory to impact the electron on the right, forcing it to alter its momentum.

   1. The electron on the right gains energy and changes spin.

5. In the end, the electrons are repelled from each other. Both change spin. The one on the left loses energy and slows down, while the one on the right gains energy and speeds up.

   1. Which one loses energy and which one gains is randomly determined. The reaction can go either way.

      1. Therefore, the reaction actually goes both ways, with a 50% probability of each, so it all balances out.

         1. Yes, physicists actually believe that.

So what really happens when two electrons approach each other? The field gradients alter their trajectories, without any direct interaction necessary. The Feynman diagram with all its virtual whatsits is a mathematical tool for describing statistical outcomes of waves and ripples in fields, not details of actual events.

Taken from an excellent summary of this topic.

Canal and Revolution

The history you don't know.

I have occasionally wondered what caused all the peasant revolutions in Central Europe in the 1830's and 1840's.  Turns out, the opening of the Erie Canal was the root cause.  It's amazing how things are connected.  Especially when new connections completely alter trade routes.

Where do you think all those Norwegian bachelor farmers in Lake Wobegon came from?

Objects and Closures

Yes, this is another short post about programming theory.  I'm a geek.  It's what I think about at 2AM.

What is the difference between an object and a continuation?

An object is data with associated code.
A closure is code with associated data.
A continuation is a closure with a goto at the end.

That's it.  There's no need to overcomplicate these things.

Three programming languages

Here are three lesser known programming languages I like, and you might enjoy exploring. They’re each the creations of one man, who are all very opinionated about how their languages should look and work.

Nim - My current working language. Syntax is similar to Python, but it’s a statically typed, compiled language. It manages memory for you (very, very efficiently), unless you tell it not to. The language embraces generics, both type and procedural polymorphism, multiple dispatch, and meta programming. Has a remarkably diverse collection of packages available.

Odin - Another statically typed, compiled language. Syntax is similar to Go, but without all the hang-ups, and even easier to read and write. (Check out the “or_else” family of error handling keywords.) You have to manage your own memory, but the language makes that fairly easy to do. Has lots of features for games and systems programming, and is ridiculously fast. Ginger Bill’s opinions are worth paying attention to.

Frink - A small language that was designed to be a general-purpose calculator for engineering and science. It attaches units to numbers, and can automatically convert between them. It has a remarkably complete and correct table of constant value and unit conversions, plus historical purchasing power of the US dollar and British pound. It will even correctly compute error bounds and significant digits without any real effort on your part. It runs on the JVM, so it’s quick and flexible. It even has graphics built in.  (Yes, it's named for the mad scientist on The Simpsons.)

How bank loans actually work

I don’t have a problem with a modest rate of interest for loans. The time rate of money is a real thing, and opportunity costs always exist. I have a problem with banks (and that includes credit card companies) charging exorbitant rates of interest for – nothing.

The money (most) banks lend doesn’t cost them anything except their normal operating expenses. They make the money up on the ledger books and credit your account with it. Funds ex nihilo. When you repay your loan, the funds return to nothing – except for the interest you pay, which is profit to the bank. From digital bits the loans were formed, to digital bits the loans are retired. Banks can lend infinite amounts of money, because the reserve ratio (how much they have on hand compared to how much they lend out) is currently 0%. To reiterate, the money the bank lends you doesn’t come from your neighbor’s savings accounts. It is made up out of thin air and keystrokes on a computer. The banks have, in effect, been allowed to print their own digital currency, in unlimited quantities.

“As announced on March 15, 2020, the Board reduced reserve requirement ratios to zero percent effective March 26, 2020. This action eliminated reserve requirements for all depository institutions.” 

When someone defaults on a loan, the banks actually loses nothing but the value of the interest not paid. The bank then writes off the default as a business loss, as if the money had been real. You see, on their books, the money loaned out (which came from nothing and nowhere, remember) becomes an asset, not a liability. So when you default, the value of the asset becomes a loss, the quarterly revenue is reduced on the books, and the bank pays a bit less in taxes.

Yes, this also means that when you pay back your loan, the bank’s assets decrease – except for the interest and fees you paid with real money, which is pure profit.

But what about the money in your savings account? Isn’t that also an asset? Of course it is. The reason it’s not readily available, despite the bank handing out loans from nothing at all, is that the bank uses 90% of your deposit money playing in the stock and bond markets. But that’s a different topic.

Data structure taxonomy

There are quite a few standard data structures in programming. I’m not even talking about complicated things like red-black trees. These are basic data structures.

Sequence / Array / Vector
Keys: Numeric. May be fixed or flexible in number.
Values: One per key. May be fixed or flexible in type.
Pick: Any.

Struct / Record
Keys: Named. Fixed in number.
Values: One per key. Type variable per key.
Pick: Any or all.

Tuple
Keys: Numeric. Fixed in number. May be named.
Values: One per key. Type variable per key.
Pick: Any or all.

Table / Dictionary
Keys: Variable in format. Flexible in number (usually).
Values: One per key (generally). May be fixed or flexible in type.
Pick: Any.

Set
Keys: Variable in format. Flexible in number (usually).
Values: None (usually).
Pick: Any.

Enumeration
Keys: Named, fixed in number (usually).
Values: None (usually).
Pick: At most one.

Union / Tagged Union
Keys: Named and/or Enum. Fixed in number.
Values: One per key. Fixed in type per key.
Pick: At most one.

Linked List
Keys: None (usually).
Values: Completely flexible in number. Type may be fixed or flexible.
Pick: Iterate through the elements one value at a time.