
# I built an MCP project - here's my thoughts

## What's MCP?
[The Model Context Protocol](https://www.anthropic.com/news/model-context-protocol) is Anthropic's way of allowing models to have agentic capabilities. You can create an action (only bounded by your creativity) and then have the model execute it.

This is a way to standardize the interactions between a model and a given capability, so that interoperability is fairly painless.

## Final result:

I made a simple test project that fetches CIA world factbook from [this excellent open source data set](https://github.com/factbook/factbook.json?tab=readme-ov-file).

The final result is published here:
[Github](https://github.com/Joshuatanderson/factbook-mcp)
[NPM](https://www.npmjs.com/package/factbook-mcp).

You can try it out on your laptop if you'd like, and have claude installed.

First, run this: ```code ~/Library/Application\ Support/Claude/claude_desktop_config.json```

Edit the file to add the following:

```json
{
    "mcpServers": {
      "factbook-mcp": {
        "command": "npx",
        "args": [
          "factbook-mcp",
          "--yes"
        ]
      }
    }
}
```

Close Claude, and re-open it. Now, you can ask it to do something like `fetch me some information about the Netherlands,` and it will be able to get the information from the CIA world factbook, instead of having to rely on its training data!

## Lessons learned
I won't try to explain everything I did in this project, but I will note the things that I wish I'd known at the start.

### Limiting scope
Use a language you know well.** I used Go at first - but trying to learn the MCP was a challenge, and working with an unfamiliar language made it very difficult to know exactly what to do. Next time, I'll limit myself to using either a new concept, or a new language, but not both.

### MCP basics
Using a language that already has support from Anthropic makes it much easier to get started. Typescript and Node are the two that are currently in their docs (as of Dec. 2024).

There are two main ways to use the MCP protocol. You can either make a server that receives requests and responses, and listens on a port the (SSE section of the SDK), or you can make something that runs locally, with the stdio sction of the library.

If you use the stdio section, as I ended up doing, you'll want to either use it locally, or publish it as an npm package.

## Miscellaneous tips
- Use the [quickstart](https://modelcontextprotocol.io/quickstart/server) as a guide. It's actually pretty solid.
- At first, I wanted to make an endpoint that would return a list of capabilities, or other syntax guides. The sort of instructions or documentation you'd want to give to a model to use your tool properly should be in the "resources" section of the MCP, not a tool. 
- Keep in mind that you can only return text and blobs as your content. Responses in MCP are very tightly typed, and it will fail without a type match.

## Final thoughts
It only took me about 4 hours to get this working. Most of this time was spent trying to figure out how to deploy it into a way that anyone could access it, and if I did it again, I could probably do the same project in an hour. 

Anthropic's MCP seems well thought out, and has great docs! Give it a go, and let me know what you build!


Building my first MCP project


# I made 2024 Bingo Cards

[The Cards.](https://www.canva.com/design/DAF3W51c8sk/1xS5xaLYsa5zDiSiXU8STQ/edit?utm_content=DAF3W51c8sk&utm_campaign=designshare&utm_medium=link2&utm_source=sharebutton)

Now is the time to make your cards, and hope we don't get a bingo blackout on 2024.

If you want some, grab some. I'm sharing my template here. Feel free to copy, steal, or fork this in any way you like.

May the new year be glorious, and may few squares be ticked.

## If you wish to fork this:

[Here are the cards.](https://www.canva.com/design/DAF3W51c8sk/1xS5xaLYsa5zDiSiXU8STQ/edit?utm_content=DAF3W51c8sk&utm_campaign=designshare&utm_medium=link2&utm_source=sharebutton)

I made this list using [canva](https://canva.com), and the "Bingo Cards" plugin on that service. I generated a few cards, but feel free to add more. Grab one, and if you know me, text me a picture of your card! Together, we'll celebrate (or not) how wild the new year was.

Also, be nice, and don't create anything hateful or in bad taste, or I'll have to restrict access to this.

## Full list of current options, if you're curious

New Pope 🔄
Pandemic ☣️ 1M+ 💀
Biden 🚫 Frontrunner
Aliens Confirmed 👽
Trump Jailed 🔒
Santos Elected 🎭
WW3 💥
🌎 Leader 💀
Balkan war 💥
AI Cult 🤖🙏
Dwayne 🪨 Candidacy 🇺🇸
BP-style Disaster 🛢️🐦
New Island 🌋
Astronaut Explosion 💥🚀
New 🇺🇳 Member
Kanye Wears 卐
Will Smith Smiths 🗡️
Simpsons Ends 💀
Melee 💥 in Kashmir
🇷🇺 Invades Again
🇷🇺 / 🇺🇦 Peace 🕊️
🇵🇸 Annexation

### My card

Just in case, and so I can point to this at the beginning of 2025:
![](https://scribedbystardust.com/img/2024-bingo-card-personal.png)


2024 Bingo Cards


"Scripture interprets scripture"[1](#sources) is a popular claim in American Evangelical circles. It is also a lie. It's based on an assumption called univocality, which is very commonly made, but promotes an understanding of the texts that comprise the bible that goes against how the texts were intended by their authors, and denies them their own voices, in favor of centering the cultural biases of those doing the harmonizing.

## What is univocality?

I'll let biblical scholar Dan McClellan define univocality:

> Univocality means a single voice. It is the dogma that holds that the scriptures (Bible or Quran), as the inerrant and/or inspired word of God, represent God’s consistent and unified position and message. It does not contradict itself.[2](#sources)

## Problems with univocality

Univocality smuggles in a lot of assumptions, each of which could easily have entire books written about them. Here's a few.

### Biblical texts disagree with each other

This happens all of the time. There are no shortage of Biblical accounts that describe events or values in ways that seem contradictory on face.

For example, Judas is described as having died from hanging in the book of [Matthew](https://www.bibleref.com/Matthew/27/Matthew-27-5.html). In The book of Acts[4](#sources), he instead fell, and exploded.

Many other examples of disagreements can be found in the scriptures. I'm not going to list them. Instead, let's look at the options we have for trying to resolve contradictions.

#### Harmonizing accounts

In many settings, from courtroom dramas to memoirs, we have accounts that slightly (or drastically) disagree, even about recent events. You might have a murder be seen by two eyewitnesses, one of whom claims to have seen a blond man shoot the victim, and one who swears that the murderer had red hair.

If you are ideologically pre-committed to the factuality of all accounts, you can **always** harmonize them if you try hard enough. The man could have had red hair on one side of his head, and blond on another. He could have put on a wig in between gun shots, or realized that his first shot wasn't lethal, called a fellow criminal, and another man delivered the killing blow.

Even if you add other witnesses claiming very different combinations of stories (the murderer used a knife, there were three victims instead of one, the murderer had highly visible neon clothing), you can always harmonize accounts, if you are willing to try hard enough.

#### Holding accounts in tension

However, a better thing to do when dealing with contradictory accounts of an event is to hold them in tension. This doesn't mean that the accounts can't be different viewpoints of the same event (maybe there really were two murderers, and smoking bullet casings could be found in different locations).

But not all contradictory events should be harmonized, just because they can be. A sequence of implausible harmonizations of events is just that - implausible, even if technically possible.

#### Back to Judas

A common narrative to try and harmonize the contradictory accounts of Judas's death is to say that both accounts happened. Luke simply neglected to mention the hanging, and Matthew just wasn't interested in the grotesque explosion. Instead, Judas presumably hung himself, rotted for a week or two, and then dropped from the rope and exploded.[5](https://answersingenesis.org/contradictions-in-the-bible/how-did-judas-die/)

This is possible, but implausible, and a modern attempt to harmonize an event with multiple competing narratives early in the Chistian community's history. Papias, a Greek church father who lived ~ 60 CE - 130 CE, gives two completely different accounts that both disagree with Luke and Matthew, saying that Judas had his body swell up grotesquely, urinating pus and worms, until his head was so swollen that he couldn't fit through a thoroughfare intended for wagons. After he died, "to this day no one can pass that place [where he died] unless they hold their nose, so great was the discharge from his body."[6](#sources)

That's obviously not the same story that Matthew is telling (although it may be a drastically more lurid version of Luke's).

### Polyvocality

Regardless of if you view the texts of the Bible as divinely inspired, or just texts, univocality is an assumption that does not allow authors to say what they are trying to. Rather, it forces you to subbordinate some authors to others.

This has the practical effect of privileging some authors and viewpoints, while marginalizing others. In American Evangelical Christianity, this is frequently done by centralizing teachings from Paul (and perhaps especially the pseudo-Pauline letters in the New Testament) at the expense of the more social-justice oriented prophets, or the strict concern for the law of the Priestly texts.

I would argue that a more honest understanding is to let authors say what they wish, and to live with the contradictions that follow however you would like.

Paul believes in celibacy (and maybe monogomy, but it's better not to). Song of Solomon celebrates love outside of a socially sanctioned relationship (quite graphically).

John believes in one god. Much of the old testament implicitly recognizes a pantheon, and places Yahweh in a more-or-less privileged place amongst them. One of the more explicit recognitions of this is [Psalm 82](https://app.logos.com/books/LLS%3A1.0.710/references/bible%2Besv.19.82.2?registration_source_host=biblia.com&tile=left&zzls=2eMKcTcKNwrEOw4IwEEPDv8OFc8KlwqprNgQrE2zChMKhasKvw61JScKuJMOXIhTDpcOfCcKdw5jCrGfDmS9jwqfCmFgCTMOXYMOtA8K5w5PCusKew4XDuz7CjAnDpsKRK8OUJcO0wp5gw5DCqsKIS20QwqXChAbCr8KNw6IHJsOjAMO3w5rDnUjClcODXMKHw4gWScKiWhgLLyNPTMOjwoXDkmDDkVjDrEzDr8Krwox0wrTDk8OmwpxFwqnCh8OKw67Cp8KJPC%2FCilLCmsKMflDDnivDk8K4w5HCn3DCjSzCkcK1w6bCrsKUZ8O5AgrDoEXCiQ%3D%3D), which features Yahweh as passing judgement over a high council of deities (replacing the Semitic high deity El by seizing his traditional role, and promoting Yahweh in his stead).

Paul believes in faith that counts for righteousness. James disagrees, and says it is works that play that role, even using the same passage to illustrate his point (likely in direct response to Paul).

It is neither helpful nor intellectually honest to insist that all authors, facts, and viewpoints in the bible harmonize, any more than it would be to attempt to harmonize the sayings of Gandhi with Winston Churchhill and Conan the Barbarian.

### Sources

[1](https://www.christianitytoday.com/ct/2022/february-web-only/bible-reference-tool-old-testament-scripture-interpretation.html) Christianity Today, 2022

[2]Dan McClellan, "On The Myth of Scriptural Literalism", (https://danielomcclellan.wordpress.com/tag/univocal/)

[3] Matthew 27:5-8

[4] Acts 1:18-19

[5] Answers in Genesis, "How Did Judas Die?" https://answersingenesis.org/contradictions-in-the-bible/how-did-judas-die/

[6] Moss, C 2019, 'A note on the death of Judas in Papias', New Testament Studies, vol. 65, no. 3, pp. 388-397.
https://doi.org/10.1017/S0028688519000080 https://pure-oai.bham.ac.uk/ws/files/77433625/Moss_A_note_on_the_death_of_Judas_in_Papias_New_Testament_Studies_2019.pdf

[7] [Psalm 82](https://app.logos.com/books/LLS%3A1.0.710/references/bible%2Besv.19.82.2?registration_source_host=biblia.com&tile=left&zzls=2eMKcTcKNwrEOw4IwEEPDv8OFc8KlwqprNgQrE2zChMKhasKvw61JScKuJMOXIhTDpcOfCcKdw5jCrGfDmS9jwqfCmFgCTMOXYMOtA8K5w5PCusKew4XDuz7CjAnDpsKRK8OUJcO0wp5gw5DCqsKIS20QwqXChAbCr8KNw6IHJsOjAMO3w5rDnUjClcODXMKHw4gWScKiWhgLLyNPTMOjwoXDkmDDkVjDrEzDr8Krwox0wrTDk8OmwpxFwqnCh8OKw67Cp8KJPC%2FCilLCmsKMflDDnivDk8K4w5HCn3DCjSzCkcK1w6bCrsKUZ8O5AgrDoEXCiQ%3D%3D)


The Univocal Lie


## Gaiapunk

Gaiapunk is an aesthetic of human flourishing for a world that is adrift. In perilous times, new, positive visions of the future are needed.
Gaiapunk is unapologetically pro-human, egalitarian, and optimistic.

## We hold these things to be worth striving for:

That we want to live in a society where all humans are deemed worthy of food, shelter, and dignity.
That it is good for humans to be able to pursue cosmopolitan values to the extent that they do not interfere with the rights of others, and without coercing others into their will.
That given that humans may be the sole intelligent species in the cosmos, that we bear the light of consciousness to as much of the galaxy as possible - without trampling on native life forms that we may find on other worlds, or our own.
That the path for a successful human species lies through optimistic embrace of technology that is designed to empower us.
That despite the promise of technology, misused technology is an existential threat to our happiness and even existence.
That no human can thrive alone. A thriving individual requires a thriving community.
That nature is neither to be exploited nor frozen in time for preservation, but to be intelligently sculpted into something more beautiful than we have today.

## We long to create a galaxy where

Humans enhance our bodies and minds beyond the current limits of our species.
Death and disease are memories.
We treat our fellow humans and other conscious life around us with empathy, not exploitation.
We spread throughout the stars.

## Aesthetics

These are merely my takes on a Gaiapunk aesthetic - I expect this will evolve, and invite everyone to steal, tweak, and otherwise modify this vision.

### Colors

Gaiapunk paints the world with joyful, celebratory colors from nature. It is the green of foilage, the gold of sunlight, and warm reds, browns, yellows, and pinks.
It is an explosion of hues and shades, but unlike cyberpunk, it favors hues that favor peace and mental health, rather than burn into your retinas at night.

### Architecture

Gaiapunk architecture is:

-   organic > angular
-   incorporates and sculpts nature, rather than paving over it or leaving it untouched
-   centers community-based, walkable communities, rather than individualistic housing.
-   Blends abundance with beauty. Public and private spaces are filled with fruit trees, herbs, and other edible plants, some of which we have today, and some engineered to provide light at night, or to perfume an area.

### Music

Gaiapunk music is a riot of different genres:
Relative to other music, it is more likely to feature:

-   Patterns that are good for your mental health. (Less screamo, more meditation)
-   Sounds from nature
-   Collective effervescense
-   Blending new synth sounds with traditional acoustics from a variety of cultures.

### Technology

Gaiapunk technology: - encourages body modification and genetic engineering. Trees glow pale pink and gold in the dark. Some implant tails into their bodies, some gain claws for climbing, some receive gender affirming surgeries, and others receive synthetic lungs to dive more deeply for longer. - uses energy responsibly, cleanly, and unapologetically. We long for clean skies, clean water, clean air, and clean bodies. We also long for nuclear powered starships, bullet trains, and space elevators. These goals are not in opposition. - is responsibly designed. gaiapunk technology does not immerse people in devices and digitally-fueled depression, but helps them to live in the real world more fully.
It is less screens, and more holograms. It is not the gaudy advertisement clusterfuck of facebook or the rage machine of Twitter, but the collaboration of Wikipedia, VR calls with friends that make you feel like you're in the room, and the synergy of brainstorming in a quest for truth.

### Gaiapunk societies

Gaiapunk societies prioritise a baseline of dignity for every person first, and then prioritises healthy communities, allowing for individual growth in the directions that individuals prefer.
We condemn the following things:

-   Conspicuous consumption: Gold-encrusted food, bottles encrusted with crystals, and massive yachts for one individual hoard resources that should be used to uplift our fellow human from poverty into a life of flourishing.
-   Exploitative technology: designing technology is a sacred trust, and must prioritise the wellbeing of the user against profit.
-   Tribalism: We are a single species, and must stick together. We recognise our evolutionary tribal instincts, and will strive to minimise their impacts on the way we treat our fellow humans.
-   Exploitative corporations: Cooperation is essential for societal flourishing, and for achieving big goals. But it is wrong to set personal profit above the wellbeing of other people. They must both be positive.


A Gaiapunk Manifesto


i remember being a brittle jar
filled with blackest oil and ink
but now the glass has cracked

and i inhale sunlight.  
even the oil slick shines


Leaving Depression


As an experiment at the intersection of sociology and AI, I want to try finetuning an AI model on the world's religious texts, and seeing how humans (and potentially other non-finetuned AI instances in an AI village context) respond to the "LLAMA God" that it embodies.

## What I've done so far.

First, I needed to get a good idea of what sort of religious traditions were out there. There is a lot of ambiguity on the line between philosophy and religion. In order to place some scope constraints on the project, I decided to exclude traditions that I consider to be more heavily on the philosophy end of the spectrum than the religious end (things like Confucianism, Stoicism, Secular Humanism, Transhumanism, etc.).

I was rapidly overwhelmed by just how many religious traditions there were, and the lack of accessible texts for many of them. Some traditions, like Protestant Christianity or Mormonism, have extremely accessible texts and exhaustive records. Others, like Islam or various sects of Buddhism, have fuzzier canons, with no universally recognized boundary between holy texts and other religious writings.

Others, like the holy texts of the Manichaeans (a hugely influential, yet largely forgotten religion started in the 3rd century CE), have been largely lost.

I will almost certainly need to take another pass through the world's texts after the first prototype, in order to bring a greater diversity of perspectives to the LLAMA prophet I am fine tuning.

Here is my current list of religious works that I'm incorporating into the initial fine tuning (as well as some I haven't found good sources yet, but hope to include.) If you know a good source for the missing ones, please let me know!:

## Abrahamic

#### Judaism

-   [x] Talmud
-   [ ] Midrash

### Christianity

-   [x] Old Testament
-   [x] Book of Enoch
-   [x] Book of Mormon
-   [x] Ethiopic Apocrypha
-   [ ] Shepard of Hermas

#### New Testament

-   [x] Catholic Apocrypha

#### Gnostic

-   [x] Corpus Hermeticum
-   [x] Gospel of Judas
-   [x] Gospel of Thomas
-   [ ] Gospel of Mary

### Islam

-   [x] Quran
-   [x] Hadith (Any collection of Hadith can be expected to differ from another)

## Eastern

### Hinduism

-   [x] Bhagavad Gita

### Buddhism

-   [x] Pali Canon

### Sikhism

-   [x] Guru Granth Sahib

## 'Pagan'

-   [x] Hymn to the Aten
-   [x] Poetic Edda

### Other

-   [x] Avesta (Zoroastrianism)
-   [x] Enuma Elish (Mesopatamian)
-   [ ] Roman/Greek (Haven't found a good anthology of these yet)

### Culty Bonus

-   [x] Urantia
-   [x] Bahai
-   [x] Dianetics

### Extinct, non 'pagan'

-   [ ] Manichaean (there are some preserved works, though many are lost)

## Next Steps

-   **Finetune and deploy the model!** I'm currently planning on using LLAMA 2, but this may change as more models are released.
-   **Second training run.** Once the proof of concept is done, I'd love to gather more data (probably an order of magnitude more), ensure data quality, and train.
-   **AI and human interactions.** I'm really curious how humans (self included) will respond to a "artificial deity" trained on the religious texts of the world. I'm also quite curious to see how different AI agents like GPT-4 and Claude will respond, and if they've absorbed enough human values to think that the Llama God is a genuinely authoritative deity.
-   **Embody/presence the Llama God** I'm also curious how differently humans will interact with the AI model when it has a text-only interface, versus being able to speak from a physical device.

Stay tuned!


The LLAMA God, Part 1


I want to be a 1000 year old vampire, and get there as quickly as possible.

The archetype of the being who has fought in revolutions, created inventions, and seen and experienced an impossible amount of things isn't one that is wholly fictional.

People like [Da Vinci](https://en.wikipedia.org/wiki/Leonardo_da_Vinci), [Newton](https://en.wikipedia.org/wiki/Isaac_Newton), and [John Von Neumann](https://en.wikipedia.org/wiki/John_von_Neumann) fit an overwhelming blaze of activity into single lifetimes, transformed whole fields (or multiple fields), and would be a hell of a party guest.

I won't claim to be an expert on any of the above people, or to have a deep knowledge of how people become polymaths. But it seems that the promising path to walk in the steps of such men is to get close to "the metal" of reality.

Leonardo wanted to learn how to paint better, so he learned anatomy. He understood it in a visceral way that can't be learned from a textbook. Newton hungered for truth, so he dove into astrology and biblical chronology, thinking for himself. (And invented calculus, which he seems to have been substantially less interested in).

Von Neumann was potentially one of the smartest men to ever live. He started game theory as a mathematical field, worked on the Manhattan project, and is revered as a founder of computer science. He also [**literally created the field of cellular automata on graph paper**](https://en.wikipedia.org/wiki/John_von_Neumann). I can't emphasize enough how insane that is. Oh, and he came up with the idea of the Singularity, and was one of the discoverers of the idea of Mutually Assured Destruction.

Obviously, these were minds that were easily one in a million (potentially far higher out of distribution, especially for Neumann).

But it seems that diving deeply into solving real problems in a variety of fields is the fastest way to move. For this, it looks like it is necessary to understand each part of the project, if not to do it yourself. If one is learning how radio works, you should know the math, and build it on your bench with wire and razor blades, rather than just reading about how to turn on a ham.

There's a lot to experience in life. And it seems like hopping straight into out of unusual experiences is the best way to do it.

So learn fire juggling! Cook haute French cuisine! Hammer a blade out of rock, and throat sing!

Deep drinkly out of the cup of life, so that you can have lived a thousand lifetimes within your own.

Even a thousand year old vampire, with the wrong attitude, could have lived a life full of nothing but television and accounting. There might not be a right and wrong between the vampire of myth that has the strength and wisdom of a hundred men and his accountant counterpart. But I know which I'd prefer to be.

As [Robert Heinlein](https://www.elise.com/quotes/heinlein_-_specialization_is_for_insects) said:

> A human being should be able to change a diaper, plan an invasion, butcher a hog, conn a ship, design a building, write a sonnet, balance accounts, build a wall, set a bone, comfort the dying, take orders, give orders, cooperate, act alone, solve equations, analyze a new problem, pitch manure, program a computer, cook a tasty meal, fight efficiently, die gallantly. Specialization is for insects.

_[Credit for this way of thinking about life.](https://www.lesswrong.com/posts/5QpufhoH2ASnppsjs/how-to-become-a-1000-year-old-vampire)_


Become a Thousand Year Old Vampire


## What makes morality?

> Is the pious loved by the gods because it is pious, or is it pious because it is loved by the gods? - Plato, [_Euthyphro_](https://mason.gmu.edu/~rcherubi/eun.htm#:~:text=Then%20the%20pious%20is%20loved,is%20for%20a%20different%20reason.)

> ... everything which [God] wills must be held to be righteous by the mere fact of his willing it - John Calvin, [_Institutes of Christianity_](https://web.archive.org/web/20220206192601/https://www.biblestudytools.com/history/calvin-institutes-christianity/book3/chapter-23.html)

> if God does not exist, everything is permitted. - Ivan, in the Brothers Karamazov, by Dostoyevsky.

Growing up in a protestant Christian philosophical tradition and culture, downstream of Calvin's belief on Plato's question, morality was defined by God. Good actions, by definition, were those that were approved of by God. This is known as divine command theory, and was the belief of a lot of foundational thinkers in Protestant theology.

## Vertical Morality

I'll be referring to this sort of top-down morality as **vertical morality.** In vertical morality, authority creates the boundary of what is right and wrong.
Murder is a sin against God, first and foremost. Poaching deer in the king's forest is an offense against the person of the king. Failing to worship the Emperor is a slight against is name.

In vertical morality, the seriousness of your offense is directly proportionate to the majesty, worth, or excellence of the one you offend. This often correlates with their perceived power.

It is this sort of logic that has the conclusion that the smallest possible white lie deserves an eternal period of torment.

> ... if I say to God, "I'm going to do this against you"—this small sin, say, spitting on somebody... you've said no to God... every sin, from the smallest to the greatest, is against God and not just against people and their consequences. And in that sense every sin is infinitely heinous. - [John Piper](https://www.desiringgod.org/interviews/are-all-sins-equal-before-god)

To which I say, **fuck that conclusion.** Imagine a two year old (or fifteen year old, if it makes a difference) child who tells a single white lie, and then gets flattened by an eighteen wheeler. Under the logic of vertical morality, not only did the child deserve to be spattered on the pavement, but they also deserve to be raped, flayed alive, and then slowly killed. And all of that, according to a Christian doctrine of eternal hell, is infinitely better than actually going to hell.

If you believe in the idea of hell as a punishment for "sinners", do not glance away from that consequence. Pause. Roll it over in your mind. Would you torture a child for a single moment because they lied to you? How hot would you heat the pincers?

Of course you wouldn't - you're not a monster. Are you better than the Christian god? Are you more merciful, or more forgiving?

## Horizontal Morality

If we don't get a barrier between right and wrong from an authority figure, where do we get it from? We can get it from an obligation to treat other beings with respect for their wellbeing and autonomy. We can treat them as we want to be treated (or perhaps as they would want to be treated).

"But that's not an objective sense of morality!" Sure, I concede that. I have no knowledge of a universal sense of ethics inherent in the universe in the same sense that _F = ma_, or _2 + 2 = 4_.

But we both agree, right? There may be some alien race with a very different view on morals, one that we may find repugnant. And Ted Bundy or Genghis Khan might disagree that they have an obligation to treat their fellow man with love. But so much the worse for the Khan!

Morality may not be objective. It may be nothing more than you and I looking each other in the eyes, seeing each other's humanity looking back, and mutually cherishing that. But that doesn't mean it's worthless. If we both look at a great work of art by Monet or Banksy, we can't measure it's beauty in pounds, or ounces. But we can feel the glow within, and appreciate it nonetheless.

Rather than viewing morality as law from above, perhaps we'd do well to view it as a piece of art we are all painting, day by day.


Horizontal v Vertical Morality


When I last published an update on this project, the UI was (mostly) finished. But the rules engine was basically non-existent. Fast forward a few weeks, and I now have a functioning game!

First, I needed to make a few UI changes in order to have the game be useable. In Hnefatafl, five squares on the board are restricted (only the king can move in them). In order to designate these as different, I ended up overlaying a background image of white-noise texture from a texture generator.

```css
.square {
    //... other code removed for simplicity
&.isRestricted {
        background-image: url('http://api.thumbr.it/whitenoise-361x370.png?background=ffffffff&noise=050505&density=50&opacity=31');
    }
}
```

I also added a system where if a user tries to make an illegal move, the square will flash red in order to communicate that a different move is needed, although the game registered their attempt.

Can't touch that.
Then it was time to add win conditions. I'd already stored the board state in a single two-dimensional array that looked like this.

```ts
const INIT_BOARD_STATE: BoardState = [
    [0, 0, 0, 2, 2, 2, 2, 2, 0, 0, 0],
    [0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0],
    [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
    [2, 0, 0, 0, 0, 1, 0, 0, 0, 0, 2],
    [2, 0, 0, 0, 1, 1, 1, 0, 0, 0, 2],
    [2, 2, 0, 1, 1, 3, 1, 1, 0, 2, 2],
    [2, 0, 0, 0, 1, 1, 1, 0, 0, 0, 2],
    [2, 0, 0, 0, 0, 1, 0, 0, 0, 0, 2],
    [0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0],
    [0, 0, 0, 0, 0, 2, 0, 0, 0, 0, 0],
    [0, 0, 0, 2, 2, 2, 2, 2, 0, 0, 0],
]
```

I used 0 to represent an empty space, 1, for a light piece, and 2 for a dark piece. I eventually added a 3 in the very center to represent the king.

The biggest issue I ran into while writing the rule engine code is that I forgot the fact that although Jotai (the state system I am using) has updates that are promises, they should not be awaited, since they are based on React hooks internally. Instead, I needed to detect state changes by sticking my reaction code inside of a useEffect block.

```ts
useEffect(( => {
if(!lastMove){
return;
}

    // code here checks if a piece should be captured.

}, [lastMove])
```

After being stuck on this issue for an embarrassingly long time, I finally figured it out.

The basics of the game are now complete! White can win by escorting its king to any of the four corner squares. Black wins by capturing the king.

Future steps:

1. Get the game to look good on all screen sizes. There is at least one CSS bug I'll have to work through as part of this.
2. Convert to React Native or Ionic.
3. (Maybe) Add custom art for the pieces. I have a friend who is an excellent artist who has expressed interest in helping out with this. For now, his identity will remain a surprise. :)
4. Deploy! I'd love to get it in an app store or two, as well as having a public facing web version.

I've also had the pleasure of playtesting this with a handful of friends and colleagues. We were all surprised that the game was fun, relatively simple, and challenging. It makes sense why the vikings played it for so long.

Cheers,
Joshua

P.S. If looking at code is something you enjoy, you can view the full code here. It is open source, so feel free to modify it, adapt it, or any other benevolent use that brings you happiness.

Feel free to drop me a line.


The Hnefatafl Saga, Part 2


_(Note - this is a post recovered from my last personal blog - sadly, images of the program were lost in the transition, but I'm keeping the text for posterity. Because of that, the photos for Hnefatafl are from Midjourney instead)_

A few weeks ago, I decided that I wanted to embark on a personal coding project, since after the recent app that I published, work projects had been the vast majority of my coding time. And so, I began the hunt for a board game that would be relatively easy to code but also both challenging and interesting.

Hnefatafl (pronounced ne-fə-ta-fəl or neffa-taffle), although it may look like the "qwertyuiop" produced by hitting a keyboard, is a board game with surprisingly deep roots. In its many variations, the history of Hnefatafl goes back to before 400 AD, and was popular throughout much of Europe up until the mid 1000's.

I knew that this game was going to require fairly complex state management. The natural choice to reach for here is probably Redux, for anyone who enjoys using a React-based tech stack.

However, although Redux is a mature tool, it is also comes with a painful amount of boilerplate, and I wanted to experience with something a bit lighter weight. I'd heard good things about Jotai, a minimalistic, TypeScript-oriented state framework, so I decided to give it a try.

Fast forward a few weeks, and I currently have the skeleton of Hnefatafl completed. The UI is mostly finished. The foundation of a rules engine is complete, although it still needs to be fleshed out.

After progressing this far, I have a far greater appreciation for people who make video games that are exponentially more complex.
So far, Jotai has been capable of everything I've needed. As always, working with React and Typescript has been a breeze. I've also really enjoyed using SASS (a CSS extension language) on this project. I've found it to be a lot more intuitive than LESS, its main alternative.

Future steps? Eventually, I'd love to package this as an app (either in React Native, or Ionic), and publish it. For now, however, I'll be focusing on the rules engine. After that, I'll consider adding customization options such as different color themes, and different board sizes, since there are many variants of Hnefatafl.


Home page

Posts

Welcome to my digital garden 🍄

Topics