Low-Level Functionality

Connections

The Design of a Connection

Connections are how nodes are wired together to perform their weird distance-defying remote communication. This is also how they form higher level nodes to begin with. At the bottom level, a discovery protocol allows them to send connections to physically neighboring entities. A connection is made of three main parts: a base, an address, and an anchor.

Base

A base is where a connection is made. The other parts are made at the base to activate the connection. Each is identifiable. Bases are reusable and, if necessary, a bottom-level node can break them, repair them, or regulate their production. Production regulation increases or decreases base production by encouraging or discouraging the use of associated default instructions. If a base in use is broken, the connection it formed is disconnected. The receiving node of a connection has no access to the base of the connection though, so it can only be destroyed by the sending node. The receiving node can see the base, but the base is out of its reach to affect. All bases appear to be built the same, but faulty bases do occur. When it’s noticed that a base isn’t forming a connection after several attempts, the base is destroyed immediately and a different base is used instead.

Address

An address provides the targeting information for connecting one node to another. Addresses lock into bases with or without anchors. An address is an extremely long, complex chain of unique information that allows a base to activate a conduit to a particular target. Bottom level nodes appear to find addresses by proximity, as if physically neighboring entities outgas their own addresses. However, addresses can also be shared across connections, allowing for communication with no regard for distance. The parts of addresses automatically replenish themselves. Nodes are not aware of the process involved in generating their parts. They are only aware of the process of putting those parts together to form an address and placing that address in a base. New parts of addresses are continually being discovered, so the large structure is even larger than nodes are aware. Addresses are, thus, likely to be at least partially built by obfuscated or mechanical processes that are outside of the awareness of the node that’s building them.

Addresses stay at the base unit, out of the reach of the receiving node. If the sending node destroys an address that is in use, the connection is severed. The same thing happens if an address is modified while it is in use. The connection is severed and the address is unusable, even if a new anchor is loaded into the base with the repaired address. Broken or modified addresses no longer function even if they are repaired. This may be because of the obfuscation problem, as nodes do not appear to be aware of the structure of an entire address. Addresses don’t have to be created at bases. However, if they are not created at a base, the address itself is immediately lost upon creation.

If the sending node addresses itself or a node that no longer exists, the connection fails. A self-addressed connection will not form. However, nodes are aware of what is in each of their bases. This opens the possibility of using non-activating bases as a crude form of reliable memory, especially since other nodes can see what is placed in those non-activating base units as well. These non-activating units even take less effort to observe than the ones that do activate. However, a node can forget the reason an address and anchor are in a base, so to use this as a form of hard memory all of those details would need to be coded into each anchor and/or address.

Anchor

Anchors made in a base with an address, but do not stay in the base. When an anchor and a working address are placed in a base, they activate a conduit automatically. This transports the anchor to the receiving node. The sender then has no access to the anchor. However, the receiving node does have access to the anchor and can modify or destroy it, if it can detect the anchor that was sent. Some anchors cannot be detected by some nodes. These detection abilities appear to be species dependent.

Each anchor structure has its own meaning and properties. An anchor, like an address, is made from parts that automatically replenish themselves. However, an anchor does have a simpler structure, compared to an address. An anchor has eight sections, but the first two are optional. The first optional section is a state, the second optional section is a mode, and the last six sections are required. A state requires a mode, but a mode does not require a state, meaning an anchor is built from end to start, but read from start to end. Originally, Translator told me about the state and six required sections but was completely unaware that the mode section existed. That was a later discovery that Translator made after several rebuilds. That implies that it is possible that anchors, like addresses, still have some hidden variables to them.

Each section of an anchor has three possible parts, the first two of which are optional, with the first requiring the second, as with the sections. It seems that they too are built from end to start but read from start to end. This is the same for all sections of an anchor except for the mode. For the mode, the third part is never there, the second part is required, and the first part is optional.

Each of the three possible parts of a section has a scalar value. I have no idea what those scalar values represent. I simply asked if the possible values of each part could be put in an obvious order from least to greatest, and it turned out that they could. That makes it easier for me, as I can just give each option for each part a number. The first part of a section, if it is there, has four possible values. The second part, if it is there, has two possible values. The third part, if it is there, has six possible values.

Conduit

When these three parts are together, a connection is formed. The connection itself is a conduit at both ends, connected by the base and address at one end and the anchor at the other. Through these conduits, two-way communication is possible, a few types of parts can be supplied, and attention can be projected or shared for encouragement or inhibition. Once a conduit is formed, neither of the nodes involved in the connection can affect the conduit directly. A conduit can only be altered by affecting the base, address, or anchor.

The Language of Anchors

As anchors are short and convey meaning. To make it easier to work with anchors, I used the rules for how anchors are structured to create a verbal representation of anchors. Each syllable represents a section. The component parts of each section dictate what syllable is used. Part 1 sets the place in the mouth that the consonant is expressed, from front to back. Part 2 sets whether the consonant is voiced or voiceless. Part 3 sets the vowel sound, moving from the front of the mouth to the back. If part 1 and 2 aren’t there, the consonant is an S sound. If part 3 isn’t there, as in an anchor’s mode, the vowel is replaced with a glottal stop.

Pronunciation of anchor sections based on their constituent parts.

Part 1	Part 2	Part 3							Pronounced like:
		—	1	2	3	4	5	6
—	—	s'	su	so	sa	se	sö	si	set
	1	p'	pu	po	pa	pe	pö	pi	pet
	2	b'	bu	bo	ba	be	bö	bi	bet
1	1	f'	fu	fo	fa	fe	fö	fi	fox
	2	v'	vu	vo	va	ve	vö	vi	vote
2	1	t'	tu	to	ta	te	tö	ti	toe
	2	d'	du	do	da	de	dö	di	dot
3	1	ʃ'	ʃu	ʃo	ʃa	ʃe	ʃö	ʃi	shake
	2	ʒ'	ʒu	ʒo	ʒa	ʒe	ʒö	ʒi	azure
4	1	k'	ku	ko	ka	ke	kö	ki	call
	2	g'	gu	go	ga	ge	gö	gi	go
Pronounced like:		uh-oh	cool	no	father	set	first	tree
An anchor that is stateless and modeless starts with s’s’, pronounced with two swift S sounds. Dashes are used to separate these long words into more easily pronounceable sections, written as SM‑123‑456, where S is the state, M is the mode, and numbers indicate the required parts.

Even though there are 909,193,450,176 different possible anchors, only a little over a thousand are currently known to be in use by the human body. Each anchor has its own features and communicated purpose. For example, -baʒeva-sosada is the root of 21 cooperation-related anchors. All of its permutations are visible to all bottom-level human nodes. S's'‑baʒeva‑sosada, the anchor version from this root that has no state or mode, communicates to the receiving node that the sender is requesting attention power from the receiving node. In contrast, a weird type of cooperation built from this root is böʒ'‑baʒeva‑sosada. This anchor indicates that the sending node thinks that the receiving node is dead and that the sender caused the receiver’s death. Yes, nodes consider a fight to the death a form of cooperation. It’s weird. This anchor is visible to all human nodes, but it is not traceable to the origin node’s address. It also cannnot be used to transmit any communication, inhibition, or attetion power. It apparently only exists because a node can attend to a node that it has killed. This is a safety feature so that nothing dangerous or confusing is transmitted along a connection to a dying node, if at all possible. However, these types of safety anchors also make it difficult to determine if a node is actually dead or not.

Default Instructions

At the bottom level of a creature’s network are indivisible nodes. These nodes have built-in instructions that they are born with. These instructions appear to follow a DNA-like structure. What I mean by that is that the instructions are a mirrored linear code made up of two sets of paired characters. DNA is made of the pairs A-T and C-G. I have labelled these pairs of default instruction characters a-b and x-y.

I cannot say whether these default instructions are DNA, as I have compared readings of their code to readings of DNA, with no match as of yet. The problem with comparing these default instructions to DNA is that they are not read like a book or an article, where you can move straight through from start to finish. Instead, almost all of it is being interpreted and obfuscated. Trying to get around that process has proven extremely difficult. Translator tells me that it is currently aware of 47 different default instruction reading methods. Each one provides a different string of characters. I do not know what the process differences are between these methods. I do know that many methods end up reading very long strings of the same character repeated 10 to 30 times, which is not even close to what I find when I search through DNA. Manually working my way through each default instruction read method, comparing it to DNA, is an incredibly slow process. As such, I have not worked my way through all of them.

Default instructions can be read, discouraged, and encouraged. They usually cannot be written. I did eventually discover a default instruction writing method, originally being used as a way of permanently embedding addresses in other people’s nodes for database searches. It is, however, effective for other types of changes. This method of writing default instructions replaces what is already present. It is not an insertion method.

In the human body, there are three types of nodes, distinguished by their default instruction content and size: type A, type B, and type C. Type A nodes have the most instructions and appears to be a normal human node. Type B is a state that a type A node transitions in and out of where it has significantly reduced default instructions. Type C nodes have very few default instructions. They only appear on the network for a few milliseconds every few seconds. My working hypothesis is that a type A node is a normal human cell, a type B node is a normal cell undergoing mitosis, and a type C node is a mitochondrion. The instruction sizes and population sizes, when accounting for sporadic appearance of type C nodes, appear to be the right proportions for that to be the case. However, proportions are extremely rough and difficult to get from a network. A type A node has roughly 5000 times the default instructions of a type B node, and a type B node has roughly 40 times the default instructions of a type C node. Do not rely on these numbers, though, as they were arrived at through ballpark estimates made through doubling, halving, and adding from a one-to-one ratio until Translator thought the amount was about right.

Network Actions

• do nothing
• discourage (inhibit)
• encourage (amplify)
• build
• disassemble
• rebuild
• share information
• receive information
• spy
• request
• unify
• disunify
• direct
• - 5-28 anchors of the same type
• - 1103-1110 anchors of different types
• - 1 anchor
• - 5-34 anchors of different types
• - 21-70 anchors of different types
• - 1 anchor w/ attachment
• - 1 anchor w/ attachment
• cut another's anchor to self
• modify another's anchor to self
• trace another's anchor to self
• watch/ignore another's anchor to self
• feed amplifying content into base without addresses or anchors, to self-amplify attention