Data/Information Tiering Revision

[Agnaa]

Referring to this section of the Tiering System page, and any parts of other pages to a similar effect

Please note that existing as a drawing or being made of data/information is not to be ranked at this tier, as such beings are still 3-dimensional, but on an incredibly small scale.

I think this argument is partially valid, but mostly flawed.

I'd hold it as valid for drawings, as those can typically be read as the literal drawings on the screen, without an actual world beyond that. But for the other cases, involving data, I think they're a lot more similar to books.

Take "a character approaches a rock".

• In a drawing, the character and the rock occupy distinct locations in our reality, the representation maps to the actual content in a pretty one-to-one way. When the character gets closer to the rock, the representation of those two objects necessarily get closer together. If the rock got bigger, it would have to get bigger by a similar factor in our reality too.
• In a book, this sort of link is entirely broken. With sentences such as "I was far from the rock" and "I made my way closer to the rock", while the world being represented has the character and the rock get closer in the second sentence, the actual representation of that in our reality (i.e. those words) actually get further apart. I think this goes to show that the representation of their world isn't their literal world. If the rock got bigger, that would only require some words describing that change in our world.
• From this, I think we can see that data lands is more similar to a book. For a character to get closer to a rock, some data would change, but those electrons representing each wouldn't necessarily get closer to each other in our physical world. If the rock got bigger, that would involve changing some data, but it wouldn't require a corresponding change in the size/amount of electrons representing it.

I think we should take this more nuanced view towards indexing fiction-within-fiction.

As a bonus, I think we already realise that this would be a kinda silly way to treat things when looking in the other direction; we don't downgrade verses to 3-A because reality is seen as a simulation by higher beings, under the view that such would imply their transcendental beings are only finitely more powerful than the electrons in the simulation. We do this since we realise that the data is a representation, and if they can affect a part of that as large as a timeline in the world being represented, that we should index that, not a comparison to the underlying data structure.

NOTE: This will not change whether or not profiles will get a rating for this; it will only change some characters who are 10-C for this reason to 11-A/11-C (depending on whether the difference is qualitative or not).

 

[Antvasima]

@AKM sama @DontTalkDT @DarkDragonMedeus @Mr. Bambu @Celestial_Pegasus @Wokistan @Ultima_Reality @Elizhaa @Qawsedf234 @ByAsura @Sir_Ovens @Damage3245 @Starter_Pack @Abstractions @LordGriffin1000 @Colonel_Krukov @SamanPatou @GyroNutz @Firestorm808 @Everything12 @Maverick_Zero_X @Crabwhale @Just_a_Random_Butler @DarkGrath @Dereck03 @Planck69

What do you think about this?

 

[Mr. Bambu]

I don't feel very strongly one way or the other, but after having spoken to Agnaa about this via Discord, I think it's fine. The character is represented by things in the real world that are not them- they are fictional.

So, sure.

 

[Antvasima]

Thank you for helping out.

 

[DarkDragonMedeus]

I am not entirely sure on the proposal, though I can at least see what Agnaa is trying to say about images often contradicting descriptions or interpretations of other things.

 

[DontTalkDT]

I mean, if you have a rock in a physics simulation, making it larger would occupy more memory and hence more physical stuff.

In any case, that depends on the depiction. I don't think we ever outright excluded R>F via computer simulation being a thing that can happen? At least not until recently. It just depends on whether something is virtual in the sense of being a computer program or in a more metaphysical sense. Honestly, I could see the same apply to pictures. You can either be the ink or you are something metaphysically separate just represented by the ink, similar to how you could be either the ink in a book or the thing represented by its writing.

 

[Agnaa]

I mean, if you have a rock in a physics simulation, making it larger would occupy more memory and hence more physical stuff.

Typically not, so I don't think that's a great default to use.

In any case, that depends on the depiction. I don't think we ever outright excluded R>F via computer simulation being a thing that can happen? At least not until recently. It just depends on whether something is virtual in the sense of being a computer program or in a more metaphysical sense. Honestly, I could see the same apply to pictures. You can either be the ink or you are something metaphysically separate just represented by the ink, similar to how you could be either the ink in a book or the thing represented by its writing.

Yeah, fiction can elaborate in ways that create exceptions, but I think the current default for low-info examples is bad.

I'd word the guideline using those archetypes as examples, with a note that some drawings could act more like metaphysical representations, and some data/text could actually be literally representative of the characters.

 

[DontTalkDT]

Typically not, so I don't think that's a great default to use.

In which kind of simulation that can handle realistic physics would it not be the case?

Yeah, fiction can elaborate in ways that create exceptions, but I think the current default for low-info examples is bad.

I'd word the guideline using those archetypes as examples, with a note that some drawings could act more like metaphysical representations, and some data/text could actually be literally representative of the characters.

Which formulation do you have in mind?

 

[Agnaa]

In which kind of simulation that can handle realistic physics would it not be the case?

Current ones. If we have two planets, we don't model the one with twice the radius (and thus, 8x the volume/mass) using 2x or 8x as much data.

There's many ways that sort of thing could be handled. By my estimation, the most-likely to least-likely ways would be:

• All objects keep track of that sort of thing using the same amount of data, entailing limits on the size of such objects.
• Sufficiently large objects can use arbitrarily more data, but such a thing would happen at certain thresholds (such as every power of 10 increase in a data point), and there would be no difference until those thresholds are met.
• There's multiple different object types, for different orders of object sizes, such that some need less data, but all objects in the same range would occupy the same amount of data.

At no point would every change in size result in a change of data usage. And even in the second case, where such changes occur most often, a lot of their data would end up not being resizable like that. On top of that, this increased data occupancy wouldn't scale at the same rate as the underlying thing it's representing. Being able to represent a mass that's twice as large does not even require the mass variable to occupy twice as much data; it barely needs to grow at all.

Which formulation do you have in mind?

Bit brain fried from work. I'll type something up in a few days.

 

[DontTalkDT]

Current ones. If we have two planets, we don't model the one with twice the radius (and thus, 8x the volume/mass) using 2x or 8x as much data.

There's many ways that sort of thing could be handled. By my estimation, the most-likely to least-likely ways would be:

• All objects keep track of that sort of thing using the same amount of data, entailing limits on the size of such objects.
• Sufficiently large objects can use arbitrarily more data, but such a thing would happen at certain thresholds (such as every power of 10 increase in a data point), and there would be no difference until those thresholds are met.
• There's multiple different object types, for different orders of object sizes, such that some need less data, but all objects in the same range would occupy the same amount of data.

At no point would every change in size result in a change of data usage. And even in the second case, where such changes occur most often, a lot of their data would end up not being resizable like that. On top of that, this increased data occupancy wouldn't scale at the same rate as the underlying thing it's representing. Being able to represent a mass that's twice as large does not even require the mass variable to occupy twice as much data; it barely needs to grow at all.

Thing is, the simulations you describe safe on complexity by doing a limited simulation. If you simulate planets to model their orbits you get away with density, radius, center of mass and stuff to get their orbits. However, in such a simulation you can't interact with the planets in most ways. They only know interaction by each other's gravitational field.

As soon as you include the capacity to break a rock, you need more all-encompassing schemes, like finite elements and finite differences. There the amount of data of the whole simulation is determined by the size of the space (unless you do adaptive refinement, which would be a good idea) and objects that occupy large parts occupy more of the data points in space, hence having more data.

 

[Agnaa]

If you're fully simulating all of the physics, sure, but if you're just representing it (as is the case with most digital media, such as games and videos), or if the simulation is limited (as is the case with many fictional examples, such as DDLC and Metaware High School Demo), that stops being true.

And even such full simulations would, realistically, typically operate as partial simulations until those minute details become relevant.

 

[IdiosyncraticLawyer]

Bump.

 

[Antvasima]

I personally do not mind @Agnaa 's suggestion here, but it depends on what @DontTalkDT think as well.

 

[IdiosyncraticLawyer]

Bump.

 

[Antvasima]

I personally do not mind @Agnaa 's suggestion here, but it depends on what @DontTalkDT think as well.

@Agnaa @DontTalkDT

 

[Agnaa]

Bump, what did you end up thinking of my latest post? @DontTalkDT

 

[Agnaa]

Bump, what did you end up thinking of my latest post? @DontTalkDT

@DontTalkDT

 

[DontTalkDT]

Bump, what did you end up thinking of my latest post? @DontTalkDT

Which formulation do you have in mind?

That I don't know what exactly you were suggesting at this point. Like, are you arguing that we should have different defaults on whether these media are R>F or just the physical phenomena?
Why is the amount of data relevant again? Like, how much data grows doesn't seem like a key criteria on whether it's abstraction.

 

[Agnaa]

That I don't know what exactly you were suggesting at this point. Like, are you arguing that we should have different defaults on whether these media are R>F or just the physical phenomena?

Full simulations including physics and the like should continue to be tiered at 10-C.

Partial simulations that use abstractions should be tiered at 11-C, unless there's a contradiction to it being true R>F, in which case it should be tiered at 11-A.

Why is the amount of data relevant again? Like, how much data grows doesn't seem like a key criteria on whether it's abstraction.

If the amount of data doesn't grow accordingly, that necessarily means that there's abstractions of some kind.

We're trying to put these cases into the two larger buckets of "More like each atom being its own universe" and "More like words on a page representing lower worlds".

 

[DontTalkDT]

Full simulations including physics and the like should continue to be tiered at 10-C.

Partial simulations that use abstractions should be tiered at 11-C, unless there's a contradiction to it being true R>F, in which case it should be tiered at 11-A.

If the amount of data doesn't grow accordingly, that necessarily means that there's abstractions of some kind.

We're trying to put these cases into the two larger buckets of "More like each atom being its own universe" and "More like words on a page representing lower worlds".

I see. In that case, I'm of the opinion that the entire data amount thing isn't the essential distinction.

To start with an analogy: Say there is a sentient stickman drawn with chalk on a blackboard. Which tier is that?
If it's a sentient being made from the chalk it's 10-C. If it's a being living in some 2D world on the chalkboard it's 11-A. If it's just a representation of an imaginary stickman being, then it's 11-C.

With data it's similar.

If you have a game being, the details of the game isn't essential IMO. What is essential is in which sense these beings exist.
If they exist as physical data, i.e. magnetic charges on a computer disk or similar, that's 10-C. It wouldn't matter in that regard whether it's a detailed simulation of all physics or a super simple VN. That would only change the amount of magnetic charges (i.e. higher or lower 10-C)
The 11-A case would here either happen, as you said, in the 11-C case where contradiction is present or in some rare cases where characters are the 2D beings depicted on screen.
The 11-C case would be if the characters are not embodied by the physical data, but by the ideas/information they represent. The ideas could be incredibly detailed, in which case there is a lot of corresponding data to be recorded or it could be very low detail.

 

[Agnaa]

I mostly agree, but I think my point about data is a natural consequence of those buckets. Usage and scaling of data indicates whether a character is literally embodied in that data, or whether that data represents them.

And, I'm worried that pages don't currently operate like that, and that they hand out 10-C overzealously, but the only example I can dig up right now is this one but SCP's not on-site anymore so meh.

 

[Antvasima]

@DontTalkDT

Can you help us reach a conclusion here please?