undefined | Better HN

0 pointsschiffern3y ago0 comments

>Obviously, to size would be less scary.

I was expecting this, since it's probably the most common criticism of this type of visualization.

Problem is, that analysis only looks at half the dataviz fidelity coin. It recognizes the (unavoidable) loss of fidelity in size, but it ignores the (also unavoidable) loss of fidelity in time.

In the real world these objects do not remain orbiting Earth for a few minutes (ie the interval you're likely to look at this visualization). Instead, objects above ~800 km remain in Earth orbit for hundreds to thousands of years.[1]

Mathematically this second inaccuracy tends to cancel out the first inaccuracy, therefore (presumably) making this "a lot more scary/impressive."

Strangely I always see the criticism about size, but I never see the countervailing criticism about time. I suppose it's just like you said: the size factor is "obvious", but the time factor much less so.

Eta: For some great visualizations of the space debris problem, I can heartily recommend this (slightly older) ESA video.[2] It's basically a full-length documentary crammed into 15 minutes.

[1] https://commons.wikimedia.org/wiki/File:Orbital_Debris_Lifet...

[2] https://www.youtube.com/watch?t=13s&v=9cd0-4qOvb0

0 comments

5 comments · 2 top-level

WastingMyTime893y ago· 3 in thread

> I was expecting this, since it's probably the most common criticism of this type of visualization. Problem is, that analysis only looks at half the dataviz fidelity coin. It recognizes the (unavoidable) loss of fidelity in size, but it ignores the (also unavoidable) loss of fidelity in time.

It’s a real time visualisation. There is no loss of fidelity in time.

Noticing the important distortion in size is legitimate. It’s not really a criticism by the way. It’s simply that the impression of fullness inherent to this visualisation is misleading. Space is obviously mostly empty.

> Instead, objects above ~800 km remain in Earth orbit for hundreds to thousands of years.[1]

And satellites bellow 600km are only there for a couple of years and those bellow 500km a year top. Let’s not forget that area scales with the square of radius.

> Mathematically this second inaccuracy tends to cancel out the first inaccuracy, therefore (presumably) making this "a lot more scary/impressive."

I’m guessing you mean we have to take into account the fact these objects orbit a long time when considering collisions but that’s a separate issue entirely. The two don’t cancel out at all mathematically in any meaningful way.

Don’t get me wrong, I’m not saying debris are not an issue. The new deorbiting rules are definitely a good thing.

schiffernOP3y ago

>It’s a real time visualisation. There is no loss of fidelity in time.

Unless you're going to spend multiple lifetimes watching it in "real time," there is unavoidable loss of fidelity here.

I don't mean fidelity in rate-of-time, but in duration-of-time. The total time available limits the duration fidelity, just as our eyeballs and screens limit the size fidelity.

Again it's less obvious (hence this confusion), but it's no less unavoidable.

>Noticing the important distortion in size is legitimate. It’s not really a criticism by the way. It’s simply... misleading.

Extrapolating real-time events into long stretches of time is also demonstrably misleading to humans. See: the history of scientific discoveries in geology.

Your point about "legitimacy" is right on. In data visualization the goal is finding the most useful (least misleading) transform of the data, not raw fidelity.

It's just that, for the purposes of lifetime collision probability estimation, rate-of-time fidelity is more misleading than duration-of-time fidelity (since you can't have both!).

>the fact these objects orbit a long time when considering collisions

Bingo. The "scariness" comes from (where else?) the collision probability, and our estimate thereof.

>The two don’t cancel out at all mathematically in any meaningful way.

I don't claim perfect cancellation with nothing left, just that it "tends" to cancel out (ie it pushes in the opposite direction), and that this factor was being ignored.

WastingMyTime893y ago

> Your point about "legitimacy" is right on. In data visualization the goal is finding the most useful (least misleading) transform of the data, not raw fidelity.

I made no point about legitimacy. The fact remains. It is a real time visualisation. It’s interesting to consider what’s currently flying above. It’s not a collision estimation tool.

> Bingo. The "scariness" comes from (where else?) the collision probability, and our estimate thereof.

There is nothing scary about the probability of collision however. Even when you take the very large safety margin the monitoring organisations like to take probability is very low.

1 more reply

tshaddox3y ago

> Unless you're going to spend multiple lifetimes watching it in "real time," there is unavoidable loss of fidelity here.

I don’t understand this argument at all. As stated previously, it’s a real time visualization, thus there is trivially no loss of fidelity in time. To claim that is like claiming that a live web cam of a city street has a loss in fidelity because it doesn’t show what the street will look like in 100 years.

1 more reply

rlt3y ago

Maybe I’m optimistic, but I believe within a decade full and rapid reusability of launch vehicles, i.e. Starship (if successful), will radically change what is economical in space, to the point of collecting/deorbiting space debris.

j / k navigate · click thread line to collapse