• 3 Posts
  • 1.59K Comments
Joined 2 years ago
cake
Cake day: July 15th, 2023

help-circle












  • FuglyDuck@lemmy.worldtoScience Memes@mander.xyzSay hello to Bary
    link
    fedilink
    English
    arrow-up
    1
    arrow-down
    1
    ·
    12 days ago

    …. Are you reading what I’m saying?

    Yes. For simple, common problems. You are correct.

    But sometimes they’re not running simple problems. Sometimes, the run time on servers costs money. Sometimes, there’s no value to be gained by being any more accurate- and it increases those costs.

    Now, in those times…. Are you really going to tell me that costing your organization more money without any useful gains…. Is “way more professional”?

    Also? Don’t get me wrong, that threshold is getting and higher every year. I have more computing power in my cell phone than they used to put a man on the moon.

    None of that changes that astronomers sometimes use 1 instead of pi, and that the barycenter of Jupiter-sun orbit is close enough to say Jupiter orbits the sun.


  • Sure.

    But sometimes, the problems are complex enough that solve time becomes a concern. When they’re complex enough, you start asking “is everything these precise enough to justify that” and when the answer is “no”, then you don’t do that because runtime on networked clusters like AWS costs money.

    And when you’re talking about scales that encompass the galaxy…. Well. There’s just not a lot of precision there to begin with.


  • Not when that definition of pi goes to all 300 trillion decimals that we have resolved. (To be fair, I don’t know of any that do… but eh…yeah. And I’m pretty sure it was defined by a masochist if one did.)

    That leads to unnecessary time spent calculating even simple equations. That level of precision is almost never actually needed.

    With fermi problems, usually that level of precision is moot and potentially a waste of time. (Particularly when the math is requiring some kind network cluster to do.)


  • fermi approximations happen all the time in astronomy. The numbers are frequently so large that the only meaningful quality is how many orders of magnitude it has.

    More to the point, using pi makes calculating things much harder. For example, we don’t really need a precise distance for most things; so using “3” makes the calculation unnecessarily spend time in computation.

    It’s like the old joke, “what’s the difference between a millionaire and a billionaire?” (“About a billion.”)


  • The reason being, that once you go large enough, a multiplier of three is irrelevant, and they only really care about orders of magnitude. You might be tempted to argue that that doesn’t happen inside the solar system, and you’d be right. Mostly.

    Except that astronomy doesn’t concern itself with just our system. So yes. Astronomers do frequently round to 1 because it really doesn’t matter that much in the scheme of things. (particularly talking about distances.) it’s even more so for cosmology.