//****************************************************************************//
//*********** Time Measurement and Models - October 29th, 2019 **************//
//**************************************************************************//

- Unfortunately, Professor Biddle has to leave class early ~5:15 due to some family issues; sorry about that! We'll come back to today's paper if necessary!
--------------------------------------------------------------------------------

- So, we read a paper by Eran Tal about the "epistemology of measurement" as it relates to time, and if nothing else, it's impressive how we've been able to measure time down to the nanosecond!
    - Here, he's analyzing the philosophy involved in measuring things and what we can know about stuff we measure and the units themselves - in this case, time

- So, let's start discussing what Tal says
    - What's this "conventionalist" explanation for why physical measures remain standardized and stable, and why does Tal think it's not completely adequate?
        - Historically somewhat related to logical empiricism, conventionalists believe there are actual, natural regularities in the world, but that people then use those physical things to define some arbitrary standard (e.g. "exactly 2.5 of that thing is one Gloople")
            - The one big, pragmatic decision to be made here is correlating your unit definition with some physical thing, and then everything else is pretty objective
                - Notice here we could choose ANY physical thing for our unit (human heartbeats, atomic vibrations, sea wave pulses, etc.)
        - Tal thinks this is problematic because it isn't quite how time-measurement actually works; in the short term, clocks drift out of sync with one another, and despite defining a second in terms of cesium vibrations, we can't measure an ideal cesium clock's vibrations since it'd involve the cesium in its 0 Kelvin "ground state!"
            - So, we don't actually measure the exact standard we've chosen, like this view implies; instead, what timekeepers do is that they take the average of a bunch of clocks throughout the world and use that consensus
            - It seems there's a lot more pragmatic social considerations here than we'd initially like under this theory
    - What about the "constructivists?" What do they think, and why does Tal still think they have issues?
        - Constructivists believe that EVERYTHING about measurements are socially defined, and the standards come about from a series of social differences with the regularities eventually coming from some agreed-upon techniques
        - Tal's critique, though, is that timekeeping simply works too well for it to be purely social! We have to make minor corrections, but in the broad scope timekeeping is remarkably consistent
    - To try and synthesize these 2, Tal proposes his own "model-based" view of physical standards - what is this? Why is it preferable to the others?
        - First off, what's being modeled here is the PROCESS of measuring time, including the actual measurements, any adjustments that must be made, any processes we use for making observation or things that could go wrong with that, etc.
            - So, we need to know the "ideal state" we'd like to measure in, and then have a model including all the ways our reality differs from that ideal conception (how higher temperatures affect cesium atoms, etc.)
            - So, the models try to capture both what is ideally the case and what's actually the case
                - For instance, if an instrument gives us an incorrect result, is the instrument not functioning correctly or is our model of how it functions incorrect?
        - Obviously socially-selected factors are important in these models, but they're saved from pure socialism by their reliance on natural measurements, which play an important constraining role on the model

- Alright; for class next time, we have a few readings on modeling and machine learning and all that goodness
    - Also do try to look at the recommended paper this time, since I do want to talk about it in-class
- Alright, that's where we're headed; adios!