pteromys: Wishing fluffies (fluffies)
Pteromys ([personal profile] pteromys) wrote2010-03-18 04:07 am

Sups and Sandwiches

The phrase of the night: "Analysis is about sups and sandwiches."

This post isn't about analysis, though. It's about math words. Today I want to draw your attention to the words "algebra", "field", and "exact".

First, "algebra". In the most general sense an algebra just refers to a set with some operations. Here are two specific cases that could otherwise seem pretty unrelated:

  1. An algebra over a ring is something like a vector space with a multiplication law and with scalars taken from the ring.
  2. An algebra over a set is a collection of subsets that is closed under complements, finite intersections, and finite unions.

The confusing part is how these connect to the word "field":

  1. If an algebra over a ring is a field, its multiplication is commutative, and it also has division.
  2. If an algebra over a set is a Borel field, it has countable unions and arises from some topology on the underlying set.

Meanwhile, a field of sets seems to be merely another way to refer to an algebra over a set, and the notion of a vector field appears completely unrelated. Then again it seems that all of these usages of "field" are mostly unrelated to each other.

Moving on, I did manage to relate usages of "exact"... and I wonder how much was intentional and how much was a fortunate accident:

  1. The exact derivative of a function F(x(t),y(t)) is written as:
    dF/dt = ∂F/∂xdx/dt + ∂F/∂ydy/dt
  2. An exact equation is a differential equation that, for some F(x,y), can be written as:
    ∂F/∂x dx + ∂F/∂y dy = 0
    Equivalently, P dx + Q dy = 0 is exact if the vector (P,Q) is the grad (gradient) of some function F.
  3. An exact differential form is something vaguely like a vector or scalar field that is the grad, curl, or div of something.
  4. Finally, recall the following identities from multivariable calculus:
    • F = grad G implies curl F = 0
    • F = curl G implies div F = 0
    In spaces where the converses hold, the grad, curl, and div operations give you what is called an exact sequence.

[identity profile] 2010-03-18 03:17 pm (UTC)(link)
There's also "normal". Oh man, is there ever "normal".

[identity profile] 2010-03-18 08:55 pm (UTC)(link)
Here, at least, you can connect "normal extension" with "normal subgroup", although "normal topological space" is different.

[identity profile] 2010-03-18 09:21 pm (UTC)(link)
Particularly amusing to me are "normal" and "normalized" as used to describe vectors--especially when both are relevant (e.g. in lighting methods for 3-D graphics).

[identity profile] 2010-03-19 12:37 am (UTC)(link)
Also, synonyms in English which are actually synonyms in math (where synonym = describing the same sketchy concept but different in the details). See "normal" and "regular" ...