docs: Tweaks to Your First Model

[swo]

• DRY: use #rustdoc_include where possible
• Consistency in the population size constant
• Move toward using "entity" over "agent" -- This was not an exhaustive effort; just a target of opportunity thing.
• Make the "this isn't SIR" explanation a bit clearer and more visible
• Clarify the model setup
  • Explain how the model as initially described (each person has an independent time to infection) maps onto the implementation here (there are infection attempts)
  • An alternative would be to pre-schedule all the infections, as per docs: Pre-schedule all infections #788, although I suspect that's not great for this first example?
  • Draw inter-infection-attempt times from Exp::new(FORCE_OF_INFECTION * POPULATION). The existing approach definitely works: if $X \sim \mathrm{Exp}(\lambda)$, then $\frac{X}{a} \sim \mathrm{Exp}(a \lambda)$, but this approach is a little more intuitive and it avoids the precision loss thing.
• Gave some hints in the exercises section, and made the 3rd exercise align better with the foodborne illness theme.
• Ignore incidence.csv produced by first_model

[RobertJacobsonCDC]

• Move toward using "entity" over "agent" -- This was not an exhaustive effort; just a target of opportunity thing.

I defer to the experts for the language that is most appropriate to our audience here.

But my understanding is that not all entities are agents, at least as the word entity is used in Ixa. We might have the Person entity that serves the role of agent, and then we might have other entities that serve as data sources but that aren't technically agents. A Setting entity might hold geographical data, an $\alpha$ influencing infection hazard, etc.; A Region entity might hold information about region-specific interventions / policy. Right?

[RobertJacobsonCDC]

Excellent! Thank you so much!

[swo]

But my understanding is that not all entities are agents, at least as the word entity is used in Ixa. We might have the Person entity that serves the role of agent, and then we might have other entities that serve as data sources but that aren't technically agents.

Definitely agree. I've heard people mostly use a "weak" agent definition, which would encompass Person but not Region, and but some are passionate about a "strong" agent definition, which restricts "agent" to a thing that executes some kind of local decision-making. A Person who mindlessly infects and is infected would not qualify under that standard. I like "entity" because it avoids this hair-splitting.