More analogies

The model kit is one way to picture a Knowledge Package. It is not the only one, and a different picture may land better for you.

Each analogy below shows something the others do not:

Analogy	What it shows
Model kit (from the overview)	how the parts fit together
Recipe	that someone else gets your result
Lunchbox	that nothing is missing

In the analogies, we show first a ludic and high-level example of “Knowledge Package” and “Knowledge Resources”, and then we evolve to the association with Earth Observation elements.

The recipe

Take spaghetti carbonara. A handful of ingredients, about ten minutes, and famously easy to ruin step.

Now hand it to someone who has never cooked it. Writing carbonara on a card tells them nothing. Five things have to travel with the dish.

1. The ingredients. The actual list, with amounts and the specific kind:

• spaghetti
• guanciale (not bacon)
• pecorino romano (not parmesan)
• eggs
• black pepper

2. The method. The steps in order, with the timings.

3. A photo of what the plate looks like when it works.

4. The warning. Take the pan off the heat before the eggs go in, or you get scrambled eggs.

5. A short video of that one moment, because it is easier watched than read.

Here is the part worth being precise about. Each of those five is a Knowledge Resource. The ingredient list is one, the method is one, the photo is one, the warning is one, the video is one. Everything held together as a single shareable thing is the Knowledge Package, and the carbonara itself is the application.

      flowchart LR
subgraph R["Knowledge Resources"]
  I["Ingredients<br/>(spaghetti, guanciale,<br/>pecorino, eggs, pepper)"]
  M["Method<br/>(the steps, in order)"]
  P2["Photo of the finished plate"]
  W["The warning<br/>(pan off the heat)"]
  V["Video of the tricky part"]
end
I --> K["Knowledge Package<br/>(the carbonara)"]
M --> K
P2 --> K
W --> K
V --> K
K --> O["Another kitchen"]
O --> D["The same dish"]
    

Click to expand

Five Knowledge Resources make up one Knowledge Package. Because the package is complete, the dish travels to a kitchen you have never been in.

Notice which resource people leave out. It is never the ingredients. It is the warning, and that is the one standing between carbonara and scrambled eggs.

A recipe that names the dish but not the ingredients is not a recipe. It is a menu.

From carbonara to Earth Observation

Swap the dish for your application. Nothing else about the picture changes:

Resource in the recipe	Resource in your Knowledge Package
The ingredient list	Your Earth Observation datasets
The method	Your processing scripts and workflow
The photo of the finished plate	Your results
The warning about the heat	The notes and documentation
The video of the tricky part	Training material and user stories

The scrambled-eggs problem carries over too. A package can hold every dataset and every script and still fail, because the one paragraph explaining how the data was produced was never written down.

The lunchbox

Now a lunchbox, packed the night before for someone’s school day.

One box, for one day, for one person. Whoever opens it is somewhere else, hours later, and cannot phone you to ask a question. All they get is what you put in. So you pack five things:

• The soup, in a sealed flask
• The spoon, because soup without one is not lunch
• The ice pack, so what you packed is still good at midday
• The allergy card, so the school knows what is in it
• The note saying which half to eat first

Each of those five is a Knowledge Resource. The lunchbox is the Knowledge Package.

      flowchart LR
subgraph R["Knowledge Resources"]
  S1["The soup"]
  SP["The spoon"]
  IC["The ice pack"]
  AC["The allergy card"]
  N["The note about<br/>what to eat first"]
end
S1 --> B["Knowledge Package<br/>(the lunchbox)"]
SP --> B
IC --> B
AC --> B
N --> B
    

Click to expand

The same shape again: separate resources, packed together as one package for the person who opens it.

This analogy is not about how the parts fit together. It is about whether anything is missing.

The forgotten spoon

Soup with no spoon is not a small problem. It is lunch nobody can eat, and every other resource in the box was packed perfectly. One missing piece is enough to waste all the others.

From lunchbox to Earth Observation

Resource in the lunchbox	Resource in your Knowledge Package
The soup	The dataset
The spoon	The script or tool that reads it
The ice pack	The computational environment, so it still runs when opened
The allergy card	The licence, saying what people may do with it
The note about what to eat first	The documentation

Every one of those has a forgotten-spoon version. A dataset in a format nobody can open. A script with no list of what it needs to run. Results with no licence, so nobody is sure they are allowed to use them.

None of it is hard to include. It is just easy to forget, because you already know it and the person opening the box does not.

Next steps

1) Creating a Knowledge Package: the practical walkthrough, once the concept is clear.

2) Concepts & lifecycles: the same two units at field level, for working against the REST API. Developer material, and you do not need it to use the Hub.