Improve Serialization page

[yolile]

Currently, the serialization page includes a big section about the "Simplified single table" option. However, that is not the best option for most OCDS datasets; the "Multi-table" option is preferred. That should be even clearer in the documentation to avoid confusion and maybe entirely remove the discussion of a fully flat table from the docs.

[duncandewhurst]

In case it's helpful, we recently developed more detailed documentation of Flatten Tool's CSV format for the Open Fibre Data Standard: https://open-fibre-data-standard.readthedocs.io/en/latest/reference/publication_formats/csv.html

[yolile]

@Ravf95 suggested that maybe as part of this, we can reference https://www.rfc-editor.org/rfc/rfc4180 to indicate things like the CSV files must be comma separated.

[jpmckinney]

Good suggestion!

[jpmckinney]

Merging #1058 into this:

Integrate blog post into serialization guidance

https://www.open-contracting.org/2016/08/08/open-contracting-data-structure-spreadsheets/

Let's have a dedicated tabular formats page, as the serialization page will otherwise be quite long. Once that's done, we'll have to decide what to do with the remaining JSON content. Right now it has some "getting started" content (what is JSON) and some data-use content (Open Refine, etc.), both of which might fit better elsewhere. The last part of the page repeating the need for packages should already be covered elsewhere.

Duncan responded:

The tabular formats page should include clearer guidance and examples on representing arrays (using the /0/, /1/ etc. array syntax in column headers) and on constructing multi-table representations (including ocid and id in all sheets, including the id field of the parent object for properties of objects in arrays)