More semantics, in a structured and strategic way!

[pdehaye]

This diagram is a visual description of what we are doing right now (source) work in progress.

Each "land" has been built as a dataflow diagram, following the "meringue idea" of Applied Category Theory.

Have a look at this source, .

There are at least two things I have not explained this far:

• why the resource (data, raw eggs) are on the arrows in the horizontal planes/lands.
• why the arrows between lands are double arrows.

The two explanations are kind of orthogonal to each other, so it is worth delaying a bit.

[pdehaye]

Several questions came in: What do the down arrows mean? and What is the bottom land?

[pdehaye]

have added/changed slightly some labels, that should clear up those questions

[pdehaye]

is this clearer now, @alexbfree ?

[alexbfree]

Also 'drop' and 'transfer' could be clearer... and why do CSVs get imported to a db but json don't?

[pdehaye]

"drop" is "drag&drop", "transfer" is the transfer between bubbles. I know that requires work on both bubbles, but that is not what I want to highlight in this diagram (this is where maybe I am most different in my concerns than your architecture vision).

Have better suggestions?

[pdehaye]

have also updated the "db import side", but it is akward when you don't want to overlook important elements: I want the loop at the bottom to be clear and unambiguous, and "db import" is not the same if the import types are not the same.

maybe a way forward to make it less awkward is to see accessors as a way to go from JSON to CSV ahead of database import. is this what you guys are doing?

[valentinoli]

maybe a way forward to make it less awkward is to see accessors as a way to go from JSON to CSV ahead of database import. is this what you guys are doing?

yes

[pdehaye]

so you go through accessors from json to csv, and you go from csv to sql. nice. ok, time to redraw (quite substantially) this thing.

[emmanuel-hestia]

Hmmm we do not have an explicit JSON -> CSV step before SQL, it is all loaded directly from JSON to SQL. But it is easy to retrieve as CSV from the SQL, of course

[Amustache]

Related to hestiaAI/hestialabs-experiences#767