Re-thinking Spreadsheet Format

_This issue will remain open until most of these ideas are realised, and can be used as a reference later for a vision of what this tool should do._

I slapped together relviz with the following format, expressed in CSV:

**Sheet 1** (Name labels the edge)

```csv
User,Organisation
w0rp,Dense Analysis
hsanson,Dense Analysis
```

The headings name the type of vertex, and the rows describe an edge in one direction. There are problems with this format.

1. This format is highly specific to Excel. There's no way to also support CSV.
3. It's difficult to draw unnamed edges.
4. There's no way to create orphaned vertices.
5. The format is not extensible. (such as defining colours for vertices, etc.)

Given that relviz is still alpha or beta quality and just about literally no one knows it exists yet, we can do better. We can switch to the following format, and abandon the current one.

```csv
Node,Type
w0rp,User
hsanson,User
Dense Analysis,Organisation

Edge,From,To
member,w0rp,Dense Analysis
member,hsanson,Dense Analysis
```

We can accept the following extension later for defining Edge information.

```csv
Node,Type
w0rp,User
hsanson,User
Dense Analysis,Organisation

Edge,Color,Label
member,blue,member of

Edge,From,To
member,w0rp,Dense Analysis
member,hsanson,Dense Analysis
```

The edge data can be optional, and if edge data is not supplied we can say that the edge ID is equal to the edge label. We can allow for any number of optional additional columns for the node and edge rows to specify properties of them. In addition, we can later support settings for the graph itself:

```csv
Graph,Renderer
,Force-directed
```

We will support these groupings of CSV data in any given order. We will apply fuzzy-matching to headings so alternative spellings will work, potentially later including foreign languages.