Add JSON form of table

[rvagg]

Closes: #305

Draft for now, just putting it up for initial feedback.

• one entry per line, but still a valid complete JSON file, so it also has [ and ] lines.
• padded to make the columns line up
• "description" column is optional and is dropped where blank
• added "intcode" and "varint" (hex bytes) fields
• I also added an optional "ref" column with only the CID ones filled in so far .. I don't think left-padding this column is going to be ideal since "description" can get quite large. An option here is to add a compound object to combine "description" and "ref" and only left-pad the top-level items, "status": "permanent", "notes": { "description": "CIDv1", "ref": [ "https://github.com/multiformats/cid" ] },

What I think should be done to make this complete:

• [ ] Add a script that will compile JSON to CSV, correctly formatted
• [ ] Add a script that will load the CSV and write any updates back into the JSON - which would make for an easier workflow to add new entries where you don't want to have to deal with formatting or adding the intcode or varint entries
• [ ] Add a script that will lint both JSON and CSV and make sure they match, which we can also add to CI
• [ ] A Makefile that will run these scripts easily for any contributor and CI
• [ ] Docs to describe how to consume the data and how to contribute new entries

[rvagg]

Link to file as you would view it on GitHub: https://github.com/multiformats/multicodec/blob/rvagg/json/table.json

[rvagg]

Another alternative is to nest the 3 code values into a nested object: "code": { "hex": "0x78", "int": 120, "varint": "0x78" }. Any nested object doesn't get the padding treatment.

[vmx]

Either layout is fine, I'd KISS and just have non-nested values. This also makes it easier if you e.g. write your own tooling and you want to convert it quickly to CSV.