Describe ScanCode output 'copyrights' type

[jdaguil]

The following is an example excerpt from ScanCode's JSON output:

"copyrights": [
  {
    "statements": [
      "Copyright (c) 1991, 1999 Free Software Foundation, Inc."
    ],
    "holders": [
      "Free Software Foundation, Inc."
    ],
    "authors": [],
    "start_line": 4,
    "end_line": 7
  },
  {
    "statements": [
      "copyrighted by the Free Software Foundation"
    ],
    "holders": [
      "Free Software Foundation"
    ],
    "authors": [],
    "start_line": 426,
    "end_line": 433
  }
],

1. Why are statements arrays instead of simple strings?
2. Why are the corresponding start_line and end_line not arrays?

[pombredanne]

1. Why are statements arrays instead of simple strings?
   Because each item corresponds to an actual distinct copyright statement

2. Why are the corresponding start_line and end_line not arrays?
   Because the copyright detection is not super precise in reporting actual lines, so we only get a range of lines and within this range we report all the statements detected...

I get this is a tad confusing and eventually we should report instead an object of sorts such as:

"copyrights": [
  {
    "statement": "Copyright (c) 1991, 1999 Free Software Foundation, Inc.",
    "holder": "Free Software Foundation, Inc.",
    "start_line": 4,
    "end_line": 4
  },
  {
    "statement": "Copyright (c) 2015 The Eclipse Foundation",
    "holder": "The Eclipse Foundation",
    "start_line": 4,
    "end_line": 4
  },
 ....
],
"authors": [
  {
    "name": "John Doe",
    "start_line": 5,
    "end_line": 5
  },
 ....
]

.....

Would this make more sense? The schema would definitely be cleaner this way!

[jdaguil]

@pombredanne Yes, this makes more sense 😄

[pombredanne]

This is tracked in #112 now