Byte counting for BASIC style DATA / READ inputs

Question

Background

Many BASIC dialects idiomatically use DATA and READ commands to store and read input data from the source code. We appear to have a consensus that this form of input is valid in these languages, however we have not formed a consensus on how to score solutions that use this approach.

The Question

Given a snippet of the form shown below, how should it be scored?

Yabasic, ??? bytes

DATA 2
DATA 3
DIM primes(4)
FOR i = 1 TO 4
  READ primes(i)
  PRINT primes(i)
NEXT
DATA 5
DATA 7

Try it online!

Answer 1

All Read Bytes Should Be Scored

Snippets should be split into header, code, and footer sections, with all DATA calls in the header and/or footer. All other code will contribute to byte count.

Example

Yabasic, 68 bytes

Header (not counted)

DATA 2
DATA 3

Code

DIM primes(4)
FOR i = 1 TO 4
  READ primes(i)
  PRINT primes(i)
NEXT

Footer (not counted)

DATA 5
DATA 7

Try it online!

Answer 2

All Non-data Bytes Should Be Scored

All Non-data (that is data, not DATA) bytes, including all required DATA calls, as well as all necessary spaces, newlines, etc.

Example

Yabasic, 92 bytes

Note: all # in the below snippet represent points of user input, and do not contribute to the byte count

DATA #
DATA #
DIM primes(4)
FOR i = 1 TO 4
  READ primes(i)
  PRINT primes(i)
NEXT
DATA #
DATA #

Try it online!