Code bowling using unique characters

The aim of the challenge is to write as long a source code as possible (code bowling), every character typed gives 1 point, using the following rules:

• No symbol typed in the source may be re-used twice. That is, only one a, only one 1, only one + and so on.
• Whitespace characters do not count and may be used multiple times (space, line feed, cr, tab etc).
• Other non-printable characters may be present but do not count.
• A standard American keyboard layout/locale is assumed, see wikipedia.
• No symbols may be present inside string literals. Empty string literals are accepted (if you can manage to type one without using " twice).
• No comments may be present.
• The program must compile for the given language and run without crashing.
• No special behavior or input/output is required.

Looking for feedback:

• Is this challenge at all feasible? Too restrictive in too many languages?
• Should special exceptions be made for very common symbols like () and ;? As in allowing them but they do not count towards the total score?
• Should function-only solutions be allowed? Or maybe it should only ask for function-only solutions?

EDIT:

I suppose the only allowed characters should be the old school 7 bit ASCII, not counting whitespace characters):

 !"#$%&'()*+,-./0123456789:;=>?
@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\]^_
`abcdefghijklmnopqrstuvwxyz{|}~

Meaning that the highest possible score is 93.

Example solution (purposely naive) in the C language:

main()
{
  double cfghjkpqrstvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ = !-1234/5+6*78|90;
}

Score: 74.

Simple test tool, paste source below "Code" in TIO then check output.

(Hopefully bug-free, proof-reading is welcomed!)

posted over 2 years ago

CC BY-SA 4.0

2y ago

Lundin‭

1174 reputation 17 44 137 70