For example, this submission written in APL (Dyalog) claims to be 39 bytes. It is 39 characters, but in the TIO link provided in the post (which happens to be an older version that is two characters longer but that does not matter) it says it is 41 characters and 87 bytes. Why does the score count as [number of characters] bytes instead of 87 bytes?

This post is just one example I found but it seems to be prevalent. This doesn't seem fair to the non golf specific languages.

  • 1
    \$\begingroup\$ Because the code for those esolangs isn't encoded in UTF-8. Many have their own encoding. If you viewed them as UTF-8 characters, you'd get a lot of invalid and unprintable characters which are unpleasant to look at. \$\endgroup\$ – Pavel Nov 8 '17 at 23:30
  • \$\begingroup\$ this post answers your question generally \$\endgroup\$ – Giuseppe Nov 8 '17 at 23:30
  • 2
    \$\begingroup\$ APL is is very much so a production language, not an esolang. \$\endgroup\$ – Adám Nov 8 '17 at 23:32
  • 1
    \$\begingroup\$ This entry (mine) is prefixed by f← which assigns it to f so it can be demonstrated. it can be executed without (anon function), so valid by ppcg consensus. Furthermore, APL supports custom code page (like other code golf langs, though it isn't a code golfing language), so although the tio uses utf8 because of technicalities, this could could be saved and executed using 39 bytes when encoded in APL codepage \$\endgroup\$ – Uriel Nov 8 '17 at 23:33
  • 1
    \$\begingroup\$ @Adám It also predates Unicode, so there was a time where counting those characters as more than one byte made absolutely no sense. \$\endgroup\$ – Pavel Nov 8 '17 at 23:33

Browse other questions tagged .