L2/03-303

Date/Time:    Mon Aug 25 06:11:13 EDT 2003
Contact:      jhi@iki.fi
Report Type:  Public Review Issue

* In the "4. Pattern syntax" there are spots like

        ... removing some characters that appeared inappropriate
            for patterns ...

        ... then some script-specific characters were removed,
            along with some other characters that appeared
            inappropriate for patterns.

  These explanations appear quite vague and without more rationale
  quite arbitrary.  Even though I believe there were good reasons
  for the removals, without knowing the details the above sounds
  a bit haphazard.

* I propose adding an "Implementation Note" or something along those lines
  which mentions one real complication of introducing Unicode to identifier
  names.  That is, in many programming languages some language identifiers
  "leak" into the filesystems namespace, both as file and directory names,
  like for example in Java the class names.

  This means that the filesystems used must be capable of storing Unicode
  somehow, either by builtin support for some Unicode encoding like UTF-8
  or UTF-16 (the latter being more problematic because of embedded zero
  bytes), or by convention (for example: filenames under certain directories
  could be known to be in UTF-8).

EOF


-- -- -- -- -- -- -- -- -- -- -- -- -- -- -- --