Extension to convert symbol <-> unicode

[David Mathog]

Wishlist

Inkscape 0.48+devel, Windows XP SP3

Greek letters should be written using unicode for HTML pages, and by extension, in SVG. Unfortunately there are enough older programs around on Windows that only know about Greek letters when they appear in the Symbol font that somewhere along the line Inkscape was modified to support Symbol font. In order to support those older programs we must use Symbol font, else imported drawings end up with a bunch of nothing or question marks where they should have Greek letters. On screen it is impossible to tell what is Symbol and what is, for instance, Unicode Times New Roman. But if we use Symbol, the svg is wrong for web use.

It would be immensely helpful if these were added:

  extension -> text -> symbol font to unicode
  extension -> text -> unicode to symbol font

The first of these would map characters in symbol font to unicode (but do nothing to text in other fonts). It would let the
   user choose which font to map to, perhaps defaulting to Arial or Times New Roman.
The second would map characters in any font to symbol font if there was a corresponding character in the target font. Alpha
  to alpha, beta to beta, and so forth.

Characters not in the map would not be modified. If this only worked for Arial/Times New Roman <-> Symbol that would be good enough.

With this tool we could easily convert a drawing for import into an older program, yet still keep the master, web compatible copy, with proper unicode coding for Greek letters.

On a somewhat related point - the "almost equals to" character U2248 is really a problem. These older programs (like Powerpoint 2003) use Arial MS Unicode to represent this character, and when they export to an EMF, for instance, Inkscape can read it in. However, these same programs cannot always import their own EMF files properly - all of the unicode characters outside of the 8 bit range end up mangled in one way or another, including especially the U2248 character. Yes, this is a bug with the older programs, but if anybody knows of a way around this please share!!!

[David Mathog (mathog)]

See bug 919728. Unicode<->nonUnicode conversion code in changes.tar.gz. This was applied only for EMF reads and writes.