[Freeciv-Dev] Re: New nation: Croatians

[gs234@xxxxxxxxx (Gaute (B) Strokkenes)]

Gerhard Killesreiter <killesreiter@xxxxxxxxxxxxxxxxxxxxxx> writes:

> * "Dmitriy" == Dmitriy Genzel schrieb am Thu, 17 Aug 2000 13:57:14 -0400 
> (EDT):
> > Actually, I'd like to report that this is a problem. When using
> > cyrillic fonts I can only see cyrillic characters instead of
> > appropriate accented latin characters. That's a problem for instance
> > in Polish nationset. Should it be fixed to use 7-bit only?
> 
> No. If there is such a problem, freeciv needs to be fixed. IMHO
> cities and names should be used in their proper spelling including all 
> »funny characters«. I admit it might be difficult to tell X to use two
> different font encodings, say iso-8859-1 und -2, concurrently.

Actually, I have been thinking about this sort of stuff for a while,
mainly in connection with my i18n efforts.  There are a couple of
aspects to this:

* The locales of a given server and a given client may be different.
  At best, this results in a confusing multilingual mix-match of
  messages, at worst it results in totally garbled output if the
  different locales use different coding systems.

  There are a number of possible fixes to this problem, including: 1)
  Standardising on some common coding system, such as UTF-8, 2) Not
  transmitting text over the wire (i.e. using message ids and
  whathaveyougot) or 3) enabling the server to use different locales
  for different connections.

  No. 1 is tricky because it's hard to shoe-horn every locale out
  there into UTF-8, and doesn't solve the tower of Babel problem.
  No. 2 is a not very pleasant to maintain.  Personally, I prefer
  no. 3, which shouldn't be that tricky to implement.

* Names of e.g. cities and units need to make an appearance in all
  locales.  Unless we restrict ourselves to ASCII, this will cause
  difficulties.  We could try to gettextize all of the nation ruleset
  files, but this has various drawback: it makes for a lot more work
  for translators, and does not cover the case where someone types in
  a name of their own choosing.

  I suggest that all such names 'on the wire' should be in UTF-8, and
  that the requirement for uniqueness should be made more strict, so
  that names are now required to be unique up to non-ASCII characters.
  In other words, names like 'Ålborg' and 'Ølborg' would no longer be
  considered unique, but 'Ålborg' and 'Ilborg' would.  That way we can
  allow non-ASCII characters without confusing those who play on
  platforms where UTF-8 support is poor or difficult to implement, or
  who do no have the prerequisite fonts installed, by using the '?'
  glyph for such characters.  If we also allow '?' in all places where
  someone might be required to type in a name (i.e. the server) we
  also avoid the problem where your keyboard does not have the keys
  that you need to type in given name.

* Chatting. (I haven't used this feature much, since I do not have the
  time or the money to play online, so I don't have much experience
  with it.)  Different clients may use different locales, which might
  lead to difficulties similar to those mentioned in my first point.
  I don't know what would be the best way to solve this.  Maybe we
  should look at the solutions that ICQ, AIM and friends use.

-- 
Big Gaute (not to be confused with LG)
Loni Anderson's hair should be LEGALIZED!!