[ILUG] PHP plus Celtic languages

[Brian Foster]

  | From: kevin <kevin at cybercolloids.net>
  | Date: Thu, 19 Aug 2004 09:07:17 +0100
  | 
  | Yes, I specify
  |  <meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
  |  <meta http-equiv="Content-Language" content="kw" />
  | Much to my surprise w3c has a content language for Cornish kw=kernewek
  | It seems to work OK in Mozilla and Konqueror. To continue the pedantic
  | note .... what is the correct code to use for "small t with cedilla"?
  | if not &#355;

 now that I've had a chance to think about it and do a
 bit of checking, &#355; is correct no matter what the
 page's encoding (charset) is set as.  (so my previous
 posting was slightly wrong, apologies!)

 the page's charset specifies the encoding of the page's
 own _source_.  the &#<dec>; entity is simply a source
 notation for specifying an arbitrary codepoint (UCS
 character) to be displayed by the remote client browser.
 that notation is written in (in this case) UTF-8, but
 the meaning of the &#<dec>; does not change, not even
 if the page's source charset happens to be 16-bit
 Unicode (whose proper name is UTF-16) --- I've tested
 the &#355; notation in UTF-16 in FireFox and Opera.

 ignoring arcane technical/political debates, the UCS
 (Universal Character Set, ISO standard 10646) is simply
 a list of agreed character names and values.  (the names
 and values are both unique.)  t-cedilla has been assigned
 the value 163 (hex) and name “LATIN SMALL LETTER T WITH
 CEDILLA”.  UCS codepoints are typically denoted in the
 U+<hex> notation, so t-cedilla is U+0163.

 encodings specify how a (subset of) UCS values are
 stored.  UTF-8 says U+0163 is the two byte sequence
 C5 A3 (hex).  see the utf-8(7) man page for details
 on how U+0163 is encoded as C5 A3.   ISO-8859-2 says
 that _same_ character, U+0163, is the one byte FE.
 code page 852 (CP852) says U+0163 is the different
 one byte EE.   US-ASCII (ISO standard 646), ISO-8859-1,
 and ISO-8859-15 all do _not_ have any storage for
 U+0163 at all (it is out of domain).   UTF-16 says
 U+0163 is the one 16-bit word 0x0163, and in UTF-32
 it is the one 32-bit Word 0x00000163.

 so &#355; is just a notation like U+0163.  and that
 notation --- the ampersand, the hash, the digit 3,
 the two digits 5, and the semicolon --- are themselves
 characters in the page's charset.  what the client's
 browser reads is those six characters, which it then
 understands means U+0163.  it then looks up how to
 display U+0163 in the current language (kw (Cornish)),
 and presents (draws) the glyph ţ.

 easy.  ;-)

cheers!
	-blf-

p.s.  a great on-line resource for UCS names and values,
      and some notations (like &#355;) and other details,
      is the Letter Database at the Eesti Keele Instituut
      (Institute of the Estonian Language):

           http://www.eki.ee/letter

      the `gucharmap' tool (note that is spelled _with_
      a `u' (U+0075, “LATIN SMALL LETTER U”) can also
      be helpful.


  | On Wednesday 18 August 2004 22:38, Brian Foster wrote:
  |   | From: kevin <kevin at cybercolloids.net>
  |   | Date: Wed, 18 Aug 2004 11:21:50 +0100
  |   |[ ... ]
  |   | Cornish uses some accents including t-cedilla in words such as
  |   |
  |   |  conveţhaz - Verb, to understand
  |   |
  |   | I can write this using codes in UTF-8 like conve&#355;haz  [ ... ]
  | 
  |  uh, not exactly.  “&#355” does not (cannot)
  |  represent literal UTF-8 per se.  (it _is_ the
  |  UCS codepoint value for U+0163, which is
  |  “LATIN SMALL LETTER T WITH CEDILLA”, which
  |  apparently is the character you want.)
  | 
  |[... remainder of my reply was slightly confused  ;-(  -blf ...]
-- 
«How many surrealists does it take to    |  Brian Foster      Montpellier,
 change a lightbulb?  Three.  One calms  |  blf at utvinternet.ie      FRANCE
 the warthog, and two fill the bathtub   |    Stop E$$o (ExxonMobile)!
 with brightly-colored machine tools.»   |        http://www.stopesso.com