Re: Microsoft Layer for Unicode on Windows 95/98/Me systems

Sam Hobbs <samuel@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx> schrieb im Beitrag
<ujS2xNQVFHA.3244@xxxxxxxxxxxxxxxxxxxx>...
> In the context of Windows, code pages are very relevant. Most fonts
provided
> by Windows are not Unicode fonts. The one or two fonts that are Unicode
> fonts are very large and until recently disk space was costlier than it
is
> now. Therefore it was unreasonable to provide a variety of Unicode fonts.

> The non-Unicode fonts use code pages.

Sorry, but:

a) Most fonts provided by Windows (from Windows 95 on) >are< indeed
'Unicode fonts'.
b) I can not see any coherence of 'Unicode fonts', disk space, and this
thread on the whole. Like all other fonts a 'Unicode font' may be large or
small in size. Like with all TrueType or PostScript fonts this does not
depend on whether it is a 'Unicode font' or not, but on the glyphs that the
font defines.
c) Last but not least: In fact there isn't anything that could be called a
'Unicode font'. A TrueType font may hold several character encoding tables.
If the glyphs of a font may be accessed by Unicode character encodings
depends on whether the font holds a table for this.

> See: http://support.microsoft.com/default.aspx/kb/816039 (KB article
816039,
> "FIX: Code Point Comparison Semantics for SQL_Latin1_General_Cp850_BIN
> Collation"). At the botton it says "Code point is a numeric value in
Unicode
> encoding, or in a code page that corresponds to a character.".

Yeah, "or"! The Unicode character encoding does not make use of "code
pages" (unless there is someone - and presumably there is since there is
always someone - who likes to call a Unicode codepoint range a code page).
In the context of single byte character encodings "code page" defines the
relation of the font's character set to one or more human scripts. The term
"page" is used because single byte character sets provide a frame of of
only 256 characters into which all human scripts have to be mapped somehow.
Unicode on the contrary already includes all (or almost all) human scripts
hitherto known. The code point space of Unicode is subdivided into code
point ranges and not in "code pages" because the code point ranges are very
different in size. A "code page" is used to specify to which human script
glyph a single character code shall be mapped. The "code page"
specification is necessary because the respective character code may be
mapped to several different human script glyphs, i.e. the same character
code is used in different "code pages". This isn't the case with Unicode:
Each code point by its value already includes the information to which
human script it belongs and which human script glyph it is encoding.

> See: http://www.unicode.org/glossary/index.html, the "Glossary of Unicode

> Terms" in the Unicode.org web site. It has a definition of "Code Page".
> Since the definition is in the Glossary of Unicode Terms, the implication
is
> that the term is relevant to Unicode.

That's really neat: Because a term is within the Unicode glossary it has to
be relevant to Unicode. Ho, ho, ho! The glossary also includes a definition
for the term "HTML". Is "HTML" relevant to Unicode?
Of course, the term "code page" is indeed somehow relevant to Unicode: It
is relevant if texts have to be converted from character encodings where
"code pages" are in use to the Unicode character encoding and vice versa.

> See: http://cpdetector.sourceforge.net ("cpdetector, free java code page
> detection"). It has a description of code pages and their relevance to
> Unicode.

But it does not state anywhere that Unicode >makes use< of "code pages".
The only relevance is if conversion from one character encoding scheme to
the other has to be done.

--
----------------------------------------------------------------------
THORSTEN ALBERS Universität Freiburg
albers@
uni-freiburg.de
----------------------------------------------------------------------

.

Relevant Pages

  • Re: Quieter glyphs than parentheses
    ... ASCII or 16-bit Unicode characters, it did not require rewriting the entire ... by non ISO8859 language scripts. ... Japanese has three popular non-Unicode-based encodings, ... display fonts is one reason I would caution against using characters from ...
    (comp.lang.lisp)
  • Re: Indian language fonts to be included as default fonts in windo
    ... Unicode MS. ... fonts, there again you have gone wrong. ... And so we have to depend on English unicode ... about keyboard requiring more than 105 keys because I want the keyboard to ...
    (microsoft.public.word.docmanagement)
  • Re: Unicode Word 2004 and Windows.
    ... > Actually, the Mac MS Unicode fonts are Times New Roman, Verdana, Trebuchet MS, ... Arial Unicode MS has all the characters of the ... I don't know how many characters Times New Roman, Verdana and Trebuchet MS ...
    (microsoft.public.mac.office.word)
  • Re: Enhanced Unicode support for "Go" tools
    ... the point to remember is that UNICODE is a _character ... It's the fonts, the OS and the application which work together ... society for the protection of French from English ...
    (alt.lang.asm)
  • Re: Unicode support in windows mobile 5
    ... You can also buy fonts, ... If you want to display unicode on your PPC, ... Generally English devices have English characters, ... I'm woking on a projects that uses unicode char. ...
    (microsoft.public.dotnet.framework.compactframework)