Adding Hebrew and Arabic support into Windows'9x
and MSIE 4.
Hebrew and arabic input in MSIE 5
and Office 2K
This page became at last not quite obsolete,
but almost has lost the significance!
MSIE 5.01 and later versions
contains support of input for "complex scripts",
i.e. Hebrew, Arabic, Thai and Vietnamese languages.
The below-mentioned recommendations and files will help
only for adding in Windows '9x the hebrew keyboard with nekudot (vowels)
(kbdhe_n.kbd), install full bidirectional support for WordPad'98, and to look into some details of multilanguage support in Windows.
At the same time the possibility of
management of the codings through the Registry
(appeared in MSIE 4) has disappeared in MSIE 5,
now for the same purpose binary editing of MLANG.DLL is needed.
In particular, for including in the list
of the encodings Outlook Express 5
"Hebrew (ISO-Visual)" (by many desirable)
it is necessary in the file MLANG.DLL to find a byte sequence
(in version 5.00.2919.6304 from MSIE 5.01 this sequence
is at address 0x1EBA0):
B66F 0000 0000 0000 **** **** 0000 0000 0000
0000 0000 0000 1712 0000 E704 0000 0E06 0030
and to replace there (with binary editor) 2 bytes
to reach the following sequence:
B66F 0000 0000 0000 **** **** 0000 0000 0000
0000 0000 0000 1712 0000 E704 0000 0F07 0030
-
Files CP_1255.NLS, CP_1256.NLS,
KBDHE.KBD (hebrew keyboard layout from Microsoft),
KBDHE_N.KBD (hebrew keyboard layout from
Thierry Frenkel
with nekudot at ctrl-alt-1 ... ctrl-alt-=)
and KBDA*.KBD (arabic keyboard layouts from Microsoft)
should be placed in Windows\System directory.
You can download
semit.zip with these files and with files HEBREW.REG
and ARABIC.REG for Registry additions.
- Unicode fonts with hebrew and arabic characters:
You can add also entries in WIN.INI in section [FontSubstitutes]
(for some non-Unicode programs, f.e. Character Map) like so:
Tahoma (Hebrew),177=Tahoma,177
Tahoma (Arabic),178=Tahoma,178
-
Adding NLS in Registry - values 1255 and 1256 should be changed:
[HKEY_LOCAL_MACHINE\System\CurrentControlSet\control\Nls\Codepage]
"1255"="cp_1255.nls"
"1256"="cp_1256.nls"
-
For appearance encoding in Outlook Express 4 value level
should be changed to 03,03,00,00:
[HKEY_CLASSES_ROOT\MIME\Database\Codepage\1255]
"BodyCharset"="iso-8859-8"
"FixedWidthFont"="Tahoma"
"ProportionalFont"="Tahoma"
"Description"="Hebrew (Windows)"
"Level"=hex:03,03,00,00
"Encoding"=hex:01,01,00,00
"WebCharset"="windows-1255"
[HKEY_CLASSES_ROOT\MIME\Database\Codepage\1256]
"BodyCharset"="iso-8859-6"
"FixedWidthFont"="Tahoma"
"ProportionalFont"="Tahoma"
"Description"="Arabic (Windows)"
"Level"=hex:03,03,00,00
"Encoding"=hex:01,01,00,00
"WebCharset"="windows-1256"
- Adding locales (values should be changed):
[HKEY_LOCAL_MACHINE\System\CurrentControlSet\control\Nls\Locale]
"00000401"="Arabic (Saudi Arabien)"
"0000040D"="Hebrew"
- Adding hebrew keyboard layout into Registry:
[HKEY_LOCAL_MACHINE\System\CurrentControlSet\control\keyboard layouts\0000040D]
"layout file"="kbdhe.kbd"
"layout text"="Hebrew"
[HKEY_LOCAL_MACHINE\System\CurrentControlSet\control\keyboard layouts\0000080D]
"layout file"="kbdhe_n.kbd"
"layout text"="Hebrew with nekudot"
- Adding hebrew keyboard via Control Panel:
Open Control Panel and Keyboard item;
Language - Add - Hebrew;
Verify that Layout is Hebrew (all available layouts
are listed in Properties).
It is possible to select other layout Hebrew with nekudot,
which permits to enter letters with vowels (nekudot) as follows:
| Ctrl-Alt/1 |
אְ |
|
Ctrl-Alt/7 |
אֶ |
| Ctrl-Alt/2 |
אֱ |
|
Ctrl-Alt/8 |
אַ |
| Ctrl-Alt/3 |
אֲ |
|
Ctrl-Alt/9 |
אָ |
| Ctrl-Alt/4 |
אֳ |
|
Ctrl-Alt/0 |
וֹ |
| Ctrl-Alt/5 |
אִ |
|
Ctrl-Alt/- |
אֻ |
| Ctrl-Alt/6 |
אֵ |
|
Ctrl-Alt/= |
וּ |
- Bidirectionality and complex scripts support in WordPad'98.
The main edit window in WordPad, named Rich Edit Control, in
the version 3 (Riched20.dll)
has full support of a bidirectionality and other complex scripts features, allowing, in particular, to enter the text in Hebrew and Arabic
(with automatic change of a kind of the Arabic letter depending on the position in a word), and also to save it in Word 6.0, RTF and Unicode text formats. Wordpad allows to input, in particular, unicode pseudographic characters, but does not save it in any formats, except Unicode text. One more interesting possibility: small Input Metod Editor, allowing input through Unicode. It is realized by two combinations of hotkeys: Alt-Shift-X will recode a character before cursor in the Unicode hexadecimal digits; Alt-X will recode a sequence of digits before a cursor in the character.
For activation of these properties it is necessary to install the 3rd version Riched20.dll (if it was not installed earlier). Archive
rich-v3.zip is contained a set of files for Windows'98, extracted from hotfix Q249973 and may be installed with the inf-file 1200up.inf.
The similar archive
rich-v2.zip is created for restoring of the version 2, though it is possible to substitute Riched20.dll by a direct rewriting in the Windows\System directory, if not any program using it.
The described feature exists only in WordPad from Windows'98.
Older WordPad versions from Windows'95 - even launched under
Windows'98 - use Rich Edit ver.1 ( Riched32.dll ),
in which Hebreww and Arabic characters may be entered, but thers isn't
any support of the didirectionality.
It is possible to try to copy appropriate files into the
Windows'95 (the mentioned updating exists also for Windows'95,
differing only by inf-files), including new WordPad, but I haven't never
realized such experiments and don't know about the possible result.
- Support for bidirectional features and complex scripts in
Office 2000.
Office 2000, and mainly Word 2000, seems may completely support for bidirectional features and complex scripts, even under paneuropean Windows'98 (that contradicts the statements that so called enabled-versions or Windows 2000 are necessary for this purpose). I do not know, what is necessary minimum, but the following is
sufficient:
- MSIE 5.01 or later versions with hebrew or arabic support is installed
(for W2K/XP - is needed system installation of the support of semitic languages as groups and Hebrew or Arabic as concrete language).
- Riched20.dll version 3 (for Windows' 98).
- Bidirectional support is installed during Office 2000 installation.
- "Microsoft Office Language settings" installed during Office 2000 installation and then Hebrew (Arabic) is included into the list of accessible languages.
After that in section Tools - Options will appear the tab
Right-to-Left, and on a formatting toolbar - switches for
paragraph direction (from left to right and from right to left).
Generally, it is possible to find the following buttons for toolbars
(from the list of All commands), that have the relation to languages and bidirectional features:
LtrPara and
RtlPara - direction of the paragraph from left to right and on the right on the left (these buttons occur automatically);
- LtrRun and RtlRun - direction in a point of input from-left-to-right and from-right-to-left;
Language - language for fragment of the text;
RunToggle - switch of the keyboard and language (with very unclear explanations).
With this all Word 2000 has the same (and more) features for complex scripts,
as Wordpad'98, but it has also some questions and misfeatures, with which I have met for Hebrew/Yiddish. In addition to that thing, that the text with the mixed
directions of input is itself not easy for perception.
- the Standard keyboard for Hebrew (kbdhe.kbd and kbdhe.dll) contains nikudot (vowels) on the upper row of the keyboard (numbers, the tilde, minus and equal-sign), but they are entered only with CapsLock pressed and by
simultaneously pressing a key Shift. On alphabetic keys with CapsLock+Shift pressed is also entered hebrew. See interactive scheme of
Hebrew keyboard layout by Microsoft.
The best known editor of keyboard layouts - Keyboard Layout
Manager knows how to handle such feature of the keyboard (checkbox
CapsLock Alternate), but for unknown reason only for Windows 2000/XP, though keyboard has this feature also under Windows'9x.
- The attempt to place nikudot on Ctrl-Alt/Ctrl-Alt-Shift combinations often fails because the majority of such combinations is already busy in Word.
- The attempt to place hebrew letters or diacritics on uppercase letter-combinations
(with shift-key pressed) fails due to strange bug (maybe strange feature), that
appears in both systems (Windows '9x and Windows 2000/XP):
Input-direction switches from left to left-to-right, and language to the main European language of the paragraph (English or Russian), irrespective of, whether the English letter is entered (as is made on the upper case at the letter-keys in the hebrew keyboard from Microsoft) or there is placed any Hebrew letter.
And then it seems, that the text is mysteriously intermixed, and the letters are entered in any other places, instead of wanted places.
The situation can be corrected by pressing a button RtlRun, but it needs to be done for each letter, that is entered with shift.
Therefore really variants of the keyboard can be done or following Microsoft in a CapsLock Alternate mode on (CapsLock-Shift), or on free Ctrl-Alt-, and even better on AltGr-combinations.
This bug is fixed in WordXP only.
That's all. Hebrew and Arabic should be appear in Fonts menu
in MSIE 4 browser and in Outlook Express 4. At now you can also switch
keyboard and write Hebrew in Outlook Express (maybe in WWW-forms ?),
in WordPad and in Word'97. But not in Word 7 (Word'95).
But left to right only, not right to left.
I don't know how to add bidirectionality, it's seems me as impossible.
MSIE 5 can support semitic languages itself (has appropriate
language packs) but without input. And MSIE 5.01 even supports input
on them, that, as is written in the beginning,
almost deprives this page of a sense.
After adding of the Hebrew keyboard (that or different way)
occurs a possibility to write on Hebrew in Outlook Express 5,
FrontPage Express and (in Windows'98 with new Riched20.dll)
in WordPad, even
right to left and with correct
disposition of nekudot (vowels)
(in Word'97 as before only left to right
and nekudot place as usual characters).
Alternatives