Re: Indian Language in AbiWord

Subject: Re: Indian Language in AbiWord
From: Andrew Dunbar (
Date: Wed Dec 26 2001 - 04:04:32 CST

 --- Savant shanti <> wrote: > Hi
> Can we write some sort of macro on AbiWord as
> provided by the MS-Word 2000?

Well AbiWord now has Perl bindings for macros but this
is not what we want in this case.

> Actually I want to work out something of this
> sort.
> Any language fonts can be used in AbiWord, so in
> order
> to use AbiWord to type a document in some other
> language basically we need a font with its keymap.

Fonts and keymaps are two separate things. A keymap
is part of the OS (or windowing system on *nix). And
passes character codes to programs. Abi interprets
these codes based on the locale, surely Unicode in
the case of Assamese. When rendering text, strings
are passed to the OS/Windowing APIs as a sequence of
Unicode values. If we're rendering with a font that
contains these values, our string will be rendered.
If not, Abi currently displays a series of fallback

> In
> Indic Languages (of which Assamese is also one) the
> problem arises out of 'Juktashara'. 'Juktashara' is
> another character which is a combination of two or
> more characters. So what we need to do is
> 'a+b'->'x'.
> Thence can we write a simple program which looks up
> an
> array, whenever 'a+b' is typed, and replaces 'a+b'
> with 'x' in document. In this way we can type a
> document in 'Indic Languages'.
> This way is just the display rendering. I mean
> works only with a particular font and gives user a
> feeling that he is working in that particular
> language.
> Please give comments and suggest ways of doing in
> such a way if its possible.

We probably don't want to handle this at the keyboard
level. Because Indic languages have not been
traditionally used on computers there are no 8-bit
standards. This has allowed the current generation
of OSes etc to come up with a "pure" Unicode method
for handling these languages. Thanks to the wonders
of OpenType we can have strings represented by a
series of "logical" characters, leaving the OS render-
ing engine to map these into glyphs. This is how
Indic languages are being implemented and Windows 2000
has a good implementation. Doing things this way
makes searching and sorting trivial. Having glyphs
stored in the strings would be a nightmare and has
proven tricky in other languages which were
implemented this way before we had OpenType.

Search around on the internet for "complex script",
"indic script", "opentype", "freetype", "pango" and
you should be able to find more detailed explanations.

Hope this helps. Andrew Dunbar.


Do You Yahoo!?
Everything you'll ever need on one web page
from News and Sport to Email and Music Charts

This archive was generated by hypermail 2b25 : Wed Dec 26 2001 - 04:04:37 CST