Date: Mon, 25 Nov 2002 11:35:47 +0100 (CET) From: Werner LEMBERG To: linux-utf8@nl.linux.org, gsinai@yudit.org Subject: Re: A Rendering Idea Message-Id: <20021125.113547.78472655.wl@gnu.org> References: MIME-Version: 1.0 Content-Type: text/plain; charset=gb2312 Content-Transfer-Encoding: quoted-printable > The idea is: > =A2=F1. Assign codes and hot spots for all possible Glyph componenents, > per script, per language system. How will you handle open-ended scripts like Urdu where the number of ligatures is changing while the language evolves? For example, I was told by an Urdu computer scientist that during a visit of Margaret Thatcher (a former Prime Minister of England) the newspapers created a new ligature for her name. > =A2=F2. Create a generic state machine thet can step through the input > unicode characters, and spit out Glyph components and their > relative hot spot positions. This is far more complicated I fear. You will need fallback algorithms for fonts which don't provide some glyphs/ligatures, etc. Some fonts have e.g. `Amacron' as a single glyph, others compose it from `A' with a macron accent. > =A2=F5. Create a generic inverse state machine. The input is > components and their relative hot spot positions and the > output is unicode stream. You can do that already by following the Adobe Glyph List (AGL) algorithm for naming glyphs. > The merits of such a rendering/font schema would be: > - It is bitmap-font-friendly Hmm, the next release of X will probably contain all bitmapped fonts in SFNT format. It is straightforward then to provide proper OpenType tables to do the same processing as with outline glyphs. Just van Rossum's freely available TTX compiler/decompiler for OpenType fonts can help here. Werner -- Linux-UTF8: i18n of Linux on all levels Archive: http://mail.nl.linux.org/linux-utf8/