Mailing List Hosted on Kabissa - Space for Change in Africa

a12n-collaboration Mailing List Archive: Re: [A12n-Collab] Re: 5 categories of African orthographies (Latin)

[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index] [Thread Index]

  • Subject: Re: [A12n-Collab] Re: 5 categories of African orthographies (Latin)
  • From: David Rowe <David_Rowe@xxxxxxx>
  • Date: Tue, 01 Jan 2008 19:29:14 +0000

The fact that the question keeps getting raised (I hear it
from others occasionally) is sign enough that there is a need to either clarify the support issues and how those are being addressed, or clarify how the system
doesn't work. Continued doubts about dynamic composition either need to be
addressed with better explanations (and real support) or alternatively - again from
a NPOV - with a real proposal that makes the justification and proposes
specific precomposed characters. This so that we can move forward one way or
another rather than recycling debates.
In order for a Unicode font to support dynamic composition, the font must contain the necessary information and there must be software that can use that information to correctly position the diacritic with the base character. Under Windows, the Uniscribe processor provides the software support necessary for an application to render combining diacritics (provided the application knows to use the Uniscribe processor). For example, Word 2003 running on Windows XP can render U+025B U+0303 U+0300 (epsilon with the tilde combining diacritic on top of the epsilon and the grave combining diacritic stacked on top of the tilde) provided that a font (such as Doulos SIL) is used which contains the information for placing these diacritics.

It is also possible to create a font that provides support for a subset of base character / diacritic combinations by including all the needed combinations as precomposed characters in the Private Use Area, and then adding substitution information (using Microsoft's VOLT) that will, for example, substitute the precomposed epsilon+tilde+grave character which I put at U+E04F when the combination U+025B U+0303 U+0300 is found in the text. I have created such fonts by using SIL's Encore Font system (originally designed for creating custom 8-bit fonts, but also capable of creating Unicode fonts) then adding the substitution information with VOLT. The Microsoft Uniscribe processor is still needed to do the substitution, so this solution should work with Word 2003 on Windows XP.

My need for a sans-serif font and a mono-spaced font was the motivation for creating fonts that support only a subset of base character / diacritic combinations. (SIL Doulos is a serif, variable width font.) When a sans-serif Unicode font that supports all the needed characters becomes available, then it can be used instead of my custom font with no change in the text of the document, just a change of font.

If the text were to contain the sequence U+025B U+0300 U+0303 (epsilon with grave and tilde), then the Doulos SIL font will render it correctly (with the tilde on top of the grave on top of the epsilon) whereas the fonts I created will not since that particular combination doesn't occur as a precomposed character (because none of the languages that this font was intended to support uses that combination).

I don't know what the situation is for Mac. For Linux, there exist already, or are in development, ways to render Unicode text which uses combining diacritics. The font needs to include the necessary information and the application and the operating system need to provide the software support for the rendering.

David Rowe
SIL Togo-Benin

[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index] [Thread Index]

Last Updated: Sat Jan 05 23:34:44 2008

a12n-collaboration is hosted on Kabissa - Space for Change in Africa

Your feedback is important. Click here to send a message to the Kabissa team.

Terms of Use | Privacy Notice | Web Site Credits © 1999-2006, Kabissa or its affiliates