Re: Industry Thai Cell-Clustering Rules

[Theppitak Karoonboonayanan <thep links nectec or th>]

On Wed, Nov 08, 2000 at 11:14:28AM -0800, Chookij Vanatham wrote:
> 
> Hi K.Theppitak,
> 
> ] Date: Thu, 09 Nov 2000 01:27:10 +0700
> ] From: Theppitak Karoonboonayanan <thep links nectec or th>
> ] 
> ] However, now I'm thinking how the Thai shaping method can be extended to
> ] other languages, such as Lao, as planned. Do you think it's generalized
> ] enough?
> For Lao, I think it's generalized enough.

I don't think so, while the shaping rules are still in C.

> ] I'm thinking of TeX's "virtual font" technique. TeX virtual font can store
> ] LIGKERN rules for composing ligatures. And it has been proved applicable
> ] to Thai shaping (without WTT clustering).
> Not quite sure if I understand this technique enough but, just make sure
> that cell-clustering shouldn't depend on this shaping "virtual font" technique.
> Right ? :)

That's right. The clustering mechanism is independent of font.
And the WTT-like table can be defined for, say, Lao.
But how about the shaping rule? Does it have to be in C?

Shaping rule for Mac/Windows extensions can indeed be described as ligatures.
I can show you an implementation of Thai TeX virtual font which enables
LaTeX to typeset quality Thai documents without touching LaTeX code.

I'm not saying the shaping can really be generalized as such, nor is C hard
coding wrong. Just trying to explore if there is a feasible generic way.

-Theppitak.