On Tue, Jul 19, 2011 at 11:39:18AM -0600, Idris Samawi Hamid ادريس سماوي حامد wrote: > On Tue, 19 Jul 2011 11:32:27 -0600, Khaled Hosny > <[email protected]> wrote: > > >The algorithm actually has a dedicated rule of this use case: > > http://www.unicode.org/reports/tr9/#Higher-Level_Protocols > > Sure. At the same time, these are exactly the kinds of rules we > _don't_ want to deify.
I'm not deifying anything, but it is unfair to criticise the algorithm for not considering certain use cases while it is actually considering them! > The people who write these things are not typographers. The bidi > algorithm mixes text-editing needs with typesetting needs. I haven't seen any evidence of this, no I see how this would be a bad thing. Unicode BiDi algorithm have been implemented and deployed by millions of users world wide, and its behaviour is what most people expect, drifting from it too much is not a wise idea IMO and I can see it making more problems than it solves. > We have to examine bidi from a typography perspective, then integrate > the parts of the algorithm that make sense in that context (no pun > intended) at the lua-extension level. Well, every time the BiDi issue is brought up, this examination (that will fix some yet to be found problems) get mentioned, yet there are no signs it is going to happen anytime soon. IMO an, allegedly, suboptimal solution is better than a perfect one that will never happen. Regards, Khaled -- Khaled Hosny Egyptian Arab
