To be more precise; Amphibious characters are positioned on the middle of a line drawn between _the vowel positions_ of the two characters surrounding it.
t Thomas Milo wrote: >> Hi Meor, >> >> Meor Ridzuan Meor Yahaya wrote: >>> So, basically we can categorize the problems into only 2: the >>> amphibious characters, and the assimilated tanween. You did not say >>> anything about it. What is your proposal about it? If i can recall >>> correctly, you were suggesting to treat the sequence fatha+fatha to >>> be equal to the assimilated fathatain, am I right? Is this still >>> your stand? Personally, I would like a new code point for it. >> >> Under the influence of your - earlier - approach I changed my >> position, already a while ago, in concurrence with Mete Kural. To >> preserve textual integrity and interchangeability with less >> elaborately spelled Qur'ans, I propose to maintain tanween in all >> positions, and instead to treat the added precision of the tajweed >> as mark-up. This doesn't have to be mark-up expressed by a separate >> layer but simply using existing code points more or less as >> intended. Vowel-plus-small_meem could be expressed by >> tanween-plus-small_meem. The off-set fathatan (as well as off-set >> kasratan and repeated dhamma's) could be expressed by >> tanween-plus-sukuun. This solution would allows for the more logical >> treatment of repeated vowel mark as equal to the imho fictitious - >> tanween character. After all, the tanween marks as separate Unicode >> characters can be traced to a typographical ligature of reduplicated >> vowel marks. >> >>> About the amphibious characters, I think it is not that straight >>> forward to support it. >> >> It can be done. Better still, it SHOULD be done. >> >>> Of course, if Microsoft says that they have no >>> problem supporting it, then there will be no problem. >> >> Microsoft? My foot. We are building it for Adobe CS2. That'll show >> them the way: they copy right and left. >> >>> I think , the >>> closest thing to amphibious currently available in Unicode is the >>> behaviour of Marks. Marks does not change the shaping behaviour of >>> other characters. >> >> Amphibious characters are close both to marks and to base >> characters. Marks float above or below a base character; base >> characters carry vowels that float above or below. Amphibious >> characters float BETWEEN characters and can carry their own vowels. >> That's why the are amphibious. Think of the famous example of >> /waliyiiya/, where one /y/ is written with superscript retroflex yeh >> - with its own kasra! >> >> If you hesitate to use this concept, you can hack a typographical >> solution by placing substitute mark characters above or below a >> tatweel bar. This will of course destroy the integrity of the >> encoding, because it then consists of a mix of context sensitive and >> context insensitive characters. Moreover, the problem with tatweel >> bars is, that even in computer fonts they are only acceptable >> between connecting characters. To simulate amphibious characters in >> this manner, you would need a non-printing tatweel allograph >> following disconnecting letters like waw, dal, etc. in order to get >> the positioning right for hamza, superscript alif, etc. When dealing >> with sophisticated (or simply "well-designed") and calligraphic >> script, one soon discovers that the concept of tatweel bar to space >> such a superscript/subscript character is not compatible with the >> use of >> ligatures - worst of all, of course, it doesn't work with lam-alef >> in any font (except yours and mine). >> >>> However, this is not exactly what's needed, because >>> marks, by current practice, should not be standalone. However, the >>> characters that we are trying to address can. So, what I'm trying to >>> convey is, we need to be very careful to define the behaviour of the >>> character.Also, we might need some proposal on how to tackle the >>> spacing problem (as compared to marks, which has no space). >> >> The "standalone" spacing behaviour in my view - and in addition to >> the printed images from Jeddah and Cairo I include observations from >> manuscript practice - is that amphibious characters are roughly >> positioned on the middle of a line drawn between (amphi-) the two >> (-bi) characters surrounding it. Whether these surrounding >> characters are spaced by a keshide is irrelevant in manuscript >> practice. >> >>> The heh goal, so basically you are saying that it should not be >>> there, right? So, is there any needs for me to include it in my >>> font? >> >> Not needed for Arabic. If you want to support Urdu, you could >> substitute a regular Heh to cut corners - or build a proper >> Nastaliq. Alternatively, you could design the necessary amount of >> hybrid nastaliq behaviour into your Nask-derived font, as suggested >> by the Unicode standard. >> >> Cheers, >> >> t >> >> >>> >>> On 3/1/06, Thomas Milo <[EMAIL PROTECTED]> wrote: >>>> Meor Ridzuan Meor Yahaya wrote: >>>>> Tom, >>>>> Good to hear from you. I remember that you are the one mentioning >>>>> that the small/superscript noon have only one occurrence in arabic >>>>> history, or something like that. Am I right? >>>>> Anyway, it is true that Indian/Pakistan tradition uses a small >>>>> noon to denote the noon sound for tanween followed by >>>>> sukun/shadda. However, it is small noon, not really superscript >>>>> noon. >>>> >>>> It's feasible that it can serve in both instances, just like the >>>> trailing/superscript retroflex yeh. >>>> >>>>> Ok, a very short list from me: >>>>> 1. The assimilated tanween. We need to finalize this, whether a >>>>> new code point will be added or some other encoding will be used. >>>>> 2. Good old hamza, especially the one over tatweel and lam alef. >>>>> 3. The superscript waw (the one occurance) >>>> >>>> Tatweel is not a grapheme. I have come to the conclusion, that the >>>> letter shaping mechanism needs a new catagory: amphibious >>>> (literally "between both", between skeleton and vowel as a >>>> category, as well as placed between two surrounding letters - with >>>> optional and separately encoded tatweel. This idea handles the >>>> problems with hamza U+0621, superscript alef, trailing/superscript >>>> retroflex and possibly even the superscript waw. Think of it, if >>>> you will. >>>> >>>>> The not so pressing issues: >>>>> 1. Spacing small alef and yeh (waw and hamza goes to above >>>>> category) >>>> >>>> Mechanically I consider all of them amphibious. >>>> >>>>> 2. Alef maksura, yeh, farsi yeh etc. The definition in 4.1 for 649 >>>>> is dotless yeh in any position (although the name remains as alef >>>>> maksura). So, what does this means? In the same document, they >>>>> said that 64A (yeh) + 654(hamza above) = 626 . This , to me >>>>> somehow does not really goes together. >>>> >>>> Here's a real legacy of earlier Arabic, for which dots were as >>>> optional as vowels. The fact that they are now encoded as part of >>>> the rasm makes it a permament problem, I am afraid. Unless canonic >>>> equivalence is accepted for [modern letter] = [rasm + >>>> points/stripes]. >>>> >>>>> I'll add more if I find. >>>>> >>>>> One quick question to you, why is the 6C1 looks like it's final >>>>> form? is the isolated form really looks like in the document? I >>>>> personally have no idea how it looks like. >>>> >>>> >>>> This so-called Heh Goal is a legacy blunder. This descending final >>>> shape is caused by the use of nastaliq script, not by the logic of >>>> the grapheme. Regular heh plus nastaliq would have yielded the same >>>> result. Moreover, the language that requires it, Urdu, accepts >>>> regular heh when it is the consequnec of use another calligraphic >>>> style, such as naskh. >>>> >>>> Cheers, >>>> >>> >>> >>> >>> >>>> _______________________________________________ >>>> General mailing list >>>> [email protected] >>>> http://lists.arabeyes.org/mailman/listinfo/general >> >> > > > >> _______________________________________________ >> General mailing list >> [email protected] >> http://lists.arabeyes.org/mailman/listinfo/general
_______________________________________________ General mailing list [email protected] http://lists.arabeyes.org/mailman/listinfo/general
