From: Peter_Constable@sil.org

Sent: Tuesday, April 17, 2001 2:43 PM


Subject: comments on N2242



>>When you say "the Dai scripts", exactly what are you referring to?


>New Tai Lue and Tai Le. But in my view, we are not going to send
>these out to ballot if the Chinese don't play by the rules and get
>(for instance) me and Ken to sign off on these. Two meetings now they
>haven't turned up for. Expensive, from my point of view. Grrr.


>I've indicated to Mr Chen that we need to see some actual lexica to
>ensure ordering info and other things. We'll see if they respond.


I haven't responded to Ken's earlier message yet because I haven't had a chance, but I'll just say quickly that N2242 is definitely a lot better than any of the earlier proposals I saw, but still has some issues. There's a big improvement in getting rid of things that were not actually used in NTL, so the inventory is very nearly right.



Here are the outstanding issues I see:


1. The ordering is way off, though - that's the biggest problem. I can

provide scans of a NTL dictionary that shows the proper order. (The order

should be more typical of Brahmi-derived scripts, based on point and then

manner of articulation, not based first on tone class.)


2. I don't see a reason to decompose the hat off the high class b, d and glottal. It's not done for the high class sonorants, and it shouldn't be done here. This is not a productive combining mark. These are three very specific combinations. They should be treated as atomic. (For the high class sonorants, the reasoning given has to do with rendering issues, which is the wrong reason. I agree that the high class sononrants should not be decomposed, but for other reasons: this is largely a simple script - certainly relative to Lanna from which it is descended - and these things are simply a set of atomic symbols in this script. I would be willing to consider a proposal to decompose the high class sonorants, but it makes very little sense to me in the case of b/d/glottal.)


3. That proposal discusses the possibility of decomposing the low class consonants that have the upward LTR arch (e.g. xx95, xx98), and dismisses that again for reasons related to rendering. It's the right decision but the wrong reason. These should no more be decomposed than the dot on Latin i should be - and for the same reasons. These are merely atomic characters, not a case of a productive modifier or a digraph.


4. The issue of subjoining the labial semivowel is not adequately discussed. This proposal makes it sound like it should merely be treated as obligatory ligation. E.g. < ka, wa > --> glyphKWA. The pros and cons of alternative approaches, particularly using a virama, should at least be considered. (I'm not disagreeing with what is proposed here; I'm just saying it needs more discussion.)


5. Another alternative for the 7 final forms would be better to use non-finals together with a virama. (This would mean adding a virama, of course, but that may also be useful for the conjoined labial.) This alternative at least needs to be discussed. The only alternative considered in the proposal is a combining mark, and again, the main argument for ruling that out has to do with rendering, which shouldn't be the basis for the argument.


6. There is no character for the front mid open vowel. I'm guessing that this proposal assumes it can be represented as an e-e digraph. That should at least have been explained. I'd suggest that a separate character should be proposed for this, however, to make this more consistent with other related scripts, such as Thai and Lao.


7. The low class ta is duplicated (xx8e, xx9a), but the high class fa is missing. The glyph at xx8e needs to be corrected.

(See attached GIF for correct shape <01170-NTL_HiFa.gif>.)


8. The glyph for digit for is Burmese; I have not seen this used in NTL, though there is a probable historical connection. The shape that should be used in NTL is close to what is used in Lanna .

(See attached GIF <01170-NTL_dig4.gif>.)


9. A couple of quibbles: I think there could be some improvements in the history. Also, here a very nice New Tai Lue font available that could have been used.




- Peter




Peter Constable


Non-Roman Script Initiative, SIL International

7500 W. Camp Wisdom Rd., Dallas, TX 75236, USA

Tel: +1 972 708 7485

E-mail: <peter_constable@sil.org>


(See attached file: NTL_dig4.gif)(See attached file: NTL_HiFa.gif)