Change requests to the table document -- pass them to me please!

[Harald Alvestrand]

Mark Davis wrote:
> You didn't look back far enough, for example: 
> http://www.alvestrand.no/pipermail/idna-update/2008-November/003021.html
>
> To that message I got no response from you, agreement or disagreement. 
> As I said in the meeting, I've also heard from others that they were 
> waiting on some of the major issues to settle before speaking up on 
> the context issues. I think we made enough progress in the meeting 
> that now is probably the time for that.
Note that I disagree with several of the points from that note (not the 
need to be precise, but I like functions like Previous() much better 
than pseudovariables like P, and I think the initial ruleset needs to 
stay in the document after publication).

I missed replying to your comments in November, but see that the 
relevant sections were revised in December, so my comments would 
probably have been moot anyway.
>
> I will add one other point from the meeting, these rules are complex 
> enough that until we get two independent implementations of the rules 
> (that agree on the results) we cannot have any confidence in them. And 
> because they are crucial to getting correct A-Labels, we need to have 
> confidence in them.
As part of those implementations, I also would like a test set; the lack 
of a test set was a major hindrance for me when attempting to 
reimplement the Unicode Bidi algorithm.

                 Harald