Breaking these apart, because they are very different topics.<div><br clear="all">Mark<br>
<br><br><div class="gmail_quote">2009/4/10 Patrik Fältström <span dir="ltr"><<a href="mailto:patrik@frobbit.se">patrik@frobbit.se</a>></span><br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
Let me just throw out some thoughts that are in my head. I just can not get rid of them.<br>
<br>
1. Stability / Storage<br>
<br>
I think it is good that we talk about M-Label, as we have been talking about A-Labels and U-Labels and we (specifically myself) *REALLY* like the fact A-Label and U-Label are defined terms. And that we have a 1:1 mapping between them. That they are stable. It makes it possible to reference those explicitly in other specifications.<br>
<br>
Now, I think as always that we always will have mappings. Applications will (and should) _ALWAYS_ try to "help" the user by trying to understand what the user want to "type". We have different keyboards, different input mechanisms etc, so mappings will exist on different abstraction layers. As Pete Resnick said some weeks ago.<br>
<br>
Anyway, I think we have to say though that what is stored, regardless of where and how it is stored MUST be the A-label/U-label. This because I think the "further away" from the A-Label/U-Label we come in the abstraction, the more divergence we will get regarding support for mappings. This is btw where I have seen many problems with IDNA2003. Some people have in fact stored what has been possible to use as "input" to the IDNA algorithm, and not the "output".<br>
<br>
Now you might say that if people did not read the IDNA2003 spec, why would they read the IDNA2008 spec, but I am not prepared on giving up due to that.</blockquote><div><br></div><div>I'm sympathetic to the goal, but the problem is what is meant by "storage": memory, APIs, modules, threads, processes, communication protocols, email bodies, IM messages, hrefs, ...? See my email of a few weeks ago. It might be better to focus on the transmitting of IDNAs</div>
<div><br></div><div>I think what we can say is programs SHOULD convert to the canonical U-Label form before transmitting to other IDNA-aware programs, and to the A-Label form before transmitting to non-IDNA-aware programs. (The above isn't formal language, but you can see what I mean.)</div>
</div></div>