Why alpha4 codes?

[Lee Gillam]

[Actually, I wanted to make sure I'd got the intended meaning correct
before sending to the list]

> [Lee replied privately, but as the content wasn't personal and relevant
> for the thread, I'm assuming he meant to reply to the list.]

> An alpha-4 ID in registered-lang should be OK. But something like
> "cy-cyde" absolutely is not.

I think we are in danger of mixing up a few issues here. These seem to
be:
	1. Using codes within the scope of RFC 3066: where a "script" code is intended, use 15924. But don't use something from elsewhere that is 4 letters. That should cover the above case.

	2. "confusion" between systems of alpha 4: provided it is clear at which point an alpha 4 is being used, such confusion should be avoidable, and hence the mn-mn example I've seen in the RFC I gueess. Here I get the impression that there is concern about the same 4 letters being used a) in 15924 and b) for Linguasphere tags. Because the purpose is different, they should not inhabit the "script" section of 3066 as I see it. Hence "cy-cyde" would be valid IF cyde identifies a script from 15924. Same as 1 really.

	3. Actual use of any Linguasphere tags: there seems to be "room" in the private use extensions if I've interpreted this all correctly so far, and this would seem to immediately remove any issue with 15924 and perhaps with the RFC needing to encompass everything for "implementers". 

Ok, have I made a glaring oversimplification somewhere here?