RE: General policy

[Masataka Ohta <mohta@necom830.cc.titech.ac.jp>]

> One possible labelling scheme when short octet strings must be labelled
> is to use ISO 2022 escape sequences, but require that all character
> set designation and invocation is done in the beginning of the string,
> before any "real" character occurs.
> 
> This is a relatively compact method that does not require developing new
> standards. It does not allow the use of the ISO-2022-JP method, however.

If you use ISO2022, it does not differ so much whether you use
an announcer only once at the beginning or you switch during
text.

> I'm not suggesting this at present, just mentioning it as one possible
> mechanism.

ISO 2022 is not so bad as an unofficial intermediate encoding method, if
it is compatible both to ASCII and to the ultimate encoding, that is, it
must be, practically, 7 bit. No 8 bit 8859 please.

In ASCII only environment, we didn't need any labelling. With the
ASCII compatible ultimate encoding, we also don't need any labelling.

But, if you introduce some official intermediate encoding, we
must support labelling, we must support conversion from the
intermediate to the ultimate encoding during the transition
period.

Moreover, during the transition period, we can't assume good
properties of encoding shared by ASCII and the ultimate encoding,
which makes the transition to the intermediate encoding quite
difficult and transition to the ultimate encoding meaningless.

So, we should not introduce an intermediate encoding.

> If DNS dies when there's an ESC in a TXT record,

It does not.

					Masataka Ohta