Re: [idn] hostname history hell

[John C Klensin <klensin@jck.com>]

--On Tuesday, 20 November, 2001 15:52 -0600 "Eric A. Hall"
<ehall@ehsco.com> wrote:

> John C Klensin wrote:
> 
>> It was also noted at the time that hyphen was a common
>> argument ("qualifier") introducer in some command languages
>> and a common introducer for negative numbers in input strings
>> in others, and hence better avoided.
> 
> The argument-introducer aspect is probably still applicable.
> Should we keep the no-leading-hyphen rule for i18n host
> identifiers? We're not explicitly tagging -GW or -TAC
> postfixes so that is not required (but may be desirable for
> compatibility).

My own personal position/ bias is that we should be as
conservative about what is permitted in an IDN as possible
consistent with meeting the requirement for identifiers drawn
from any of the world's languages or combinations of them.  Put
differently, I think the general model of the old "hostname"
rules has served us well.  I would like to think of the IDN work
as expanding that model to include additional alphabetic and
ideographic characters, rather than discarding the model and
seeing how much "stuff" we can put in.

If a too-restrictive model turns out to be a mistake, it is
possible to expand it later (just as "leading digit" was
unblocked); if we adopt a model that turns out to be too broad,
there is probably no way back.

On that basis, my inclination would be to:

	* continue to prohibit leading (or trailing) hyphens
	
	* continue to prohibit all spacing characters
	
	* continue to prohibit all punctuation characters except
	for that hyphen and the label-separating period (full
	stop, ".")
	
	* prohibit, in the spirit of the hostname rules, all
	symbol and drawing characters

We don't _need_ them for identifiers.  Some of them will, sooner
or later, run up against a legitimate command language or cause
"interesting" lexical parsing problems (even if they don't cause
problems in today's URI syntax definition).  High risks,
marginal benefit.

Just my opinion, of course, but I have these scars...

    john