Touchstones for "Mapping"

["Martin J. Dürst"]

On 2009/03/25 7:46, Lisa Dusseault wrote:
> I heard a few principles in IDNABIS this week that I'd like to
> identify.  They all relate to "Mapping", or "input assistance", but at
> any rate we're talking about pre-lookup logic.
>
> 1.  Canonical forms:  encourage labels to be stored and displayed,
> whenever possible, in a canonical form.

Yes, very much so.

> The protocol only uses the
> canonical form.

"The protocol"? If you mean DNS, then I think nobody disagrees.
For higher levels (EAI, HTML,...), it's again highly desirable,
but I don't think it can be inforced.

> 2. Conservatism: in no cases map something valid to something else
> valid!

Yes, I think provisions such as this are very good to give an outer
boundary of what is allowed for variant mappings.

> 3.  Consistency of conversion: encourage software to apply the same
> global conversions (aka mappings) if they feel it necessary to help
> users get from an invalid input to a ULabel. Some implementors will,
> however, choose to use alternate mappings no matter what we do.

There are some cases where localized mappings help (the Turkish
dotless I is the classical example). But these are few and far
between. In my impression, the discussion up to now often has
been between "we need to allow any mapping, because a single one
won't do" and "we cannot allow any variations, because they
imply chaos/security issues/...". Your point above is a very
important step in the right direction: Mapping variations may
occasionally happen, because they are necessary, but they should
be the exception rather than the rule.

> 4.  Confirmation: encourage software implementors to confirm a valid
> ULabel as the user's intended input, where possible and appropriate

As we all know, user confirmations can be very helpful, but also
very annoying and counterproductive. There may certainly be cases
where they would be helpful, but I cannot currently come up with
a specific example.

> I believe this is a reality-based approach,

Very much so!   Regards,    Martin.

but that's only insofar as
> I understand reality!
>
> Lisa
>
> --- Scanned by M+ Guardian Messaging Firewall ---
> Messaging Architects sponsors The Spamhaus Project.
>
> _______________________________________________
> Idna-update mailing list
> Idna-update at alvestrand.no
> http://www.alvestrand.no/mailman/listinfo/idna-update
>