On Wed, Jul 20, 2011 at 05:31:37PM +0200, Florian Zeitz wrote: > Am 20.07.2011 17:16, schrieb Joe Hildebrand: > > On 7/20/11 6:18 AM, "Florian Zeitz" <[email protected]> wrote: > > > >> Maybe I'm misguided, but I disagree with the "No more big tables" > >> statement. You either need at least a code point to General_Category > >> mapping and a way to tell whether a character is compatibility > >> decomposable as input for the algorithmic classification, or a table > >> containing the result of of the classification. Both seem relatively > >> large to me. > > > > Perhaps "no more big tables maintained by the IETF" would resonate better? > > > Did the IETF itself ever maintain one? > There is still the "PRECIS Derived Property Value Registry" mandated by > section 9.1 to be maintained by the IANA, which (as I understand it) is > basicaly the "table containing the result of the classification" that I > mentioned.
All stringprep processing was based on a lookup table. In theory, IDNA2008 and PRECIS are based on an algorithm. Yes, there are tables at iana.org, but those are not supposed to be normative. Granted, you still have big tables in Unicode, but not separate tables at iana.org in addition to the base tables of Unicode properties. /psa _______________________________________________ precis mailing list [email protected] https://www.ietf.org/mailman/listinfo/precis
