Browse Prior Art Database

Design to Enable Locale Tagging for the Database Manager

IP.com Disclosure Number: IPCOM000110797D
Original Publication Date: 1994-Jan-01
Included in the Prior Art Database: 2005-Mar-26

Publishing Venue

IBM

Related People

Obermann, D: AUTHOR

Abstract

The OS/2* Database Manager (DBM) currently uses the OS/2 "country code" tagging paradigm to indicate the cultural bias of a database. Unfortunately, this tagging paradigm is increasingly awkward to use and extend as it bundles language with territory. Consequently, country code tags are not used in the industry beyond DOS**, Windows**, and OS/2.

This text was extracted from an ASCII text file.
This is the abbreviated version, containing approximately 17% of the total text.

Design to Enable Locale Tagging for the Database Manager

      The OS/2* Database Manager (DBM) currently uses the OS/2
"country code" tagging paradigm to indicate the cultural bias of a
database.  Unfortunately, this tagging paradigm is increasingly
awkward to use and extend as it bundles language with territory.
Consequently, country code tags are not used in the industry beyond
DOS**, Windows**, and OS/2.

      Another cultural tagging paradigm, the POSIX "locale" model, is
emerging as the personal computer and workstation standard.  This
model clearly distinguishes between territory and language, plus it
offers the advantage of portability across a number of operating
platforms.

      This proposal provides a design and functional specification
for migrating DBM cultural tagging to the "locale" model, thereby
enhancing function and portability.

BACKGROUND AND OVERVIEW - The OS/2 country code (CC) tag, inherited
from DOS and also used by Windows, is a device whereby these
operating systems and their applications specify their preferences or
bias for culturally sensitive data and processing such as date/time
formats, numeric separators, collate scheme, monocase scheme, etc.

      If there were only one culture per country, the CC paradigm
would be an adequate model for cultural specification.  However, it
is clear that this is not the case.  Within Canada, for example,
English and French cultures flourish and insist upon maintaining
their uniqueness and identity.  While CC=002 identifies
Canadian-French conventions, Canadian-English users have no such CC,
but instead must use the US-English CC=001.  This is because there is
only one CC per country; since the CC for Canada is assigned to the
Canadian-French bias, no CC is left for Canadian-English.

      The CC paradigm also bundles language with country, or rather,
fails to provide any means at all to specify language.  Though

French, German, and Italian are all official national languages of
Switzerland, there is only one CC=041 for the country, leaving no
flexibility for cultural differences due to language.

      Finally, some of the CC values are far too broad in scope to be
useful.  For example, CC=003 is for all of Latin America.  Yet there
is no common currency for Central and South America, leading one to
doubt the accuracy and usefulness of the currency format implied by
CC=003.  Currency is just one of several attributes subject to this
ambiguity.

      Because of problems such as these, the computer industry has
failed to embrace the CC paradigm as a standard for cultural
identification, and instead has looked elsewhere for a solution.  One
proposal which has gained acceptance in the UNIX*** environment (and
which is making inroads into the PC arena) is the POSIX model of a
"locale".  This model suffers none of the shortcomings mentioned
above for CC, and due to its growing acceptance across the industry,
enhances portability as well...