-
CAW N 19
-
Title: Unicode Contribution to the Cultural Adaptability Workshop
in Ottawa
-
Source: The Unicode Consortium
-
Date: December 18, 1997
-
Action: For discussion at the meeting
Preamble
The Unicode Consortium represents a wide array of major computer and software
makers, as well as research libraries, universities and user groups, who
have an interest in a universal character encoding and the internationalization
technology that is founded on top of it. The Consortium has been in operation
for 7 years and was preceded by over 3 years of informal cooperation by
a number of its current members.
The representatives at our technical meetings are internationalization
experts, usually with direct product responsibility. Therefore it is not
surprising that implementability has been one major emphasis of our technical
approach. The other is universality. Our members are keenly aware of the
cost savings that result from well-designed, generally applicable designs.
While the Consortium began around a character encoding, members quickly
felt the need to address those areas of cultural adaptability that intersect
with its definition of a character as an entity that has inherent properties.
In doing so, our members also agreed that while consensus was possible
on the fundamental algorithms, it would be counterproductive to provide
application programming interfaces (APIs). Not only because different segments
of industry differ and differentiate themselves by the style and manner
of their APIs, but also because the traditional concept of a functional
API is misplaced in modern object oriented languages.
Since the year of its incorporation, the Unicode Consortium has been
actively and closely cooperating with ISO/IEC JTC1/SC2/WG2. More recently
we have initiated work with SC22/WG20. Many of the delegates in our technical
meetings are also representing their companies in several national bodies
or other WGs in ISO.
Our contribution below reflects the lessons from our history, both as
a consortium, but also as individual implementers and as industry.
Business issues and Requirements
I.1 Meeting Market Demands in a timely manner
Internationaliztion, the process of desiging products so they can be adapted
to the cultural preferences of users in different markets, is an obvious
and strategic requirement for industry. Industry welcomes the efforts by
ISO to examine its standardization process in this light.
I.2 Single point of access and Complete Solution
Globally, industry is experiencing tremendous pressure to increase cost
efficiency of operations. This also affects standardization. The traditional
multilevel, dispersed, and at times, ponderous process of ISO standardization
is very expensive. Participitation in it has fallen in many countries.
In order to be cost-effective, industry requires a single point of access
to the standardization process, which in turn is responsible for a complete
solution.
I.3 Role of consortia
At the same time, industry has found that consortia are an efficient and
flexible means to produce timely technical work of high quality. In the
case of the Unicode Consortium in particular, we were able to combine work
on the classic aspects of character encoding with the other elements that
are needed to provide a complete solution to implementers.
I.4 Internationlization requires specialized, rare expertise
Developing new approaches to internationalization requires a form of expertise
that is in very short supply. Above all, ISO's efforts at reorganization
should strive to make the most efficient use of this limited commodity.
II. Technical Issues and Requirements
II.1 Openness and wide application
Internationalization affects all aspects of IT, from platform support (operating
systems) to programming languages and environments to application implrmentation
and user interface design. It has proven resistant to being encapsulated
opaquely into a subsystem that can be treated as a black box, no matter
how desirable that would be.
II.2 Core features
In some applications, the core (or 'value added') features (such as text
processing or database processing) are exposed to internationalization
in a manner that is very sensitive to both quality and performance. It
is especially in these cases that a one-size-fits-all approach can lead
to unacceptable degradation. Careful attention must be given to separate
the fundamental concept from acceptable variation in implementation strategy.
II.3 Focus on technical fundamentals
Industry reqirements of Internationalization standards are that they focus
on the underlying concept, and don't burden the implementer with artificial
details. (For example POSIX-style functional bindings, that may not applicable
in non-POSIX environments).
II.4 Stability concerns
An unfortunate habit of many standards is the concept of 'deprecated' features.
In a context where implementations (e.g. device interfaces or program source
code) routinely are updated to work in new environments, such a concept
makes sense. In Internationalization it is disastrous, because it makes
the user's data obsolete.
II.5 'Write only' standards
Character encoding, language, country, and other tagging schemes or locale
definitions, all must be designed and maintained in a 'write only' fashion,
if necessary preserving duplicates or even 'mistakes', in order to preserve
exisiting data. (The same applies to text and document formats, but they
are outside the scope of this discussion).
II.7 Ability to address customer requirements
Ultimately, products, whether based on ISO standards, or not, must meet
the requirements of and within the cost parameters defined by a competitive
marketplace. Successful companies have a long history of researching local
requirements, but also of devising general solutions that can meet these
requirements in a unified way, in order to contain the costs.
II.8 Market acceptance
Sometimes, the local marketplace demands solution that are at odds with
what is officially sanctioned in national or international standards. Industry
must not be prevented from addressing its customers in these cases, nor
burdened by being required to support little utilized 'check-the-box' features.
III Organizational recommendations
III.1 New Technical Direction for Internationalization
Create a new Technical Direction for Internationalization. This TD should
collect all standardization efforts for Internationalization across IT,
and have a mandate to create standards that are applicable in a range of
contexts, including new technologies for which no existing ISO standards
exist.
III.2 Scope
The goal of the TD should be to create standards for fundamental internationalization
technologies that affect many or potentially all areas of IT, are known
to widely applicable, and for which a technical consensus has emerged.
The new TD should focus to leverage, rather than duplicate existing efforts,
and work in ways that are more responsive and more light weight.
III.3 Make greater use of Technical Reports
In areas where internationalization work is ahead of implementation efforts,
greater weight should be given to assemble expert knowledge and contributions
into technical reports. TR's can be created faster, are therefore more
likely to be available to implementers when needed, and allow eventually
to create standards in the end that can incorporate more implementation
experience and at the same time be as stable as required.
III.4 Transposition to specific areas of application
Transposition of these standards into existing frameworks of other ISO
standards, e.g. operating environments (e.g. POSIX) or particular programming
languages should be the task of focussed efforts with the intent to produce
TRs.
III.5 Better use of limited pool of expertise
Working groups and subcommittees in this new technical directions should
work more closely together. Where the same experts attend more than one
group, the work should be re-aligned to reduce the number of meetings.
It is imperative then, that the new TD for Internationalization does its
work in a way that pools the experience of all parties, industry consortia
and national bodies.
III.6 Working with industry consortia
More efficient ways to leverage the work of and expertise collected in
industry consortia should be persued, in order to make the overall process
lighter weight, more responsive to industry requirements, and finally more
timely in response to market and customer requirements. Where possible,
responsibity for a standard should be delegated to consortia who have demonstrated
the ability to produce industry-accepted standards, allowing for a much
lighter-weight process.