Re: Status of SUMO?
Adam, Philippe, et al.,
As I've said many times, I am skeptical of any attempt
to legislate content. I believe that the content of
the ontologies should be allowed to grow and change
with the needs of applications and the demonstrated
successes in practice.
The best candidates for standardization are the
metalevel features, frameworks, and terminology, which
should be able to accommodate any or all proposed content.
Metalevel standards should facilitate the merger and/or
interoperability of multiple ontologies, and we should
consider the criteria for evaluating any proposed
standards. Examples of the kinds of criteria:
1. Flexible, systematic terms that have cognate noun,
adjective, and verb forms with suitable prefixes
or modifiers for related terms (e.g., type, noun;
type, adj; to type, verb; typing, participle;
subtype; supertype; untyped; etc.).
2. Familiarity across a wide range of disciplines
that will use or contribute to the development of
the ontologies: e.g., logic, linguistics, philosophy,
computer science, programming languages, databases,
and many different fields of application.
3. Flexible methods for combining, modularizing, merging,
relating, and extracting submodules for various
theoretical and practical purposes.
4. Compatibility with the formats and requirements
for closely related fields, such as database systems
(e.g. SQL, OO databases, and various design and
development tools, notations, and techniques),
programming tools and notations (e.g., UML and related
technologies), engineering modeling (e.g., EXPRESS),
etc., etc.
Finding a single, ideal system that is completely compatible
with all of these areas is undoubtedly impossible, but it
should be possible to find large commonalities that can be
adopted while minimizing the amount of pain and suffering
necessary for converting any ontology to the standard form.
John