|
My previous postings have discussed semantics in terms of: - Requirements for better semantics
- Semantics and business processes
- Semantics and languages
- Semantics and the translation of ideas into solutions
- Semantics and common languages that avoid the need for translation
- Semantic standards (e.g. Semantics of Business Vocabulary and Business Rules)
- Natural languages
- Executable specifications
- Modeling standards (e.g. ODM, MOF)
Deeper dives into any of these areas would benefit from further clarifications on the nature of semantics. This requires revisiting the topic of how languages relate to semantics, which leads us to semiotics: - Semiotics – The study of language signs and symbols, which consists of three sub disciplines:
- Semantics – The study of meaning.
- Pragmatics – The study of how signs and symbols in languages relate to the people using them.
- Syntax – The study of grammar.
This high-level semiotics taxonomy provides a fertile ground for exploring semantics. Given the growing emphasis on the so-called “semantic Web” and the increasing recognition that we need to move beyond syntax-based processing to semantically enabled business systems; i.e. systems that have some understanding of the larger context that surrounds the execution of instructions. It turns out that context is a key concept: - What is context? – Context is information about the environment, which includes the identity, location, state and/or behavior of other actors capable of influencing the behavior of a system.
- What does context have to do with semantics? – Pragmatics deals with context; i.e. how people use language in a social context. This impacts the meaning of language, i.e. it impacts semantics. Syntax also deals with context. Some languages (e.g. Russian) are inflective wherein words have many different endings to indicate how they relate to other words. Different word endings represent context at a level that impacts the rules for building grammatically valid statements, which impacts meaning; i.e. semantics.
- What are context-aware systems? – Context-aware systems adapt their behavior based on context. Using context in this way is similar to using metadata to better understand data. Two examples illustrate:
- Metadata – The source and currency of credit scores are examples of metadata, which helps in judging the reliability and accuracy of credit scores. Metadata is focused on making data more meaningful, i.e. metadata is about data.
- Context – The location of a device is an example of context, which helps in providing more relevant services. Context is focused on making system behavior adaptable to circumstances and therefore more consistent with system goals, i.e. context is about behavior.
- What is the relationship between semantics and ontology? – Semantics is not only an element of semiotics but also of ontology. Ontology was discussed in previous postings in terms of ontology languages and meta-languages. In philosophy, ontology is the study of existence or reality. In IT systems ontology deals with the structure and behavior of systems. While taxonomies are classification systems that represent relationships (e.g., peer, parent/child, etc.) between real or virtual artifacts, ontologies specify these relationships in more formal terms using classes, subclasses and associated methods capable of supporting logical operations such as computation and inference.
For organizations that wish to adopt semantic technologies, a key task is to define a strategy and implementation road map that details how to progress from existing syntax-centric computing models to more advanced models that support semantic processing and semantic integration. These advanced models will involve: - Grammar
- Controlled vocabularies
- Taxonomies
- Thesauri
- Ontologies
Each of these topics requires further discussion in future postings.
|
Data Governance