TL;DR
This paper argues that the names we give to computer commands and objects (lexical semantics) are as crucial for usability as syntax and discourse. It reviews existing research on how these names affect user understanding, memorability, and interaction, proposing a featural framework to analyze these phenomena and bridge the gap between linguistic theory and practical system design.
What Problem Does This Paper Address?
While much research in human-computer communication focuses on syntax and discourse, this paper highlights the often-overlooked importance of lexical semantics – the study of word meanings – in command languages. The names given to commands and objects significantly impact a system's usability. Cryptic, misleading, or unmemorable names can frustrate users, even if the underlying system logic is sound. This paper aims to synthesize existing research on this topic and provide a framework for analyzing and improving command language design.
Methods
The paper doesn't introduce a novel algorithm or method but rather synthesizes existing research. It categorizes studies based on the phenomena they investigate:
- Suggestiveness: How well a command name hints at its function. This is analyzed by comparing semantic features of the name and the action.
- Memorability: How easily users recall command names. Studies explore the impact of name distinctiveness, frequency, and compositionality.
- Congruence and Hierarchicalness: Examining how the relationships between command names mirror the relationships between the actions they represent (e.g., inverse actions having inverse names) and how hierarchical structures in actions are reflected in names.
- Interaction with Syntax: How command names can influence or be influenced by the grammatical structure of commands.
- Non-words as Names: The use of special characters or icons and how their surface properties can map to meaning.
- Naming Objects and Categories: Research on how users describe and name files and categories, including the "Yellow Pages problem" of describing objects.
The paper proposes a featural framework to analyze these phenomena. This framework categorizes features of names and actions into four classes: semantic, pragmatic, syntactic, and surface features. The core idea is that distinctiveness between names and referents prevents confusion, while similarity between them facilitates understanding.
Datasets
Not discussed in the paper.
Key Insights and Ideas
- Names Matter Immensely: The choice of words for commands and objects is not a trivial detail; it directly impacts user learnability, memorability, and overall system usability.
- Beyond Syntax: While syntax and discourse are critical, the "words" of a command language have their own complex semantic landscape, mirroring natural language challenges.
- Predicting User Behavior: The "suggestiveness" of a command name can be analyzed by comparing feature sets of the name and the action, allowing for predictions of user guesses about command functions.
- The Power of Distinctiveness: Distinctive command names are easier to remember and less prone to confusion than generic or overly similar ones.
- Structure is Key: Congruent and hierarchical naming conventions (where relationships between names mirror relationships between actions) significantly improve learning and understanding.
- Bridging Linguistics and Design: The proposed featural framework offers a way to connect linguistic theories of semantics with the practical needs of human-computer interface designers.
Key Results
The paper summarizes findings from various studies:
- Suggestiveness: Experiments suggest that similarity calculations between feature sets of command names and actions can predict user guesses about command meanings.
- Memorability: Studies indicate that less distinctive, more general command names are harder to recall and lead to more frequent use of help menus. Conversely, computer-oriented or more specific names are better remembered.
- Congruence: Congruent names (e.g., insert/delete) are easier to learn than non-congruent ones.
- Hierarchicalness: Hierarchical names (e.g., move forward/move backward) are easier to learn than non-hierarchical synonyms.
- Naming Objects: Studies on file naming found that users often employ simple organizing paradigms, with a low likelihood of two people using the exact same main content word to describe the same object.
Limitations and Future Work
- Framework's Scope: The author acknowledges that the proposed featural framework is simple and has mixed parentage, intended more for practical design application than a formal linguistic theory.
- Syntactic and Pragmatic Features: The paper notes that the role of syntactic and pragmatic features in naming requires further investigation, particularly how selection restrictions and command language rules interact.
- Complexity of Natural Language: While command languages are becoming richer, they still operate under constraints not present in natural language, and the full implications of this are an ongoing area of research.
- Future Directions: The author suggests that linguistic theory can be practically useful for system designers, and the complex lexical phenomena in command languages may also be of interest to linguists. Further work is needed to explore the full potential of the featural framework in these domains.
Summary updated