Speech graffiti habitability: What do users really say?. Tomko, S & Rosenfeld, R In Proceedings of the Sigdial 2004 Workshop 5th Annual Meeting of the Special Interest Group on Discourse and Dialogue, pages 81–84, January, 2004.
abstract   bibtex   
The Speech Graffiti interface is designed to be a portable, transparent interface for spoken language interaction with simple machines and information servers. Because it is a subset language, users must learn and adhere to the constraints of the language. We conducted a user study to determine habitability and found that more than 80% of utterances were Speech Graffiti-grammatical, suggesting that the language is acceptably learnable and usable for most users. We also analyzed deviations from grammaticality and found that natural language input accounted for the most deviations from Speech Graffiti. The results will suggest changes to the interface and can also inform design choices in other speech interfaces.
@inproceedings{tomko_speech_2004,
	title = {Speech graffiti habitability: {What} do users really say?},
	abstract = {The Speech Graffiti interface is designed to be a portable, transparent interface for spoken language interaction with simple machines and information servers. Because it is a subset language, users must learn and adhere to the constraints of the language. We conducted a user study to determine habitability and found that more than 80\% of utterances were Speech Graffiti-grammatical, suggesting that the language is acceptably learnable and usable for most users. We also analyzed deviations from grammaticality and found that natural language input accounted for the most deviations from Speech Graffiti. The results will suggest changes to the interface and can also inform design choices in other speech interfaces.},
	booktitle = {Proceedings of the {Sigdial} 2004 {Workshop} 5th {Annual} {Meeting} of the {Special} {Interest} {Group} on {Discourse} and {Dialogue}},
	author = {Tomko, S and Rosenfeld, R},
	month = jan,
	year = {2004},
	keywords = {Human-Machine Speech Communication},
	pages = {81--84},
}

Downloads: 0