ÉVALUER L'EFFICACITÉ DES DOCUMENTS TECHNIQUES PROCÉDURAUX : UN PANORAMA DES MÉTHODES

This paper takes a panoramic view of the methods employed in testing the usability of technical documents. Three different kinds of approach are distinguished that, between them, embrace the multi-faceted aims of different kinds of evaluation (i.e., usability verification, choice-support and troubleshooting). These three approaches have been described by Schriver (1989) as (a) document-focused ; (b) expert-focused ; and (c) user-focused. This paper outlines the main charactenstics and ways of proceeding for each set of approaches, and discusses their strengths and weaknesses. Document-focused methods focus on the document itself (and use measures such as readability formulae). Such measures can be used without recourse to experts or to readers. This advantage, however, has limitations. Readability formulae, for example, are only suitable for assessing the text, and they do not consider linguistic factors (e.g., terminology, syntax, and information organisation), para-linguistic factors (e.g., information layout), extra-linguistic factors (e.g., information presentation format), contextual factors (e.g., device affordance) or individual factors (e.g., prior knowledge). However, they can be used to compare different versions of the same or similar documents. Expert-focused methods focus on expert readers (and use measures such as technical review). Such methods are useful for assessing the content of a document, because of the experts' knowledge, but experts sometimes lack the ability to assess the suitability of the text for non-expert readers. User-focused methods focus on these actual readers. Such methods include both process and outcome measures. Process measures rely on behavioural tests (such as performance tests, and analyses of reading behaviour) as well as verbal protocols. Outcome measures include comprehension and memory tests and users y judgements. User-focused methods are useful for revealing misconceptions in the design of a document and in its suitability for the user, but they are rarely suitable for judging the content of a document. The review concludes by suggesting that an integrated approach that combines expert-and reader-focused methods is a precondition for the effective design and evaluation of technical documents. L'évaluation de la qualité et/ou de l'utilisabilité des documents techniques procéduraux (guides d'unlisation, modes d'emploi, etc.) reste souvent absente du processus de conception, au risque de produire des documents i