Early in the article, the authors declare that little current research focuses primarily on visual communication. Perhaps this is due to the divisibility difficulties their own case studies highlight, and because evaluating visual communication in stark separation from usability is an undertaking of doubtful practicable worth.
If the authors sought to isolate efficacy of visual communication, it appears that doing so through ‘usability’ testing was not the optimal choice. Under usability testing, it is general ‘usability’ that is evaluated, not a specific element within usability, unless opinion regarding an element is coerced from subjects. If it is the visual aspect of usability alone that must be tested, then ‘usability’ demands redefinition at the outset, otherwise the issue of methodological appropriateness arises.
This article has limited direct application to my work scenario since I have no input in the creation of UIs.
What is of most academic interest is the authors’ fleeting mention of users’ reporting preference for UIs that were more difficult to use. Such an observation, which is antithetical to many design precepts, if correctly investigated, stands to yield very useful information. To explore this paradox, future studies might involve users building prototype interfaces autonomously, using a restricted set of icons or, ideally, creating them freeform.
On usability, I would be reluctant to endorse the authors’ questioning methods, or at least to make them central to the testing, primarily for this reason: if subjects are asked simply which UI they prefer – ‘A’, or ‘B’, they might answer ‘B’, but deeper questioning often proves futile, since non-technical users have a tendency toward laconic and circular responses. For more lucid feedback, a superior method would bolster subject responses with quantitative data derived from counters and other automated logs running behind a subject-manipulated system. Such a combination would illuminate both correlation and discrepancies between usability (evidenced by impartial data) and subject preference (derived from interviews).
For technical communicators whose work includes visual communication and who agree with the researchers’ opinion regarding the need to evaluate visual communication independently, this article has some worthwhile implications, such as the employment of a testing method that prioritizes visual communication over other aspects of usability.
For theoretical rather than practical purposes, there is much to be extrapolated from the researcher’s recommendations. For general applicability however, because the write up contains ambiguities and the methodology is open to question on a number of counts, the author’s conclusions form a weak basis for adoption of their recommendations. I would therefore caution against abandoning the integrity of usability in favour of isolating one of its elements – visual appeal – for exclusive examination.
In sum, it is because the issue of visual communication resides in the affective domain (i.e. do viewers like what they see?) that its testing cannot be simple (Wang, 2000). Hence, any usability methodology must be tolerant of user whimsy, and there will be gender-, culture-, and other audience-specific variables to be considered.
The discussion of primary value elicited by this article is that of the necessity of a broad rethink of visual communication within the field of technical communication:
- Graphics must account for a comprehensible mental image metaphor and ‘look-and-feel’ is meaningful within usability only if defined along the lines of ‘quality of appearance characteristics and effective interaction sequencing.’ (Gurak, 2003.)
- Creation of visuals, such as icons, still lacks design standard regulation (Gurak, 2003; Burnett, 2005).
- Technical rhetoricians, because their discipline favours scientific notions of images as neutral objects and also because scholastic tradition privileges textual discourse, must resist lapsing into uncritical practice when analyzing and creating images. Until recently, when technical communicators discussed visual design, they typically referred to the formulaic placement of visuals within text. This perspective relegated visual communication to secondary status and often caused it to be lumped in with page layout or design (Salinas, 2002).
- The possibility that graphics can be instructive (as well as aesthetic) independent of textual accompaniment merits examination. Design of images, especially the strategic composition of visual representations incorporating cultural / conceptual significance, must be wrested from the domain of the graphic artist and brought within the sphere of scientific technical communication.
- Although aesthetics strongly influences information design, designers remain reluctant to acknowledge its significance (Kostelnick, 1998) and rhetoricians whose principal medium is verbal text are similarly shy of embracing its potential.
- Interactive graphics are the prime candidate for study by technical communicators, but it is computing specialists who shape the conventions of and standards for visual interfaces. Technologically oriented expertise usually downplays aesthetics (Preece et al, 2002) because its protocols stem from a design language distinct to that of technical communication and other schools of context-oriented task analysis.
- Since visuals are seen nowadays as intrinsic components of technical communication, technical communicators are influencing onscreen layout, UI design, and graphics (static, active, and interactive). This responsibility is swelling as Web pages and other electronic platforms become the typical media of technical content and the information these must deliver grows increasingly complex (Jackson, 2000).
 Especially true when testing UIs.
 In this article, the term ‘visual communication’ is itself polysemous: it refers to pictures (icons and other onscreen elements) – but is more usually a holistic term covering the spectrum of graphic design precepts.
 The wizards and online help that I help produce are designed around standardized, basic templates that are adherent to the Windows model. They feature blue-grey color schemes for frames and bars, and black on white for text. Visual elements in print and CD-ROM manuals consist of formulaic line art elucidating potential sticking points within procedures. All visuals, online and paper-based, are minimalist and composed of three colours at most.
 For illustrative purposes I included in the appendix a short, representative extract from a usability questionnaire transcript.
 However, there is scant research to support the authors’ argument for independent evaluation of visual communication, and what stands to be improved by such separation is never fully stated (although we must assume it is improved usability via superior visual appeal).
 For researchers seeking hard support for robust alternatives to conventional usability testing, these studies provides little of insight or empirical value. The holes in the methodology are too numerous and the authors’ sui generis conclusions too few and fragile.
 For example, engineering students preferred a visual metaphor that resembled an operation panel over a book-themed metaphor (Hailey and Hailey, 1998).
 Most technical communication textbooks written since the mid 1980s contain chapters on visual communication (covering typography, graphics, color, and, since the early 1990s, Web design).
 Despite these fields traditionally belonging in the realm of human-computer interaction (HCI).