Technical Translation: Usability Strategies for Translating Technical Documentation

Indirect Observation 
In contrast to direct observation where users perform tasks with an observer 
present in the same room, indirect observation involves users performing 
tasks without the presence of an observer. This type of observation gener-
ally incorporates some form of recording mechanism, be it in the form of 
audio, video or software recording or some combination of the three. 
Audio Recording
Audio recording
can be useful when combined with verbal protocols. It in-
volves recording what a user says during the course of an experiment from 
general comments or spontaneous outbursts to verbalised thought processes 
in the case of think-aloud protocols (TAP). Audio recording and TAP are 
generally of greatest use in formative evaluations as the wealth of qualitative 
information can provide significant insight into the way the users interact 
with and perceive the interface. Particularly in the case of evaluations in-
cording with field notes on the events as there are no clues other than what 
users say and they may not always verbalise problem areas. It can be argued 
that TAP can place additional cognitive loads on subjects and can interfere 
with the way they perform tasks (see the discussion of attention in Chapter 
3). 
Video Recording 
Video recording or logging
counteracts both the problems of direct obser-
vation and also the problems of audio recording in that it does not require 
the observer to be present in the same room as the user and it provides a 
permanent record of the experiment while allowing the observer to see 
what the user did at a given point in the experiment. It can also be used in 
conjunction with some form of verbal protocol. By positioning cameras in 
182
volving interactions with software, it can be difficult to match the audio re-

---

Approaches to Empirical Evaluation
a variety of locations in the room where the test is being conducted it is 
possible to capture a wide range of data such as what the user types on the 
keyboard, what appears on the screen, whether the user refers to the user 
guide, as well as the user’s body language or facial expressions. With mod-
ern video camera technology, it is also possible to record reasonably high 
quality audio as well, thereby negating the need to synchronise different 
types of data. 
There are, however, certain problems associated with video logging. 
While this method provides valuable and comprehensive information, it 
only does this for as long as the user stays within the camera’s field of view 
(Dix 1998:428). An obvious solution would be to position the camera 
closer to the user but then we risk undoing the benefits of indirect observa-
tion with the obtrusiveness of the camera. Conversely, hiding the camera 
and filming users surreptitiously raises certain ethical and legal issues and is 
to be avoided (Faulkner 1998:123; Dumas & Redish 1993:206). A simpler 
solution is presented by Dumas & Redish (1993:225). Rather than hoping 
that users stay in the correct position during the test, Dumas & Redish pro-
pose placing pieces of adhesive tape in an L-shape on the desk to indicate 
where the documentation must be placed. In addition to ensuring that the 
document stays within shot (if necessary) this strategy also constrains the us-
ers’ movements and ensures that they stay in more or less the same position. 
If recording events on the screen, careful positioning of cameras, the 
user, the screen and lighting is necessary unless there is some mechanism for 
connecting a second monitor (Dumas & Redish 1993:224, 384). In such a 
scenario, the second monitor could be placed in another room where the 
observer records the images on video. But the issue of data synchronisation 
can be quite problematic where more than one camera is used (Dix 
1998:428; Preece 1994:619). At the very least some form of on-screen time 
code along with a mechanism for starting all of the cameras simultaneously 
will be necessary. This problem can be alleviated at least partially by 
through the use of software logging.