Technical Translation: Usability Strategies for Translating Technical Documentation

The Low-Capacity Channel 
Linking sensory memory to STM is what is termed the “low-capacity 
channel” (Downton 1991:23). This channel serves as a conduit for informa-
tion passing from sensory memory to STM. In practice, however, this 
channel has a low transfer capacity, something which is evident from the 
difficulty we experience in paying attention to many different sources of in-
formation simultaneously. In addition to transmitting information, this 
channel also converts the information from its raw, physical and unproc-
essed state into symbolic representations which can be used in STM. In-
deed, this is where perception occurs. The limited speed with which this 
information is converted helps to explain the low capacity of the channel in 
general. This limitation means that the channel is very prone to overload-
ing. 
Short-Term Memory (STM)
A popular way of explaining the concept of STM is to describe it as a 
“scratchpad” or as RAM in a computer (Dix 1998:28; Hill 1995:19). STM 
is responsible for storing information that we are currently using. It is 
where we carry out all of our memory processing, encoding and data re-
trieval. STM allows us to “do” things with information. We can also filter 
information here and discard information which is no longer needed. 
Card 
et al.
(1983:38) argue that STM (or working memory as they call 
it) is really only an activated subset of information stored in long-term 
memory (LTM). While it is true that STM obtains some of its input from 
LTM, e.g. stored knowledge, procedures etc., information passed on from 
sensory memory also provides STM with its input. 
In contrast to information stored in sensory memory, information in 
STM is stored in the form of symbolic representations or schemes (Coe 
1996:72). However, like sensory memory, information is stored in STM 
temporarily. The information is lost, overwritten or replaced after 20-30 
seconds (Downton 1991:24), although with practice information can be re-
tained for several hours (Coe 
ibid.
). That information is only stored tempo-
rarily is due to the limited capacity of STM. In 1956 Miller posited that the 
capacity of STM is seven chunks plus or minus 2 chunks. This “7±2” rule 
is universally accepted as fact (Faulkner 1998:34; Coe 1996:72; Downton 
lustrated using the following sequence of numbers: 
116 
1991:23; Dix 1998:28) and is generally true for most people. This can be 
il

---

Memory
0352765994 
The average person may find it difficult to remember each digit in this se-
quence. However, if we group the digits into smaller sequences as we 
would with a telephone number, each sequence can be treated as a single 
chunk: 
035-276-5994 
So, instead of remembering ten separate pieces of information, by chunking 
the information we reduce the amount of space required to remember 
them. An interesting property of chunks is that what actually constitutes a 
chunk depends on individual people and the content of their LTM (Card 
et al. 
1983:36). According to
Downton (1991:24) the number of chunks 
which can be stored is independent of the amount of information each 
chunk contains. We can, therefore, combine small chunks to form larger 
chunks and so on. For example, letters (small chunks) form words (larger 
chunks) which can be combined to form sentences (even larger chunks) 
and so on (Faulkner 1998:73). With sufficient practice and rehearsal in 
STM, several sentences can be treated as one single chunk.