Abstract a computing system, he therefore inherits the

Abstract

As long as we can remember humanity always followed its
goal towards evolution. Nowadays, evolution mainly takes
form through technology. For the last decades the race to
develop more performing and complex systems has never
ceased. However, as Ted Nelson said, “the good news about
computers is that they do what you tell them to do. The bad news is that
they do what you tell them to do” 1. And as you might already
know, Humans are far from perfection, and despite countless
security layers we still make errors. This paper focuses on the
errors that have been made by humans with computer
interfaces. We will describe some inadvertent errors then
provide some techniques that could be used in order to
prevent these.

Introduction

Computer interfaces get more complex over time and despite
how powerful they might be, they still require human
management. When someone create a computing system, he
therefore inherits the responsibilities over his work. Any lack
of control or negligence might end up in an incident affecting
others and this is precisely why Humans errors cannot be
tolerated. Errors are not an active choice and can most of the
time be identified, predicted and redesigned. We will base our
analysis on Rasmussen’s Skill Rule Knowledge (SRK) model
2 and a paper from the Health and Safety Executive (HSE)
division of Great Britain 3 to define the factors that lead to
Human errors.

ASSIGNMENT 3 1

Problem description and Inadvertent errors

Accidents occur everyday because of Human errors even
though some are attributed to human errors when ‘all other
factors have been eliminated’ 4. Everybody is entitled to make
mistakes since technology is getting more performing and
complex to use over the years. Nowadays it is actually
considered as ‘normal’ to make these. The majority relies on
what Nash Grier said “As humans, we’re going to make mistakes. It’s
what makes us human, and most of the time, the most effective way of
learning is from a mistake”; however it is critical to not
underestimate how such a small error can impact many
people, even millions.

Figure 1: Human Failure Types

According to the figure 1 above, they are two main types of
inadvertent error, action errors and thinking errors.

ASSIGNMENT 3 2

Skill-based

The action errors are defined as skill-based mistakes. They
happen if the person’s focus is redirected. As the International
Journal of Research in Engineering and Science 5 suggests this
behaviour leads to ‘semi and fully-automatic elapsing actions’.
This can be explained with familiar tasks; for example a
developer omits to secure a security layer, it is a lapse failure
type. However if the developer move a switch down rather
than up, it would be a slip failure type since it is a commission.

The best way to counter these errors would be to implement
systems able to detect malfunctioning or incomplete elements
such as alarms. Since the designer is most likely not the person
who manipulate the software, it is essential to create a
communication between the machine and the user. The
objective is to avoid mistakes by telling the user something
isn’t right as follows “Error 442: security layer n°3 not secured”. An
other simple possibility would be to ensure the task to
somebody who is highly-trained and has experience with the
tool he’s using.

Rule-based

Belonging to the thinking error category, rule-based errors are
decision-making failures. They represent actions that don’t
need any knowledge to be completed. Rule-based behaviour
comes from the mis-application of a rule; the person bases his
act on a remembered rule. According to the Data Center
journal 6, Robert McFarlane, principal and data center
design expert at Shen Milsom stated that “Reputable studies have
concluded that as much as 75% of downtime is the result of some sort of
human error”. One of those error is described as ‘overloading a
circuit by plugging in a server’ by the website ComputerWeekly
7. The employee carried out the task believing it to be the

ASSIGNMENT 3 3

right thing to do when it actually was not. He mis-applied the
rule which states that servers must be running at all times.

Knowledge-based

As the previous point, this behaviour is part of the thinking
category. The difference is that knowledge-based errors come
from an error of judgment. A computer professional might
misdiagnose an issue due to insufficient information or a lack
of experience resulting in a mistake.

Based on the previous example about datacentre, an employee
could pull the wrong power cord because he doesn’t know
what each cord does. As the article from ComputerWeekly
website suggests, ‘Unplanned downtime can lead to financial loss as
well as reputational damage to businesses’ 7 making it the biggest
internal threat to the business. As a result companies ‘implement
a redundancy plan or create a datacentre resiliency strategy’ 7. It is
now common, even essential for such companies to develop
back up plan. In addition, improving diagnostic tools, training
and sharing knowledge within the business are very strong
recommendations to counter thinking errors.

Summary

In this paper we saw that human error is the result of three
types of mistakes: skill-based, rule-based and knowledge-
based. Some are more difficult to counter whereas others can
be easily avoided if the right systems are implemented.
Unfortunately as Paul Inett, vice-president of Enlogic Europe
said, “you’ll never eliminate human error …, but you can make choices
in both technology and training that will help to reduce its severity and
impact” 7.