Proof in Computer Science
Content
Charles Benitez
Professor Espinosa
Middle Circle
17 October 2013
Proof in Computer Science: Usability
Proof is a universal idea that is necessary in providing evidence for credible claims.
While the general idea of proof remains the same everywhere, the proof and evidence found in
individual fields can vary widely. In the field of computer science, most methods of proof are
mathematical in nature. There are two main aspects of computer science offered at FIU: the
logical and scientific track that deals with the technical and mathematical aspects of computer
science, and the software development track that deals with software engineering (cis.fiu.edu).
The distinction provided by FIU accurately reflects on the field of computer science as a whole.
Most of the proof in the first field is highly technical, and while it is useful for experts in the
field, it is not useful for the general public. The use of computers by the general public has
grown exponentially in the last decade, and methods of proof that involve computer science
and the general public are found in the software development track of computer science. These
methods are generally more abstract and comprise of the involvement of the end user and the
effects of computing on them.
The method of proof that will be focused on is usability. Usability is a method of
determining how “usable” a computer system is. It measures the system’s effectiveness,
efficiency and the overall satisfaction of the user (usability.gov). Usability is measured primarily through
usability testing, which allows developers direct feedback on how real users use the system. This
Benitez 2
is an important aspect of software development because even though the program may work
completely and perfectly, if the user has no idea on what to click in order to make the program
work, has the program actually been completed successfully? Usability is the main way for
computer science to come away from the technical aspects and jargon and allow the
perspectives of non-experts to come into play. However, this also adds a great deal of
vagueness and difficulty when it comes to providing evidence. It is simple to prove that a
program works correctly. If a program is meant to find the average of 100 distinct values,
checking the output of the program to see if the average is correct would prove whether or not
a program works. However, if the program was written for use in a middle school Algebra class
and the user needs to convert every number entered into the program into binary code first,
the program has proven that it works, but it has not proven its usability.
The main challenge to this is simple: how do you objectively quantify something as
subjective as the end user experience of a product? We would never know with 100% certainty
that a product is free from usability problems, because that would involve the feedback of
every single user that will ever use the product. This is where the concept of the Rule of Five
attempts to draw a line and make usability testing more manageable. The Rule of Five basically
states that no more than five users are necessary in order to find an acceptable majority of
usability issues. The main idea behind it is that there are a finite number of usability issues that
can occur. When you have a small amount of users, each user’s data brings in more insight, but
there is inevitably going to be overlap in your findings. Usability is very costly and timeconsuming, especially when it is used to test many aspects and intended outcomes of products.
Therefore, the Rule of Five states that after five users, most of the relevant data on the usability
Benitez 3
of a product will have been found, and after five users the small amount of new data does not
outweigh the cost and difficulty of conducting the tests (nngroup.com).
The validity of this method of proof is difficult to say. It is my argument that usability is
not a valid method of proof, but it is acceptable for use solely because there does not seem to
be a sustainable and practical alternative in computer science. The Rule of Five is a great way to
make usability work in practice, but in theory, the Rule of Five is basically admitting that we will
never be able to truly prove that a program is usable, and gives us a guideline on where it is
acceptable to give up. This makes it easier to perform usability tests, but it draws from the
validity of usability as a method of proof. There are so many different factors when it comes to
usability that, in practice, it becomes impossible to keep track of them all. Users of computer
systems come from different age groups, different cultural and ethnic backgrounds, different
degrees of education, different levels of computer literacy, etc. Usability testers have to
consider their target audiences, and this becomes another excuse to let developers give up on
making a program usable to a certain demographic.
The concept of usability has many parallels to the scientific method. In science, when a
hypothesis is tested, it is never said that one has proven a hypothesis to be correct, but only
that one has failed to reject the hypothesis. Science is therefore a never ending quest to find
knowledge. This is why usability does not meet the same level of validity as a method of proof
as the scientific method does. Usability allows one to say when it is acceptable to stop looking
for problems, while science never gives up. When a counterexample is found in science, a
paradigm shift might occur in order to allow this new knowledge to come in. When a
counterexample is found in usability, such as Adobe Photoshop not being useful to the blind, it
Benitez 4
is simply disregarded because the blind are not part of the target audience. While in practice it
would be virtually impossible to make a photo manipulation program usable to the blind,
usability concedes to the fact that the hypothesis has been rejected, but opts to continue to
accept it. This is why usability is unacceptable as a method of proof, but is necessary for
computer science to function practically.
I took part in a usability test for a Technical Writing in Computer Science course last
semester, and the firsthand experience I gained through this serves as a great example for
usability. A usability test was conducted for the Society of Technical Communication Palm
Beaches Chapter, and the point of the study was to determine whether or not the website was
usable to new clients (sites.google.com). Five users were tested in accordance with the Rule of
Five, and most of the users agreed that the website was not usable to new users interested in
Technical Communication. There are many technical reasons as to why the website was not
usable, but relevant part of the usability study to this assignment is not what was presented by
the study, but what was rejected and was not included in the study. A majority of the users
were either interested in Technical Communication or were able to feign interest enough to be
able to find usability issues with the site. However, there was one troublesome user that could
not have cared less about Technical Communication and simply found the entirety of the
website boring and uninteresting. This data was ignored because the usability test was geared
towards individuals interested in the field, and therefore the lack of usability to this specific
user was not considered.
This specific example, however, only serves to prove that usability is not a valid form of
proof. The user was not impressed whatsoever by the website, and the website was in no
Benitez 5
means usable to her. However, instead of attempting to make the website usable to this user
and accept that the user has proven the website to not have usability, the user was completely
ignored and omitted from the findings. In reality, though, a website for a local branch of a
group of a very specific branch of computer science cannot be expected to woo uninterested
and apathetic college students of unrelated majors. Yet, the sheer fact that usability admits that
a claim is false and continues to state that it is true further proves that usability is not in itself a
valid method of proof.
Benitez 6
Works Cited
"Palm Beaches Chapter." STC PBC. N.p. Web. 17 Oct 2013.
<https://sites.google.com/site/stcpalmbeaches/>.
"Undergraduate Programs." School of Information and Computing Sciences. N.p. Web. 17 Oct
2013. <http://www.cis.fiu.edu/programs/undergrad/>.
"Usability Evaluation Basics." Usability.gov. N.p. Web. 17 Oct 2013.
<http://www.usability.gov/what-and-why/usability-evaluation.html>.
"Why You Only Need to Test with 5 Users." Nielsen Norman Group. N.p. Web. 17 Oct 2013.
<http://www.nngroup.com/articles/why-you-only-need-to-test-with-5-users/>.
Professor Espinosa
Middle Circle
17 October 2013
Proof in Computer Science: Usability
Proof is a universal idea that is necessary in providing evidence for credible claims.
While the general idea of proof remains the same everywhere, the proof and evidence found in
individual fields can vary widely. In the field of computer science, most methods of proof are
mathematical in nature. There are two main aspects of computer science offered at FIU: the
logical and scientific track that deals with the technical and mathematical aspects of computer
science, and the software development track that deals with software engineering (cis.fiu.edu).
The distinction provided by FIU accurately reflects on the field of computer science as a whole.
Most of the proof in the first field is highly technical, and while it is useful for experts in the
field, it is not useful for the general public. The use of computers by the general public has
grown exponentially in the last decade, and methods of proof that involve computer science
and the general public are found in the software development track of computer science. These
methods are generally more abstract and comprise of the involvement of the end user and the
effects of computing on them.
The method of proof that will be focused on is usability. Usability is a method of
determining how “usable” a computer system is. It measures the system’s effectiveness,
efficiency and the overall satisfaction of the user (usability.gov). Usability is measured primarily through
usability testing, which allows developers direct feedback on how real users use the system. This
Benitez 2
is an important aspect of software development because even though the program may work
completely and perfectly, if the user has no idea on what to click in order to make the program
work, has the program actually been completed successfully? Usability is the main way for
computer science to come away from the technical aspects and jargon and allow the
perspectives of non-experts to come into play. However, this also adds a great deal of
vagueness and difficulty when it comes to providing evidence. It is simple to prove that a
program works correctly. If a program is meant to find the average of 100 distinct values,
checking the output of the program to see if the average is correct would prove whether or not
a program works. However, if the program was written for use in a middle school Algebra class
and the user needs to convert every number entered into the program into binary code first,
the program has proven that it works, but it has not proven its usability.
The main challenge to this is simple: how do you objectively quantify something as
subjective as the end user experience of a product? We would never know with 100% certainty
that a product is free from usability problems, because that would involve the feedback of
every single user that will ever use the product. This is where the concept of the Rule of Five
attempts to draw a line and make usability testing more manageable. The Rule of Five basically
states that no more than five users are necessary in order to find an acceptable majority of
usability issues. The main idea behind it is that there are a finite number of usability issues that
can occur. When you have a small amount of users, each user’s data brings in more insight, but
there is inevitably going to be overlap in your findings. Usability is very costly and timeconsuming, especially when it is used to test many aspects and intended outcomes of products.
Therefore, the Rule of Five states that after five users, most of the relevant data on the usability
Benitez 3
of a product will have been found, and after five users the small amount of new data does not
outweigh the cost and difficulty of conducting the tests (nngroup.com).
The validity of this method of proof is difficult to say. It is my argument that usability is
not a valid method of proof, but it is acceptable for use solely because there does not seem to
be a sustainable and practical alternative in computer science. The Rule of Five is a great way to
make usability work in practice, but in theory, the Rule of Five is basically admitting that we will
never be able to truly prove that a program is usable, and gives us a guideline on where it is
acceptable to give up. This makes it easier to perform usability tests, but it draws from the
validity of usability as a method of proof. There are so many different factors when it comes to
usability that, in practice, it becomes impossible to keep track of them all. Users of computer
systems come from different age groups, different cultural and ethnic backgrounds, different
degrees of education, different levels of computer literacy, etc. Usability testers have to
consider their target audiences, and this becomes another excuse to let developers give up on
making a program usable to a certain demographic.
The concept of usability has many parallels to the scientific method. In science, when a
hypothesis is tested, it is never said that one has proven a hypothesis to be correct, but only
that one has failed to reject the hypothesis. Science is therefore a never ending quest to find
knowledge. This is why usability does not meet the same level of validity as a method of proof
as the scientific method does. Usability allows one to say when it is acceptable to stop looking
for problems, while science never gives up. When a counterexample is found in science, a
paradigm shift might occur in order to allow this new knowledge to come in. When a
counterexample is found in usability, such as Adobe Photoshop not being useful to the blind, it
Benitez 4
is simply disregarded because the blind are not part of the target audience. While in practice it
would be virtually impossible to make a photo manipulation program usable to the blind,
usability concedes to the fact that the hypothesis has been rejected, but opts to continue to
accept it. This is why usability is unacceptable as a method of proof, but is necessary for
computer science to function practically.
I took part in a usability test for a Technical Writing in Computer Science course last
semester, and the firsthand experience I gained through this serves as a great example for
usability. A usability test was conducted for the Society of Technical Communication Palm
Beaches Chapter, and the point of the study was to determine whether or not the website was
usable to new clients (sites.google.com). Five users were tested in accordance with the Rule of
Five, and most of the users agreed that the website was not usable to new users interested in
Technical Communication. There are many technical reasons as to why the website was not
usable, but relevant part of the usability study to this assignment is not what was presented by
the study, but what was rejected and was not included in the study. A majority of the users
were either interested in Technical Communication or were able to feign interest enough to be
able to find usability issues with the site. However, there was one troublesome user that could
not have cared less about Technical Communication and simply found the entirety of the
website boring and uninteresting. This data was ignored because the usability test was geared
towards individuals interested in the field, and therefore the lack of usability to this specific
user was not considered.
This specific example, however, only serves to prove that usability is not a valid form of
proof. The user was not impressed whatsoever by the website, and the website was in no
Benitez 5
means usable to her. However, instead of attempting to make the website usable to this user
and accept that the user has proven the website to not have usability, the user was completely
ignored and omitted from the findings. In reality, though, a website for a local branch of a
group of a very specific branch of computer science cannot be expected to woo uninterested
and apathetic college students of unrelated majors. Yet, the sheer fact that usability admits that
a claim is false and continues to state that it is true further proves that usability is not in itself a
valid method of proof.
Benitez 6
Works Cited
"Palm Beaches Chapter." STC PBC. N.p. Web. 17 Oct 2013.
<https://sites.google.com/site/stcpalmbeaches/>.
"Undergraduate Programs." School of Information and Computing Sciences. N.p. Web. 17 Oct
2013. <http://www.cis.fiu.edu/programs/undergrad/>.
"Usability Evaluation Basics." Usability.gov. N.p. Web. 17 Oct 2013.
<http://www.usability.gov/what-and-why/usability-evaluation.html>.
"Why You Only Need to Test with 5 Users." Nielsen Norman Group. N.p. Web. 17 Oct 2013.
<http://www.nngroup.com/articles/why-you-only-need-to-test-with-5-users/>.
