Plaintext
Teaching Erlang through the Internet:
An experience report
Stephen Adams
sa597[AT]kent.ac.uk
School of Computing
University of Kent
Today functional programming languages are seen as a practical solution to the difficult problems of
concurrent and distributed programming. Erlang is a functional language designed to build massively
scalable and fault tolerant applications. This paper describes the authors’ experiences delivering a
massively online open course (MOOC) on the FutureLearn platform.
1 Introduction
The unexpected lifespan of Moore’s Law seems to be coming to the end, soon programmers will no
longer be able be rely on increases in individual processor speed to power increasingly complex soft-
ware [6]. Fortunately at the same time multicore and multithreaded technology is on the rise. Imperative
languages with their implicit global state and low level concurrency primatives make writing concurrent
programs very difficult.
Functional programming languages offer solutions to these problems. However computer science
education is predominately based around imperative languages. In 2014 of the 39 top undergraduate
Computer Science programs only five taught a functional programming language to their first year stu-
dents [4]. Many learners are turning to alternative education platforms to discover functional program-
ming.
1.1 Erlang
Erlang1 is a functional programming language designed to be massively scalable and highly fault tol-
erant [7]. It was originally developed in 1986 by Joe Armstrong, Robert Virding, and Mike Williams
at the Computer Science Laboratory at Ericsson Telecom AB [2]. Erlang’s core design tenets include
lightweight process, that communicate through message passing. Erlang also boasts a “let it fail” er-
ror handling strategy, where processes either succeed or fail and other specialised processes handle the
error [1].
Packaged along with the Erlang language is the Open Telecom Package or OTP. OTP is a set of
libraries and design principles for developing distributed, concurrent, and fault tolerant systems [7].
OTP has been described as a large part of the reason Erlang is so good at these domains [5].
1.1.1 Erlang at the University of Kent
Though the University of Kent’s introductory programming modules are taught using Java, “Functional
and Concurrent Programming” is a mandatory stage 2 Spring term module taught since 2014. CO545’s
1 Named after the Danish mathematician Agner Krarup Erlang the inventor of queueing theory.
c Adams
Submitted to:
This work is licensed under the
TFPIE 2017
Creative Commons Attribution License.
2 Teaching Erlang through the Internet
principle language is Erlang, however Haskell is also covered to a reasonable degree towards the end of
the course. Erlang was chosen as the principle language because of its relatively small syntax and its
message passing concurrency makes a stark contrast to Java’s threads.
Additionally since teaching Erlang makes Kent a bit unusual our students are not only well trained
in principles but are also capable of working with Erlang during their year in industry placements. Some
companies have begun taking our students for this reason2
In early 2017 the authors had the opportunity to supervise two massively open online courses (MOOC)
on the online learning platform FutureLearn, Functional Programming in Erlang 3 which took place from
February 20 to March 10 2017 and Concurrent Programming in Erlang4 (April 3 to April 21 2017). This
paper will describe our curriculum and approach to teaching functional programming in a MOOC context
in section 2. It will also cover our experiences teaching the courses and who participated.
2 Curriculum and Pedagogy
Our first attempt at teaching a MOOC happened in the Spring of 2015 as part of the University of Kent’s
“Beacon Projects”5 which were a part of the University’s 50th anniversary celebrations. We ran a three
week pilot program. This pilot MOOC was hosted through the University’s Moodle site. The pilot
course’s materials were adapted from the first part of CO545. The pilot course involved short, two to
twenty minute video lectures, along with assignments for students to complete. We also brought in Joe
Armstrong co-creator of Erlang, and Frencesco Ceesarini the technical director of Erlang Solutions to
produce “master classes” which focused on describing how Erlang was used in the real world. The
master classes were made available to the students at the end of the pilot course6 .
We recruited participants in the pilot MOOC from the University’s current computing students and
through social media associated with the Erlang community. Overall over five hundred people signed up
for this pilot course. At the end of the pilot course a few things became clear based on feedback we had
from our learners. Moodle, though not designed for the delivery of MOOCs was did an adequate job.
Most of the people we surveyed were interested in taking another MOOC course on Erlang regardless of
what platform it was delivered on. At the same time learners indicated that social learning via discussion
was important and a dedicated MOOC platform would have more features to support this type of learning.
Based on our experience giving this pilot MOOC we decided to develop a further three weeks of
material and move our courses onto the FutureLearn platform.
Functional Programming in Erlang, the first course on FutureLearn, covered the same materials as
the pilot course but we were able to expand it to use FutureLearn’s support for quizes, tests, and peer
assessments. We also developed a second three week group of material, meant to follow the first course,
about Erlang’s concurrency and fault tolerance features.
2.1 The curricula
Functional Programming in Erlang was designed with someone who was familiar with at least one other
programming language though not necessarily a functional one. Our goal then for this first course was
2 The web security company Alert Logic for example: https://www.alertlogic.com/
3 https://www.futurelearn.com/courses/functional-programming-erlang/1
4 https://www.futurelearn.com/courses/concurrent-programming-erlang/1
5 https://www.kent.ac.uk/beacon/about.html
6 See: https://goo.gl/WxDHiB, https://goo.gl/RITxtd, and https://goo.gl/Wazkuc
Adams 3
to give someone unfamiliar with functional programming a basic introduction to Erlang and make them
ready to tackle the second course on Erlang’s concurrency features.
The three weeks of the course are:
• Getting started with Erlang
• Lists in Erlang
• Advanced functional programming
The first week began by introducing Erlang in fairly abstract ways. We covered some of its history
and the ways that Erlang (as well as many functional languages) differ from imperative programming,
evaluating expressions and immutability, versus procedural steps mutating a global state. We then cov-
ered the basics of Erlang’s syntax, including its basic data types and how pattern matching occurred. The
final activity in the first week introduced the idea of functions and recursion (both head and tail).
The second week was focused entirely on lists. Learners were first introduced to list syntax and how
lists were constructed, pattern matched, and used in functions. Once they became familiar with the list
syntax we looked at common strategies for defining Erlang functions. Finally they were tested on their
knowledge with an assessment.
The final week introduced the learners to higher ordered functions and the week’s second activity
involved modelling the game rock-paper-scissors. The final activity of the course linked to the master
classes so that people could continue their learning if they wanted and asked what people’s general
opinion of Erlang was.
Following a three week break the second course on concurrency was run. Like the first course
Concurrent Programming in Erlang was three weeks long. This course’s outline is:
• Concurrency - nuts and bolts
• Concurrency - making code robust
• Scaling it up
The first two weeks cover Erlang’s built in features and how these features affect the design of
concurrent systems. The third week on scaling systems introduces the learners to the Open Telecom
Platform (OTP) the set of middle-ware libraries for developing highly distributed systems that is included
with Erlang [7].
The first week explained Erlang’s concurrency system, the actor model. Processes are lightweight
to create and destroy and communicate with each other through message passing. Erlang’s messaging
system is asynchronous, each process has a mailbox and will handle messages in the order they are
received. The second week then covered Erlang’s unique error handling philosophy, “let it fail.” Erlang
processes are expected to either succeed and if they can’t they should fail immediately without doing any
error handling [1]. Specialized supervisor processes should be designated to monitor other processes.
When children processes fail the supervisors can then take the appropriate action, such as error reporting
or simply restarting the failed process.
The course finished with a bit of material about OTP. Many of the concurrent design patterns we
had our students implement in the first two weeks of the course are actually standard templates that OTP
defines.
4 Teaching Erlang through the Internet
Continent FP Erlang Concurrent Erlang
Africa 438 84
Asia 1,122 291
Australia 115 43
Europe 2,504 970
North America 1,008 406
South America 392 145
Table 1: Enrolment by continent
2.2 FutureLearn’s Course Creation
FutureLearn provides many ways to configure and deliver online courses. Every future learn course is
broken down by week, and each week is further broken down into “activities.” Each week is used to
suggest the pace that learners should aim for, and activities organise steps. Steps are the smallest section
of a course.[3]
There are several different types steps that a course is build from. Each step typically hosts some
materials and a discussion section for learners to have a conversation about that step’s content. Broadly
speaking steps can be divided into two different types which could be described as teaching steps and
doing steps. A teaching step’s primary purpose is to convey information to the learners, textual articles
and audio/video steps are examples of this. Doing steps are more interactive. They actively work to
engage learners to do something, whether that be participate in a discussion, take a multiple choice quiz
or test7 , do an exercise, or submit an assignment for peer review and review other learner’s work. Every
step type is designed to support active learning. Doing steps require some learner interaction to pass,
however, in teaching steps the learner may optionally engage in the discussion but this isn’t required to
complete the step.
3 Who participated
In the end 5,642 people enrolled in Functional Programming in Erlang and 1,965 people enrolled in
Concurrent Programming in Erlang. The majority of the learners came from the UK and the USA but in
total 149 countries were represented.
Where users were located was determined based on their IP addresses. A survey was also sent out to
everyone who enrolled on the course and from the people who responded we can roughly determine the
ages of our learners as seen in table 2.
3.1 Prior experience
Beyond the basic demographic information we prompted the learners to introduce themselves on the
first step of each course and to briefly describe why they were taking the course what their previous
programming experience was. In over 300 responses to this question we can determine that the learners
were familiar with over 40 different programming languages. The top ten languages that people claimed
to have used before are shown in table 3
7 Quizzes allow for infinite attempts whereas learners only have three attempts at each test question
Adams 5
Age Range FP Erlang Concurrent Erlang
<18 6 2
18-25 82 29
26-35 130 54
36-45 77 41
46-55 64 41
56-65 40 20
>65 18 6
Table 2: Enrolment by age
Language # of People Language # of People
C/C++ 52 Haskell 26
Java 51 Erlang 11
JavaScript 49 Scala 11
Python 41 Elm 9
Ruby 31 Clojure 8
Haskell 26 F# 4
PHP 23 Scheme/Racket 4
C# 14 ML/OCaML 2
Erlang 11 Idris 1
Scala 11 Emacs Lisp 1
Table 3: Ten most popular languages Table 4: Functional Language Popularity
Our learners’ experience seems very much in line with broader industry trends. Seven of the lan-
guages in table 3 also appeared in the ten most popular programming languages in the May 2017 TIOBE
index [8]. The three exceptions to this are Haskell, Erlang, and Scala who’s TIOBE ranks are 38th, 41st,
and 31st respectively.
It makes sense that our learners would have a significant interest in functional programming lan-
guages prior to joining our course. Table 4 shows the number of people with experience in functional
languages only. A very common reason why learners wanted to take our course was that they wanted
to improve their understanding of functional programming techniques independently of the language the
course was taught in.
4 What happened
This section will describe our experience running both of these courses. We had previously ran the pilot
MOOC so we weren’t completely new to the concept of MOOCs but this was our first time using the
FutureLearn platform in particular and these courses were an order of magnitude larger than our pilot
course.
6 Teaching Erlang through the Internet
4.1 Participation
If you don’t have experience running or taking an online course before it may seem unmanageable for
two people to run a course with over 5,000 people in it, but of those people enrolled only a fraction
even begin the course and a fraction of those finish it. FutureLearn categorises learners by how far they
progress through and how they interact with the course. The categories are:
• Joiner - Total number of people enrolled on the course
• Learner - Joiners who have at least viewed one step
• Active Learner - Learners who have completed at least one step
• Social Learner - Learners who have left at least one comment
• Fully Participating Learners - Learners who have completed all tests and at least 50% of the steps
We have summarized how many people fit into each of these categories for our two courses in table 5.
Learner Category FP in Erlang Concurrent Programming Notes
Joiners 5,642 1,965
Learners 3,858 1,117 68% & 57% of joiners
Active Learners 2,683 676 70% & 61% of learners
Social Learners 586 142 15% & 13% of learners
Fully Participating Learners 374 40 10% & 4% of learners
Table 5: Learners in each category
Even though relatively few people comment it’s still too much for our small team to respond to every
question. We actively encourage learners to answer each other’s questions and give informal feedback
about each other’s exercises. We would only comment when something needed to be said about the
course materials or there seemed to be a widespread misunderstanding though this was fairly rare.
One thing we noticed from the pilot MOOC is that it was difficult for us to participate in discussions
without stopping the discussion8 . This time we decided that at the end of each week we would record
a short video talking about issues that came up throughout the previous week. This meant that the
discussion sections were for learners to interact with each other and our thoughts would come through a
different medium.
4.2 Learner Feedback
After each exercise and assessment and in the final step of both of the courses we requested feedback so
that we can continue to improve the materials for these courses and our approach to MOOC creation in
general. People were more than happy to give us feedback at most opportunities.
4.2.1 Workload
Both of the course descriptions estimated that learners should spend about five hours a week to com-
plete the course. It seems that most people spent a lot more time than that. Interestingly this was seen
8 Maybe people felt that our comments “settled” the issue?
Adams 7
as a positive thing during the functional programming course but more of a negative thing during the
concurrency course.
When designing the second course we wanted to make the exercises much more open ended so that
learners could explore more. Many of the assignments and examples in the second course were centered
around a “frequency server” example. The server would allocate and deallocate numbers when they were
requested through message passing. Throughout the course the frequency server was expanded through
both the video lectures and student’s exercises.
During the second week we wanted the learners to attempt to build their own version of a supervisor
process that would clean up the system if the server failed. Up to this point we had spoken in fairly
abstract terms about supervisors.
The feedback we got was that this exercise was much to open ended. Students were unsure what
we really wanted and without previous concrete examples of supervisors they were left aimlessly trying
things not sure what we wanted.
This seemed to be the key difference between the workload feedback we received in both courses.
The first course had a high workload but learners felt that they were always working on the right thing,
it just took longer than expected (they just asked that we update our estimates). For the second course
learners didn’t enjoy just trying things without knowing if what they had was a good solution.
5 Conclusion
Overall people were very positive about their experiences with the course. This is of course tainted by
survivorship bias because we hear from people who completed the course (or at the very least looked at
the final step). The percentage of learners who ended up in each category from our courses was lower
than the average all FutureLearn courses as seen in table 6.
Category FutureLearn Average FP Average Concurrency Average
Learners 50% of Joiners 68% 57%
Active Learners 81% of Learners 70% 61%
Social Learners 38% of Learners 15% 13%
Fully Participating Learners 21% of Learners 10% 4%
Table 6: Average number of learners in each category
Our courses had more people become learners (those that at least looked at a step) than FutureLearn
as a whole but the number of people who went on to either complete a step (active learners) or comment
at least once (social learners) was lower than FutureLearn’s averages. This may mean that the workload
is turning people away.
5.1 Future Work
The functional programming course was popular enough that many people requested we run it again. We
have scheduled another three week offering of the course to begin on the 29th of May 2017.
A much requested addition to our course was more information on the OTP framework. Given that
both our courses had high workloads we have begun planning a third course specifically about OTP. We
are currently investigating developing this course along with an industrial partner and we are discussing
with FutureLearn about incorporating these three Erlang courses into a single “program.”
8 Teaching Erlang through the Internet
Today functional programming is being seen, more and more, as a practical solution to design of
highly distributed and concurrent programs. Many traditional educational institutions are still teaching
imperative programming first and so people are turning to alternative opportunities to augment their
knowledge. In this paper we have described our methodology and experience teaching two massively
online open courses with the functional programming language Erlang.
Acknowledgements
The authors would like to thank Mark O’Connor, Distance Learning Technologist at the University of
Kent for his partnership throughout this entire enterprise. His advice and mentoring on preparation and
recording as well as editorial and production help made these courses possible. We also would like
to thank Claire Lipscomb our contact at FutureLearn for her dedication and prompt responses to our
questions.
Finally we must thank all of the people who have participated in all of our MOOCs. Their hard work
and honesty is much appreciated.
References
[1] Joe Armstrong (2003): Making reliable distributed systems in the presence of sodware errors. Ph.D. thesis,
The Royal Institute of Technology Stockholm, Sweden.
[2] Joe Armstrong (2007): A history of Erlang. In: Proceedings of the third ACM SIGPLAN conference on
History of programming languages, ACM, pp. 6–1.
[3] FutureLearn (2017): FutureLearn Partners. https://partners.futurelearn.com. Accessed: 2017-05-
13.
[4] Philip Guo (2014): Python is Now the Most Popular Introductory Teaching Lan-
guage at Top U.S. Universities. https://cacm.acm.org/blogs/blog-cacm/
176450-python-is-now-the-most-popular-introductory-teaching-language-at-top-u-s-universities.
Accessed: 2017-05-13.
[5] Fred Hebert (2013): Learn You Some Erlang for Great Good!: A Beginner’s Guide. No Starch Press, San
Francisco, CA, USA.
[6] Tom Simonite (2016): Moore’s Law Is Dead. Now What? https://www.technologyreview.com/s/
601441/moores-law-is-dead-now-what/. Accessed: 2017-05-13.
[7] Ericsson OTP Team (2017): Erlang. http://www.erlang.org. Accessed: 2017-05-13.
[8] TIOBE: TIOBE Index for May 2017. https://www.tiobe.com/tiobe-index/.