Why is UoM not dumping Master-Slave terms in computer science?



If you’re not involved in computer science in some way, you might not know that it’s common for the term “master/slave” to be used to refer to a variety of concepts in computer science. The term has a range of uses, but one of the main ones is where there is a primary copy of some data (the “master”) and various replicas of that data (referred to as “slaves”). When I was taught to use this term as a computer science undergraduate in 2010 I was taken aback. It seemed unclear and unnecessarily upsetting. Why were we making casual references to the horrific practice of slavery in a lecture on databases?


The origins of the use of “master - slave” as a technical computer science term are unclear. The first use I could find was in a paper from 1959, which uses it as a metaphor alongside the more self-explanatory “primary - secondary” when discussing a method of linking two computers together. It comes up again in 1964 when researchers at Dartmouth College (an Ivy League university in the United States) use it to describe part of their system for allowing different parts of a computer to communicate with each other, known as the Dartmouth Time Sharing System. Then, in 1996 a company called Scientific Data Systems released a new computer, the Sigma 7. At the time there was only one competitor in terms of computing power, the IBM 360, so the Sigma 7 was a big deal in the field of computer science. The term “master - slave” was again used to describe the way the Sigma 7 dealt with allocating resources, possibly due to familiarity with the term from the Dartmouth System. It has been used in the field of computer science ever since, for an increasingly broad range of concepts.


The earliest widely reported pushback was in November 2003, when the Los Angeles County Office of Affirmative Action Compliance sent a request to equipment manufacturers that they purchase from not to use the words “master” and “slave” in reference to computer hardware. A representative of the office described “an avalanche of complaints from vendors and the general public” including emails telling him he should be fired.


In 2014 Django, a popular technology used in web development replaced uses of “master/slave” with “leader/follower” and received a lot of complaints about this change, in which the phrase “political correctness” is predictably prominent. Python, a popular programming language, removed the term in 2018 to similar backlash. In July 2020 the term was replaced by “primary/replica” by the developers of MySQL, a widely used database program.


Some small and hard fought steps are being taken towards removing the word “slave” as a term that is casually peppered throughout technical documents, but many in the field of computer science still cling to it. “It’s a clearer metaphor”, they wail (it isn’t). “Removing it is a waste of time”, they spend countless hours arguing. “It’s a technical term that we can’t change”, they say to the people currently changing that term. “You’re just being politically correct”, they type over and over, as if something being blatantly racist isn’t a good enough reason to make a tiny, easy change.


How about the University of Manchester? At least one lecturer in the Computer Science department is still teaching the term “master/slave”. After reaching out to the UoM Equality and Diversity team regarding the issue but receiving no response, I wrote the lecturer an email asking them to consider teaching a different term, along with a link explaining its problematic use. The response I received stated “this is a terminology that is used in the field,” as if that leaves them powerless to teach one of the other terminologies also used in the field. Removing this terminology from our vocabulary is nowhere near the only thing that needs to be done - it is a symptom of the larger problem of racism and colonialism which permeates the institution of academia. Professor Nalin Thakkar, Vice President for Social Responsibility said in a statement on Black Lives Matter that  “Our role in addressing issues of race and ethnic discrimination is therefore not a neutral one – the knowledge we produce and share with others will play a profound role in achieving the sort of change that is necessary.” Why insist on using “slave” rather than terms that are clearer and more neutral? We hope the computer science department at the University of Manchester will reflect anew and take the opportunity to make a positive change. 




 Useful Resources:


https://www.cnet.com/news/master-and-slave-tech-terms-face-scrutiny-amid-anti-racism-efforts/


https://cdm.link/2020/06/lets-dump-master-slave-terms/


https://www.wired.com/story/tech-confronts-use-labels-master-slave/


https://www.vice.com/en/article/8x7akv/masterslave-terminology-was-removed-from-python-programming-language








Comments

Popular posts from this blog

Establishing and holding on to anti-racist space in the University of Manchester: The Ahmed Iqbal Ullah RACE Centre

Resisting Carceral Feminism

Decolonise Our School of Arts, Languages and Cultures