As the BLM movement gains some much needed momentum, we have some terminology that we are overdue for replacing. Fortunately, there’s lots of great work happening on this already. Unfortunately, everything I’ve seen has been happening in isolation and different projects are coming to different answers, which undermines one of the key advantages of the old terminology: It was universally understandable regardless of context.
I’ve created a survey to understand what is most natural to the most people across different contexts so that we can have a resource to point to when making these decisions. My hope is that if we can nail this, it will be one less point of friction to making the change happen.
It would be really cool if you could contribute your opinions. And sharing it to get a wider audience would be amazing.
We should also recognise that there are several projects that have been sensitive to this from the beginning, or made the change when they became aware of it.
[Update 2020-08-23: Re-worded the intro to make it more approachable.]
As we scamble to do something about this, something is kind of baffling to me. All of the conversations I’ve seen about what terminoloy to use instead are isolated conversations on a project-by-project basis and the results are often not the same. One of the things that worked really well with the old M/S terminoloy was that you could go from any context, to any context, and have a pretty good idea what those terms meant in the new context without knowing anything else about the project. Yet the solutions that are floating around, are all over the place. Eg
Can you tell immediately which ones are in which roles? Some of them are pretty easy to take a good guess at while others are less so. More importantly, what happens when you are talking about one project (Eg MySQL), and in the second half of the sentence, you are talking about another project (Eg rabbitMQ) explaining how their activities related to each other in your infrastructure? For an SRE/DevOps/Sysadmin, it’s not uncommon to talk about 3 or more technologies in a single sentence when discussing how things fit together. Do you respect the terminoloy of each, and get it right first time, every time? Now we’re talking about another project (Eg Redis). How are you doing? Are you keeping up? You’ve just realised you messed up a couple of sentences back. But your collegues understood you, …right? How many other places did you make a mistake? What happens when the results of this conversation get in to production? Are we going to be writing to a replica, that is essentially a black hole from the perspective of the validator, er um, leader?
As you look at a single example like this, it’s easy to think “well, someone will spot it and question it or fix it.” But multiply this across the thousands of assumptions we have to make about out infrastructure as we build and maintain it, and something is going to break badly. The consequences vary from “whoops that’s embarrasing, but we caught it in time” through to data that is never backed up and lost forever, because it was never in the correct state, or a data leak that exposes private customer data, and ultimately destroys the reputation, and later the viability of the business.
We need consistency of these terms. Disagree? There’s a place for that in the survey below. This is the only question where I’m telling you my opinion before you get in to it. You don’t have to agree, but if you don’t, I’m really interested to see your reasoning in the comment box. I do think that we should give a free pass to projects that are already using alternative terms. They have lead the way, and we should give them credit for that…. Ok two questions! ;-)
The primary focus of the survey is to get an idea of which terms we should be focssing on. And give projects/teams something they can point to when choosing their terminoloy.
There are a few questions where there is a list of things to choose from. I’ve tried to think of all the sensible answers. But I will have missed some. So in each case, there is an “other” option where you can specify your suggestion. Particularly in the early weeks of the survey, I’ll check these, and add them to the options to prevent spelling mistakes from creating duplicates.
The survey is here.
I’ll post updates on the results here every every month, or more often if there is reason to do so. My intention is to run the survey for a few months, but I’ll see how it’s going. If there is still lots of genuine activity beyond that, I’ll likely keep it going a while longer.
I want to have a resource that I and others can point to: “Here’s what other people seem to be thinking, let’s do that” with the goal that the industry as a whole will eventually land on one standard rather than be fragmented by tool or context.
People working in tech, with technologies that have a relationship between two or more roles where either the “master” or “slave” terms have typically been used.
I’m not collecting, or interested in, any private or contact information.
You need to be logged in to google for the google form to determine that you haven’t already submitted an answer. But I will not receive this information in any way.
While creating the Google Form, click the Settings bar and turn on the option that says ”Allow only one response per user.” When the unique option is enabled for a Google Form, respondents will have to sign-in with their Google account to access the form. Their email address won’t be recorded in the response sheet but Google Form will not allow another entry from the same Google Account.
This does mean that the survey is restricted to people who have a google account. But considering the target audience for this survey is the tech industry, most of which will already have a google account, or can easily get one, I’m ok with that.
There are some optional demographics questions at the end.
For the Black Lives Matter movement, http://antiracismforbeginners.com/ has links to lots of resources for how you can get involved.
For this survey, please share it https://forms.gle/7Feaszh2NbYZqgop8 or this blog post http://www.randomksandom.com/betterTerms/ . The more responses we can get, the more statistically significant it will be.
The title photo is of a floppy drive cable from the late 80s/early 90s. There’s a bit more that’s interesting to see about this cable in that the drive B connectors are out of view of this photo. But I wanted to focus on the twist, which is what was used to denote which set of connectors represented drive A, and which connectors represented drive B.
The two drive A connectors shown here are
I have seen exceptions to those usages, but they were incredibly rare.
Nostalgia nerd has a really good write up about why there is the twist and how it works. But for the relevance to this blog post, is that it denotes which set of plugs go to a primary thing, and which go to a secondary thing.