Pushing The Button
This is a retropost. Written contemporaneously in 2020, but published four years after the events. It's May 2020 as I write this. I'm typing to capture the moment. Right now, I've no idea what the impact is.
This is the exact moment, on Thursday May 7th, I hit the Big Red Button - three of them! - to open source the UK's COVID-19 Beta test app.
It was thrilling and terrifying. We'd spent the last few weeks getting ready to open source the repos and then, at the last minute, it all went wrong. The plan was to launch on Tuesday - but fate conspired against us.
The problems fell into three main areas:
-
Threats and personal safety. This was probably the highest profile code release that we'd ever done. There were already people grumbling online that the people writing the code were "traitors". Did we want to expose our people to that sort of personal abuse? What if they were targets of phishing attempts?
-
Redacting history. Probably the most contentious issue. We all wanted to release everything from the very first commit. Would that reveal anything dangerous? Had someone slipped and accidentally committed an API key they shouldn't?
-
Communications. The other most contentious issue! The department were in "crisis comms" mode. Everything was delayed. No one had reviewed the blog I'd written, there was no pre-arranged plan in place for this sort of thing. Understandable really - this was a tiny piece of a much larger puzzle. But it was still frustrating to wait for people to be ready for us to publish.
We took the pragmatic approach. We took a snapshot of the code, thoroughly scrubbed it of all identifying information and secrets, and prepared to release it. Then we waited. And waited.
Every time we thought we had the go-ahead, there was another delay! There was a strict comms schedule. We couldn't launch now; it would interrupt that other announcement!
I was asked to help rewrite bits of the announcements. This led to some memorable questions from the comms squad. How can you explain to the average user...
- what "Source Code" is?
- why the Android code is different from the iPhone code?
- who are "Git Hub"?
And, the kicker? All these questions came in while I was on a conference call with a bunch of government ministers! The joys of multiple monitors!
It was interminable. I sent texts which went unanswered. Emails. Phone calls. Just a few minutes more. Any moment now. We need to wait for...
And then!
"Can we launch ASAP?"
Yes! The email I was waiting for. But I am a paranoid and cautious Fraggle. Was that "Launch now!" or "Can we launch now?"?
So I sent a reply. "Just to confirm - do you want me to publish now?" And waited.
And waited.
I got an email from my boss "Launch now!"
And a second later, from comms: "Please hold off - no go. Will call you shortly."
How I longed to press that button. I could say that I only saw the first email... No. Maybe. No.
An eternity. During which time I casually glanced at Twitter and read all the angry messages from people demanding the release of the code.
The call came. "Publish it - but don't tell anyone." Weird flex, but OK.
I called my very-patient wife into my home office. I wanted the moment captured. She opened her camera. A few clicks, and it was done.
I did a little dance. Let all of the tension out of my body. And waited for the hate to roll in.
It didn't. The response was... positive! Yes, there were grumbles, but so many people were fulsome in their praise that it was overwhelming. Congratulatory tweets and emails did the rounds, and I had a nice cold ale.
I took the bank holiday weekend off. Well, I obsessively read all the tweets, answered questions about my blog post, and kept half-an-eye on GitHub. I'm not good at relaxing.
Has it worked? Did we make the NHS more open and transparent? Did open source win the day? Did the beta test work? Were lives saved? Or was it a damp squib?
As I write this, we're still in the eye of the storm. Perhaps, when this post is published, we'll know the answers.
Ben Smith said on social.lol:
@Edent I bloody love these posts.
Dr. David McBride said on mastodon.social:
@Edent At least in my bit in the world, this was incredibly valuable work. When the app became available to the general public in September, it had already been subject to detailed peer review. And that built trust, which was vital because one big hazard at that point was trust in government.
That peer review meant that techies like me could tell their mates that they didn't need to worry, that installing the app was both safe and helpful.
You enabled that. 🙂
frankieroberto.github.io said on frankieroberto.github.io:
Week 4: Meeting the users
Matt Jukes said on newsletter.digitalbydefault.jobs:
Internet of Public Service Jobs is a weekly list of vacancies related to product management, agile and design in…you guessed it…‘public service’ curated by @jukesie
More comments on Mastodon.