March 30, 2021 — The CDC needs to move to Git. The CDC needs to move pretty much everything to Git. And they should do it with urgency. They should make it a priority to never again publish anything without a link to a Git repo. Not just papers, but also datasets and press releases. It doesn't matter under what account or on what service the repos are republished to; what matters is that every CDC publication needs a link to a backing Git repo.
Git is the "Global Information Tracker". It is software that does three things that anyone can understand 1) git makes lying hard 2) git makes sharing the truth easy 3) git makes fixing mistakes easy.
Because the CDC's publications are currently full of misrepresentations, make it very hard to share the truth, and are full of hard to fix mistakes. Preprints, Weekly Reports, FAQs, press releases, all of these things need links to the Gits.
The whole world now builds on Git. The CDC is far behind the times. Even Microsoft Windows, the biggest proprietary software project in the world, now builds on Git.
Git is an open source, very fast, very powerful piece of software originally created by Linus Torvalds (the same guy who created Linux) and led by Junio C Hamano that makes extensive use of principles from blockchain and information theory.
The CDC's GitHub account has 169 repos and 10 people (and I'm told many hundreds more Git users). I would immediately promote every person working on these repos. (There are probably one or two jokers in there but who cares, it won't matter, just promote them all). Give them everything they need to be successful. Give them raises. Tell them part of their new job is to get everything the CDC is invovled with published to Git. This is probably really the only thing you need to do, and these people can lead it from there.
Provide a hard deadline announcing that you will stop all funding for any current grant recipient, researcher, or company doing business of any kind who isn't putting their sponsored work on a publically available Git server and linking to it in all releases.
The CDC has 10,600 employees, so buying them all 20 worth of great paper books on learning how to use Git would only cost 201,200. For the most part, these are highly educated people who are autodidacts and can probably learn enough with just some books and websites, but for those who learn better via courses or videos you can budget another $30 per person for those. Then budget to ensure everyone is paid for the time spent learning. We are still talking about far less than 1% of the CDC's annual budget.
Because the CDC not only failed at it's mission by not stopping COVID, but it continues to mishandle it. Mistake after mistake. Miscommunication after miscommunication. I just shook my head looking over an amateur hour report that they just put out. It's sad and their number one priority should be to regain trust and to do that they need to focus on the most trustworthy tool in information today: git.
I'm adding two very clear and specific examples to illustrate the problem. But my sense is the problem is prolific.
For young children, especially children younger than 5 years old, the risk of serious complications is higher for flu compared with COVID-19. @ CDC
This statement appeared on the CDC's website for more than a year. As it should have. Every big dataset I've looked at agrees with this, from the very first COVID-19 data dump in February 2020.
I started actively sharing and quoting that CDC page in August 2021. Coincidentally or not, within days they removed that quote. There is no record of why they made the change. In fact the updated page misleadingly states "Page last reviewed: June 7, 2021", despite the August edit*.
To recap, they quietly reversed the most critical piece of contextual advice on how parents should think about COVID-19 in relation to their children. No record, no explanation. (In case you are wondering, the data has not changed, and the latest data aligns with the original statement which they removed. Perhaps the change was made for political reasons).
The second example is well documented elsewhere, but the CDC changed their online definition of the word "vaccine", again perhaps for political reasons. That sort of thing seems like the kind of change that maybe should have some audit trail behind it, no?
I used to take it for granted that we could trust the CDC. That made life easier. Health is so important but so, so complex. I would love to trust them again, and would have more confidence if they were using the best tools for trust we have.