Thread by @astrofrog

Researchers who use Macs, PLEASE don't upgrade to MacOS X Big Sur or buy a new M1 mac yet: a short thread. (1/n)

I know it's hard to resist when a little pop up shows up to suggest upgrading to the latest MacOS X, and oooooo shiny new laptops, maybe I can replace my obsolete 7 month-old laptop (2/n)

(You were also probably looking forward to being able to spend a couple of work hours avoiding work while your computer updated, but it's ok, I won't tell 😉) (3/n)

But here's the thing: unless you mostly use your work laptop for browsing the web and doing emails, you probably use a whole bunch of more advanced software, including (I am guessing) some open source software. (4/n)

I'm going to bet that the very large majority of maintainers of open source software haven't had advance access to either Big Sur or new macs with M1 chips, both of which are significant updates/changes. I certainly haven't been able to test any of the packages I maintain. (5/n)

I know of several major Python visualization packages that in fact do not work at all on Big Sur at the moment, and users are running into issues even just installing Numpy in certain cases (github.com/numpy/numpy/is…) (6/n)

Python wheels (binary releases) of most packages won't work at all on the new M1 macs, and you'll find yourself suddenly having to build many Python packages from source. (7/n)

If you upgrade now, you will likely be frustrated at the number of things that will break or not work anymore - and you will likely be angry at the wrong people, by blaming the maintainers of the packages you use. (8/n)

If you want to blame anyone, blame Apple for yet once again creating a lot of extra work for developers - especially for people who volunteer time to maintain open source projects. (9/n)

Your research computer should be a stable work environment. There is very little in Big Sur that you actually NEED to be a more productive researcher. So resist the temptation to upgrade until you know that your favourite research software will actually work on it. (10/n)

Likewise, while the new Macs are supposed to have blazing fast processors, things are only faster if they work, and some of your favourite research software won't work. (11/n)

So the bottom line is to wait a bit - at least a few months, if not more. Maybe a better way to think about it is to update once you actually need to, rather than when things are new and shiny! (12/n)

That's all I had to say. Thank you for coming to my TED talk. (13/n)

Epilogue: yes, I know that upgrading to Big Sur might work just fine for some, but I also know that it has broken things for others. The key message here is that by upgrading *this early* you are taking unnecessary risks (and this is true of every major OS X version)

And yes I know that upgrading to the new Apple hardware is far more risky - in fact it is almost guaranteed to not work smoothly unless all you do is browse the internet and check emails!