Yes

No.

More specifically, if you're willing to hand-wave away certain things as being outside of the "center of the open-source software-release ritual", then you can get a 'yes'.

But in addition to source-distributions mentioned in the essay, which are "too tiny a minority", two other issues are:

1) If you define "git" as "pervasive" then you can declare that git is the solution. In my field, some of the tools are still only accessible via tarball/zip. Two examples are the IUPAC InChI distribution, from http://www.iupac.org/home/publications/e-resources/inchi/dow... , and VMD from http://www.ks.uiuc.edu/Development/Download/download.cgi?Pac... . PyMol, a more popular free software visualization package, is also not developed using git.

2) If you define "open-source software-release" to exclude fee-based free software distribution (perhaps it it also "too tiny a minority"?), then it preferentially excludes certain flavors of 'open-source software'. For example, I write free software, and distribute it for money. A tarball is an easy way to identify the delivery of the contractually obligated product. I can also send it via email, vs. setting up a private git server and getting accounts set up for my customers.

How much free software development is part of the hidden world of non-publicly accessible development? I don't think anyone really knows.

So if your free software baseline assumes "pervasive git", a public development repo, and no cost to access the code, then sure, the answer is a "yes". Until then, it's a "no."

And my examples show that baseline is only a subset of free software development, with no clear idea of how large it is.