What is Modal?

A quibble with an otherwise good article:

is free software just structurally unable to break through and build solutions that can win?

History has shown that, in the right circumstances, it can win. ...

In about a decade they went from there not being a usable free software operating system to GNU/Linux being a competitive alternative that ended up taking over the server OS market.

I think Free software is bottlenecked on design, in contrast to implementation ...

i.e. I think there is still truth to the assertion that "open source is good at copying/implementing a design; not at coming up with one"

Look at say Forgejo -- it didn't create Github, but it seems to be very successful at copying it (which is good! I want to use it)

---

By the time Linux was being developed, Unix was already a 20+ year old design. It was proven to be successful

Linux was a project that added wide hardware support and a distributed development model to Unix. It didn't do much OS design, and some of the designs it did are pretty bad (arguably worse than the state of the art, and require multiple incompatible mechanisms to converge on approximately right, like sandboxing/security)

---

Likewise

• Rust has a coherent design, but it was incubated at Mozilla for over a decade. I think a lot of that was people sitting down in the same room together
• Postgres has a coherent design, but it started from a commercial company Ingres (led by Stonebraker)
• The more I learn about CPython, the more I am amazed that the core design was all there in CPython 0.9 in ~1990 or so. (A primary example is the way PyObject* works, the M x N waist, which is something Perl, PHP, and JavaScript don't have)
  • The next 35 years was building on the same design. (e.g. Python 3 changed nothing about the core design)
  • I believe this design may have come from the ABC language, which was developed at a research organization in the Netherlands. Guido said that Python was basically ABC ...

I think it is very difficult (but not impossible) to do a new design in a purely distributed development model .. maybe git is a counterexample. It was based on BitKeeper and such, but it was still a new design that was hugely successful (compared to CVS and subversion)

Ironically git is about distributed development, so maybe that's the reason it worked :-)

---

Basically I think it's hard to compete with "big tech/VCs get a lot of talented people to sit in a room together". But it looks like they do understand that

For the past year or so, we've been running the "Boiling the Ocean" event series in Berlin.

I just wouldn't look to GNU/Linux as an example of a new type of software architecture gaining wide adoption -- it was a proven architecture