* Posts by robmacl

1 publicly visible post • joined 26 Dec 2023

War of the workstations: How the lowest bidders shaped today's tech landscape

robmacl

Blast from the past

Kept reading because I wondered where you are going, but I certainly didn't expect to end up at Dylan. I worked on Dylan back in the day, and before that on CMU Common Lisp, which was a public domain implementation of CL, in the end running on conventional hardware. We aspired to have a Lisp Machine experience.

It's hilarious these days when I see Chrome using 64 meg to implement an empty tab, and back then we were really self conscious about our unreasonable use of 4 meg for an entire IDE.

It was a highly productive environment at the time, and you could certainly spin alternative histories around it. As to what happened, Gabriel's good news/bad news paper is a solid take. At the highest scale economics is the best lens.

What you don't mention is that each wave of crappier computers was 3x-100x cheaper than what had come before, and consequently so many more were made and used. A crappy computer is so much better than no computer that it spread like crazy.

The rise of Unix workstations was in parallel with PCs, and initially PCs were not powerful enough and software not sophisticated enough to give a mainframe replacement experience. But the explosion of the PC dramatically shifted the center of gravity of software development away from the universities and research labs by the late 80's.

In the trenches, as an implementor, the arrival of popular GUI frameworks was a big problem. There was just too much going on out there for us to hook it up to Lisp in an artful way. We made tools that users could use to do the hookup, but the GC memory model created problems, and the APIs were big and every growing. There just weren't enough people who cared and knew how to do it.

Gabriel's good news/bad news take is solid, and worth a read. See also "The Cathedral and the Bazaar" by Eric S Raymond. Ultimately the self reinforcing network effects of greater market share and mind share are too much to overcome. Too much is happening somewhere else, and you are stuck in a backwater.