Chapter 7 was a rant about X11.
It's interesting how the nagging about X11 memory usage is similar to today's nagging about Electron-apps and web-browsers.

Quote:Even today, the X server still turns fast computers into dumb terminals. You need a fairly hefty computer to make X run fast — something that hardware vendors love.

We know today that the idea of running graphic programs over the network is lacking as it can't use graphic acceleration. So many X extensions have been made to bypass its initial design and use direct graphic rendering.

On that note, I had never in my life ran `xauth` in its interactive mode, I had no clue it even existed. That whole section reminds me of the thread we had about all the ways to start an X11 server.

The rant definitely comes from a while back, a time where fonts where still hosted in the X11 server. These were ugly and couldn't even use anti-aliasing because of patents.

There's a point to be made about the fact that there aren't any standard GUI toolkit. The author calls it "user interface Tower of Babel". We do have the ICCCM "standard" but even today it is hard to respect and comply with, along with EWMH as a subset.

Looking back at it, in my opinion, that's an issue about the lack of overarching power-structure that would dictate who can do what and what is standard and what isn't. There aren't authorities enforcing and deciding for others, there are rough standards because it's mainly driven by open source and free contribution, at least today.
We can clearly see how people react today when someone tries to enforce their vision on others with projects from RedHat... Not everyone likes this.

On the other side, this can give rise to the "cognitive black holes" that the author talks about, because of the ever-increasing complexity and fracture. "Feature creep".

The resource location section killed me. It's definitely true, a resource could be in a million places for X11. At least now there are libraries that will search all the places for it. But the reverse would mean having a central registry, so back to the previous point.

And don't worry about all the "XFillRectangle" and "XDrawRectangle" issues, now we have Wayland and you'll deal with OpenGL constructs instead.

Chapter 8 about csh, pipes, and find.

Again with the theme of "it was good in the past but what it is today doesn't fit the case". Also, in relation with feature-creep.

The author argues that leaving flexibility in the hand of the users slowly but surely will lead to failure. I'm not sure about that. Imagine the opposite scenario, where we are handed the whole pristinely designed OS from above and are not left with any room for improvement, nor allow any other devs than those from a certain group to touch it. Is that good, back to the previous point? I'm not sure, but I think the author would be fan of the centralization movements happening today.

Quote:When was the last time your Unix workstation was as useful as a Macintosh?

These days I have no idea...
Mac was indeed made to please non-techie users while Unix was targeted at developers. I'm not sure of that either.

I love the argument that the shell is mostly unused. This just opens up opportunities for ways in which it could be upgraded. But that also doesn't mean we should remove it.

Chapter 9 on Programming

I think the author likes lisp a lot.