It's hard to write something that will blow peoples' minds when you're writing in a genre that's had decades of writers mining the same material. But we ought to beware of survivor bias; the stories we remember from the Golden Age are just the ones worth remembering. Most of the stories that got published back then were derivative and extremely crude. Today, in contrast, most stories that get published are derivative but very competently crafted. I guess that's progress of a kind but in a way it's almost depressing.

I think the most recently written mind-blowing sci-fi (or perhaps weird fiction) novel I've read was China Mieville's *The City & the City*, which tied with *The Windup Girl* in 2010 for Best Novel Hugo. I was impressed both by the originality of the story and the technical quality of the writing.

I recently read Ken Liu's translation of Liu Cixin's *The Three Body Problem*, which I enjoyed. In some ways it reminds me of an old Hal Clement story in which the author works out the consequences of some scientific idea in great detail, but the story also deals with the fallout of China's Cultural Revolution and the modern rise of public anti-science sentiment. So this is a foreign novel which doesn't fit neatly into our ideas about genres of science fiction. It's got a foot in the old-school hard science fiction camp and foot in the new wave tradition of literary experimentation and social science speculation camp.

This was a KDE Wayland bug? I've never encountered in X11 with KDE 5.

There's actually a proposal before the Fedora steering committee to change the default desktop in Fedora from Gnome to KDE. It's quite a serious proposal: https://fedoraproject.org/wiki...

I don't think it will gain traction, though.

Doesn't sound like Fedora is for you!

Fedora is, and always has been, a desktop-focused distribution. It's not used in the data center (or shouldn't be!)

There are lots of spins of Fedora. I use the Mate Desktop spin. Just means the default desktop environment is different.

I'm pretty sure you know all that. Well played. We all rose to the bait!

That issue has nothing to do with IBM. This issue with controlling source redistribution has a been a longstanding and challenging issue with RedHat going back more than 10 years. It just came to a head recently when RH stopped pushing their spec files to git. Several RH folk are on record as saying this move was not influenced in anyway by anyone in the parent company. In fact they didn't even inform IBM execs of this nor would they have needed to since they are their own business unit.

Can't blame IBM for a controversial business policy that goes back many years.

ALSA was complicated. ALSA is a kernel driver + userspace library. Kernel driver just talks to the hardware, and you'll have mixing only if the hardware supports it.

Trouble was that hardware mixing quickly went out of favor. ALSA implemented dmix, which turns out to be library based so it doesn't work for anything that's not talking to ALSA specifically, so old apps are out of luck. There were also troubles with old school apps trying to open /dev/dsp directly and blocking out everything else, ALSA configuration being inscrutable, and apps using /dev/dsp, ALSA, jack, artsd and esd all existing simultaneously. That era was just a terrible mess if you had anything other than say a 100% KDE or 100% Gnome system, and even then it was tricky. Sound just broke seemingly at random depending on what managed to grab the sound card.

Pulseaudio introduced sanity by gobbling everything up and supporting all that junk at once.

Well, we disagree, there. I see it as having gone through so many iterations and having so many players involved now it's lost much of it's ideological purity and, with that, some of it's power, reliability, and usefulness due to the resulting loss of cohesion. It's backsliding into trying to become Windows NT, completely with 'svchost.exe'.

Ideological purity is to me a flaw, especially if you mean holding on to old things invented in the 60s. Cohesiveness is good, paralysis is not.

Also, as a sysadmin my job is to solve problems, not to bask in ideological purity. So, eg, I really like that systemd makes it really easy to ask for the logs of 3 particular services last weekend at 3 AM, and have them properly interleaved with one command. On SysV systems, any time spent on figuring out where it logs, if the log was rotated, and whether it's in .1 or .5.gz, and assembling a combined report is wasted time.

I think shell scripts are awesome and they do tons of real work for my clients and my own company's infrastructure. They are indispensable to me. Why be annoyed that someone else is getting some mileage out of a technology you don't happen to use?

You misunderstand. I don't mean shell scripts, I mean bash. Bash is crap. I want something more like PowerShell in spirit, only that's not quite it either. I want a shell where: every command is 100% precise. All arguments are always exact. There's never confusion with spaces, quotes, newlines, or anything else. IFS doesn't exist. You never need to work around command line lengths by hand. "Do this to $list" always works, whether 10 or a million files. It should offer excellent process management, along the lines of "Launch up to 10 simultaneous instances of (command) on (list of files), call (this) on command fail and (that) on success" as a natively supported primitive.

All commands should output something Powershell-ish. Not plain text, but flexible objects with datatypes. I don't want to parse stuff with AWK, I just want to ask for $File.CreationDate.Year. The format should be perfectly delimited, no splitting by spaces or by ":", and allow adding additional data without breaking every script. Every datatype should be transmitted perfectly. A date is always the exact date the software knows of internally, not something truncated to seconds to look pretty on the console.

The Unix text stuff was a good start. Time to do better.

I teach several sysadmin classes, too. What I notice is that once novice folks learn shell script, they become disenchanted with unit files. However, beforehand, they are usually indifferent or show some bias toward using units. Also keep in mind nobody got rid of them, they just forced some major distros to do it. There are still plenty that didn't, including FreeBSD which I saw grow in popularity and catch a lot of disenchanted former Linux users.

Funny, I'm the complete opposite. I wrangled with SysV scripts a bunch. They're awful. Every distro has its own standards and tooling. Upgrades sometimes force you to look at a 3-way diff of some 150 line script. Some do weird and fishy things. Best case it's 50 lines of boilerplate and 3 unique lines.

Unit files do it right. Code isn't ever placed in /etc. The configuration format is maximally simple and boring. I can roll up to a systemd system and quickly see what the admin changed, because all the changes are in /etc, and the only thing to be found is the changes. Upgrades go much smoother. The boring boilerplate with start-stop-daemon doesn't exist. PID files don't exist. Services are tracked perfectly. Every possible knob can be applied to every possible unit without hacking up a shell script. Logging is done reliably for every single service.

Well, it's like in Econ 101 when you studied equillibrium prices. At $3500 the number of units demanded is small, but if you dropped that to $1000 there should be more units demanded, assuming consumers are economically rational.

There is a tech adoption curve in which different groups of people play important roles in each stage of a new product's life cycle. At the stage Vision Pro is at now, you'd be focused on only about 1% of the potential market. The linked article calls these people "innovators", but that's unduly complementary; these are the people who want something because it's *new* whether or not it actually does anything useful. This is not irrational per se; they're *interested* in new shit, but it's not pragmatic, and the pragmatists are where you make real money.

Still, these scare-quotes "innovators" are important because set the stage for more practical consumers to follow. Perhaps most importantly, when you are talking about a *platform* like this people hungry for applications to run on the doorstop they just bought attract developers. And when the right app comes along the product becomes very attractive to pragmatists. This happened with the original IBM PC in 1981, which if you count the monitor cost the equivalent of around $8000 in today's money. I remember this well; they were status symbols that sat on influential managers' desks doing nothing, until people started discovering VisiCalc -- the first spreadsheet. When Lotus 1-2-3 arrives two years after the PC's debut, suddenly those doorstops became must-haves for everyone.

So it's really important for Apple to get a lot of these things into peoples' hands early on if this product is ever to become successful, because it's a *platform* for app developers, and app developers need users ready to buy to justify the cost and risk. So it's likely Apple miscalculated by pricing the device so high. And lack of units sold is going to scare of developers.

But to be fair this pricing is much harder than it sounds;. Consumers are extremely perverse when it comes to their response to price changes. I once raised the price of a product from $500 to $1500 and was astonished to find sales went dramatically up. In part you could say this is because people aren't economically rational; but I think in that case it was that human judgment is much more complex and nuanced than economic models. I think customers looked at the price tag and figured nobody could sell somethign as good as we claimed our product to be for $500. And they were right, which is why I raised the price.

I think realistically most people that read a lot have moved onto e-books. Just like physical copies of video games or music are dying, so are physical copies of books. Not only is it convenient, but realistically the cost to "publish" an ebook is effectively nothing.

Aside from proximity, Mars is better suited to colonization than the moon.

There's a minor atmosphere that makes traditional flight possible, and evens out the temperature swings a bit.
The day/night cycle is nearly identical to Earth so solar arrays are more practical (compared to lunar days lasting 28 Earth days).
The gravity is higher (38% of Earth's gravity compared to 17%), so muscle atrophy and general disorientation should be less severe compared to the lunar surface. On the other hand this does make it harder to get back off the surface and into orbit again.
While both likely have "enough" water, Mars has more.

The only thing the moon offers that's better is for resupply or emergency scenarios, Earth is just a hop and a skip away.

I've grown quite disappointed with the state of Linux over the years.

Back when I started, I had DOS and Windows 3.1/9x. Linux was troublesome (remember writing modelines for X11?), but also like magic in comparison. You want a very minimalist setup? Can do that. Want ridiculously overdone themes with fake transparency on hardware that can barely bear the load? Sure, just install Enlightenment. Want to write complex scripts? Sure, bash is far superior to crappy .bat files, and then there's Perl. And you can even email the author of almost anything that runs on the system and actually talk to them. Freaking magic, man.

The difference I suspect is that I never got into the "unix philosophy" worship. I just liked that it was more functional than DOS/Windows. I see no reason to stop at it, and if somebody comes up with something better, then good riddance. Unfortunately Linux seems to have ended its wildly experimental phase and now gotten annoyingly conservative. How are we still using bash in 2024? Annoys the heck out of me. At least we got rid of SysV init scripts, that's a start.

Powershell has its weird quirks, but after writing a lot of bash, I can't stand it. Any time I write more than 5 lines of shell script I quickly stop and switch to Perl if for personal use, or Python for longer term use.

I don't get how after so long, Linux is still cursed with such a horrible scripting language.

Powershell once you get used to it is a lot more pleasant to work in than bash, and I say that as somebody who only uses Windows when absolutely necessary.

So rather than them hating each other, they'll be united in their hatred for the UN.

Nobody wants anyone coming into their home and telling them what to do. The issue between Israel and Palestine is that both of them consider the land theirs, and and foreign interference that sides with one side will be hated by the other, and any that supports neither side or both sides equally will be hated by both sides.

The reality is that the elites of both sides want to fight . . . but realistically Israel is the side that will come out on top militarily, so the Palestinian leaders have to be willing to come to the table and negotiate. They're not getting one state, and they're not getting any historic territory back - not without land swaps anyways.

Come to the table. Draw up official borders and have nothing to do with each other. Israel doesn't control what goes on in Gaza's borders and they become an independent state (maybe united with the West Bank, maybe West Bank becomes its own separate country - who knows). After that though, any attack from EITHER side against the other is an act of war. There is no more fighting, no more trying to reclaim ancestral lands - you have your territory and you stay there in peace.

Funny, I like Fedora for pretty much the opposite reasons -- adoption of cool new features like systemd, and nice enterprise features like SELinux that are hard to find with good support almost anywhere else.

Pulseaudio was a very nice improvement over the horrors of getting ALSA configured properly. These days it's replaced with pipewire, by the way.