It's not your imagination. Proper books have approximately 1200dpi resolution. The best printing approaches 2400dpi.

Try viewing two pages side-by-side on a 1920x1080 display. Assuming you can even make out the letters, they're REALLY hard to read... even on a 17-inch diagonal display that approaches the physical dimensions of a normal computer book opened to two facing pages.

At 3840 x 2160-2560, the problem isn't quite as bad... but it's STILL kind of like comparing shitty inkjet printing to a professinally-published book in terms of resolution and legibility. The diminished contrast and blockiness has approximately the same impact on legibility as a diopter of uncorrected astigmatism.

1. eBooks take a relative ETERNITY to flip between pages, compared to a book where you can hold one or more pages between your fingers and more or less instantly flip back and forth with no lag or latency.

2. one page vs 2-up view. For literally CENTURIES, the norm for technical books has been, "diagram on one page, text explaining it on the facing page". Most ebooks shoot that paradigm to hell, either by allowing you to see only one page at a time (with non-insignificant delay when flipping back and forth), or by making almost no meaningful attempt to keep facing pages coherent (in the name of reflowabilility).

3. Digital astigmatism. 1920x1080 (or worse) just doesn't cut it for heavy reading. Antialiasing small text has more or less the same impact on legibility as a diopter of uncorrected astigmatism... turning enclosed parts of letters into blocky gray mush, and increasing the cognitive load.

What we REALLY need to make ebooks not utterly and completely suck are tablets with 300+ PPI screens that are physically at least as large as two facing pages in a typical computer book by someone like Manning or Apress... combined with storage and a processor that can do a total page-flip in under 200ms, and a vastly improved mechanism for interaction that includes some direct equivalent to dog-earing a page, or grabbing a bunch of pages between your fingers so you can rapidly flip through the book.

It's not anyone's imagination. Ebooks ARE harder to read and comprehend than "real" books... and for the most part, utterly and completely SUCK for technical content where you AREN'T reading serially, from start to finish. Ebook readers have shitty hardware whose sole design criterion is "cheap to manufacture", and sloppy layout & editing (assuming the publisher even TRIES to lay it out with any kind of formatting beyond a single paragraph).

Back in high school, I never understood how CD players could even HAVE anything less than infinite channel separation. I finally found the answer ~20 years later: first-gen CD players had two fully-independent (after reconstruction) left-right bitstreams, but shared a single DAC between them using a pair of "sample & hold" circuits. Because they did the two in rapid succession, the first channel in the pair to get output left a "shadow" in the other.

Ironically, you could NOW probably use a combination of FLAC, OFDM, and a megabyte or so of RAM (for decoding) to put CD-quality digital audio ON a LP record (by encoding it to FLAC, then modulating it DSL-style into a bunch of parallel audio frequency bands). Utterly pointless, but probably worthy of some retro-nerdy fame at hackaday.com.

In partial defense of hipsters... 21st century records are thicker & generally sound an order of magnitude better than a 1980 LP, the same way metal tape + Dolby-C recorded from a CD is almost real-world double-blind indistinguishable FROM a CD.

Back in the 1980s, LPs and pre-recorded cassette tapes sounded WAY worse than the recording medium made them *have* to. Pre-recorded tapes barely had the frequency response of a telephone. LPs had bass rolled off, amplitude compressed, and channel-separation deliberately reduced to make records thinner (ie, cheaper) and longer-playing.

Give yourself the freedom to fill a thick 45rpm 12" record with 6-7 minutes of music, and it's a very different result from packing an hour of music onto a 33-1/3 RPM LP of the same diameter (and half the thickness).

The BBC's problem is, it tries *so* hard to espouse no editorial position at all, it ends up inadvertently legitimizing actual extremists in the name of "balance". A roundtable news episode between a leader of the Muslim Brotherhood and a Haredi Zionist isn't "balanced", it's a brawl worthy of Jerry Springer that might be entertaining, but isn't "news" or "journalism".

Journalism is not objective, nor should it be... as long as it's honest about its biases AND strives for at least factual honesty & correctness.

"Bias transparency" becomes relevant when you realize that two news stories about the same event can both technically be 100% factually correct, yet present conclusions that are almost 180 degrees apart by using the same facts to argue different conclusions, or silently omitting germane details. It's why you *really* need to find 2 or 3 respectable news sources with slightly different biases, and consider the possibility that they're *both* simultaneously right & wrong.

Comparing Crew Dragon to Starliner is like comparing a DeLorean to a Yugo. Except a Yugo actually worked (mostly)... at least, for a year or two... which is more than you can presently say about Starliner.

Now, take the "hinged screen" idea, but give us THREE panels--- a 3840x2160 center one that's at least 17.3", flanked by a pair of portrait-orientation panels that are 1680-1920 pixels wide and 2160 pixels high that fold over it like window shutters. Capable of 120hz+ FreeSync, with built-in multi-display DisplayPort hub, powered and driven by a single Thunderbolt cable.

To get more value from the development & tooling investment, take the same monitor guts, and make a second variant that puts them into a bigger enclosure with proper VESA bracket on the rear, built-in power supply, and the expected assortment of desktop monitor ports on the rear. It will instantly become the "must-have" display for anyone buying an eBlaztr mini-ITX case (and quite probably drive the creation of an entire new market segment of desktop PCs built around a similar form factor, using this monitor). Or even just people who don't necessarily need their computer "on the go", but want something they can easily lug around from one place to another every few months (think: college students going home for a month at Christmas, people taking weekend trips to a vacation home, etc).

They could also use the display to spawn an entire new family of laptops. Say, one that tries to be not particularly thick... and a second one that just says, "fuck it, it's a laptop-like desktop replacement" with case big enough for a full-on mechanical keyboard, a flex bay for either a single 3.5" hard drive or sandwich of two pairs of 2.5" hard drives side by side occupying the space of a 3.5", etc." The display could snap onto the base for convenient travel.

Bonus points if the monitor uses "double-jointed" side hinges, so you can fold the side panels flat against the back (facing rearward), either turned off to save space/power, or turned on to serve as a secondary display when doing presentations.

Disasters aside, I think one of the most demoralizing moments in Boeing history was probably the first time they saw photos of Starliner's interior next to Crew Dragon's & realized their flagship product of the 21st century looked like a Soviet vacuum cleaner next to a Dyson.

(goddamn it, Slashdot! Give mobile users fucking preview-then-post! It's way too easy to accidentally submit when typing on a phone using Palmsource Graffiti!)

Anyway, Windows bends over backward to treat kernel code like a big DLL and allow old code to work in newer Windows. Linux goes to the opposite extreme, and treats every kernel-build as a 100% new and 100%-incompatible monolithic hardwired blob.

I *believe* Google (as of Android 13) is now working towards a "middle way" for Android Linux kernels that adds build-independent open-source thunking layer to closed-source binaries to make binary kernel modules a little more abstracted and forward-compatible... not quite as virtualized as Windows, but an order of magnitude better than the way Linux has historically handled the problem.

It looks like I accidentally deleted a chunk of the last paragraph. My main point was that with Linux, closed-source kernel binaries aren't even compatible between adjacent versions of the same kernel, let alone years later.

The way I understand it, Windows kernel modules use dynamic late-binding & virtual tables, while in Linux, everything is monolithic & hardwired at compile time (after autoconfig first by the build system).

Put another way:

Windows: the kernel consults a virtual lookup table, and jumps through hoops to allow old code to t

Not to mention, with Linux, you're basically fucked at update time if you're dependent upon any binary-only kernel modules. Windows might *grumble* about using old drivers, but if you're absolutely *hellbent* on doing it, you can sometimes even coax 64-bit Windows 11 to use a 16-bit parallel-port SCSI interface meant for NT 4 (by rewriting the .inf file, and doing some registry-hacking). If co

As opposed to a new PC with 400+ watt RTX video card?

Has anybody in a patent-infringement lawsuit ever tried to challenge the constitutionality of contemporary IP law by making the legal argument that:

1. The wording of Article I, Section 8, Clause 8 makes protection of intellectual property contingent upon it actually promoting the sciences and useful arts.

2. Not only does contemporary IP law fail to achieve that goal, it actively and demonstrably harms its achievement.

And... if so.... did the challenge fail because:

a) The US Supreme Court heard the case, and issued a ruling that there's no requirement that IP law actually promote any public purpose or provide any benefit

b) The USSC heard the case, and ruled against it on the grounds that the defendants failed to make a case that IP law in their specific case harmed that goal

c) The USSC heard the case, but ultimately ruled against the defendants on purely procedural grounds that didn't really address the matter of the law's fundamental constitutionality

d) The USSC declined to hear the case, allowing the lower court's ruling to stand by default?

It seems like the present Supreme Court majority's professed belief in the sanctity of strict constructionism could work in favor of a legal theory arguing against the constitutionality of contemporary IP law. If, in fact, the Constitution's plain language is to be taken as inerrant literal truth at face value, then it's hard to see how any reasonable reading of the clause could conclude it does NOT make the constitutionality of IP law contingent upon the law actually promoting the advancement of science and the useful arts.

If the same justices who argue in favor of strict constructionalism suddenly did a 180 and ruled that it doesn't apply in this case, it would seriously undermine the perceived legitimacy of every other ruling they've issued that has strict constructionalism as its professed legal basis.

Try reinforced concrete using proper rebar, ICF forms, and a concrete roof deck. 95% of tornadoes are ET3 or below... and reinforced concrete (roof and all) plus Florida-code missile-rated impact-glass windows will make it through those with barely a scratch. EF4 & 5 are monsters... but even in "tornado alley", they're exceptionally rare.

If Kansas had Dade County building codes, most tornadoes would barely make a dent (in neighborhoods built to Dade County standards) instead of destroying neighborhoods wholesale.

And yes, my own house (in South Florida) is reinforced concrete & has suspended concrete slabs for the roof & second floor. I practice what I preach :-)