That's my point, the newer NVidia chips are defective here. The 940MX didn't hijack the external ports to save ¢2 on the BOM of a $2K machine

(my i7 is the gen 7, yours is probably gen 6 based on its being matched with the 940MX, plus Lenovo didn't see fit to sell the proper T460p with the real quad core i7 around here (France) — I'd have gone ugly machine (lenovo) if it was in the race at all, but it wasn't. Buying out of state is/was a no-no in this case (keyboard layout, warranty, tax status [work machine], etc.).

(next time, ... I'll see what the market delivers, but will probably be more willing to trade raw performance for less clunk. Still weeping my 2012 Samsung 900, it's a shame they exited the market. Still a notch above today's zenbooks in terms of form factor and finish.

The damn NVidia "optimus" laptops without hardware mux.

My laptop (2017 Gigabyte Aero 14W) has an nvidia GTX 1060 that I never use (the i7's internal GPU is perfectly adequate to run IDEA IntelliJ, YouTube's videos, and GNOME's desktop animation at QHD 2560x1440p. PLENTY ADEQUATE). So I never, ever waste 25+W running the damn nvidia chip, except for ONE thing: enabling the external outputs.

Yep, if I want to display something on a beamer, I must run "intel-virtual-output -f" in a clunky shell away.

Older laptops/chips at least had the good grace of having a real HDMI multiplexer on the external outputs, now it must be blitted through the nvidia chip.

Fuck you, NVidia. Fuck you again.

(and yes, Gigabyte, you failed. If you sold the same machine without the nvidia chip, even at the exact same price, I'd have bought it without hesitation. But yours was the only 14" option with a real (really quad-core) i7 and ability to have 32 GiB of RAM, under 2kg, this winter).

AVE is a blend of TGV and ICE depending on the stock vintage

<blockquote>You can simply put several signal posts in a row, and read the red/amber/green streak that goes by the cabin.</blockquote>

well, you can. But the cab-signal system also gives you an advantage, as you have a continuous

A missed Square (double-red, absolute stop, whatever it is rendered as in your neck of woods) should trigger an immediate emergency stop of all trains in the vicinity, cab-signal or not, anyway.

<blockquote><blockquote>Some slower but WAY busier lines also need to get away with the old block system, in order to reduce the spacing.</blockquote> <p>Ah, the good old "throw safety out of the window to increase profits" way of managing things.</p></blockquote>
Well, no. The system removes (actually <b>disables</b> unless a non-equipped or faulty train comes) the static, side-mounted block system; <b>replacing</b> it with a dynamic, moving block system. Each train knows where the previous train is, its own speed (obviously) and the speed of the previous. And of course, the brake distances.

SACEM (and the like) computes the safe stopping distance, and can cause everything from slowing down all the way to hitting the brakes, in order to keep the safe distance held.

The "5 meters" (actually, I saw a couple times even closer) obviously can happen only when one train is stopped (in station) and the next train approaches. Nowadays, they've tuned the system with more space, not because it was particularly unsafe, but because it was occasionally freaking out passengers... The damn thing has been working, cramming LOTS of commuter trains daily for 20 years. It's working fine, thank you very much :)

Now about joining the trains, it might work, until you want to do things such as stop in station or drive fast between the stations...
Obviously, when trains run their 'normal' 70-80km/h, the software spaces them several hundreds of meters apart. AT LEAST (dunno what the emergency braking distance from 80km/h for a MS61 or MI2N is, but the commercial deceleration is enforced).

you can't actually read the signals, when the train travels > 250km/h. Even in 1980, designers of the TGV (270 then, 320-350km/h now) knew this, and the signalling is done using what is called <i>cab-signal</i>, which puts the display within the cabin.

Some slower but WAY busier lines also need to get away with the old block system, in order to reduce the spacing. In Paris, the two primary suburban lines (RER A and B) use what is called 'permissive' spacing, (SACEM on A, KVBP or KCVP on B), in order to reduce space between trains -- SACEM can space trains under 5 meters apart under stressed conditions.

But the key point of these advanced signalling systems is that the train-spacing software MUST be perfect. Not just "bug-free, we tested and deployed and ITIL'd the thing to death" but "mathematically proven bug-free". And even that doesn't cut it. Read up on how the SACEM hardware works, for instance. Or on the "Methode B" used to design the SACEM and the SAET (the latter of which powers automatic lines such as M14 and now M1 in Paris. SAET can safely take even a 110 year-old manually driven train within the robotic shuttle traffic, and get everyone safe there).

Back to China, perhaps the strike broke some communication line, making the position of the stopped train 'unknown'. But if that happened, someone much worse must have happened as well.

Perhaps, by cutting corners everywhere, they've also cut on the provably bug-free programming which one MUST use to build the train-spacing software. THAT, if that happened, is criminal.

Perhaps they've cut corners on brakes. Or whatever.

Hopefully for them, that's a fixable bug....

SuperPhenix, not Phenix, failed because of a construction defect, which the builder (a close friend of the right-wing, and strong influencer of the left as well) managed to hide until the thing was taken out of commission.

Add on that some pressure by the ecologists, when the left needed them to have half of the power, and bam, the device was taken down.

From an industriel point of view, it did produce electricity, it could have worked fine was it actually built up to spec. Too bad its burial means no progress is going to happen on molten-salts reactors either.

PS: the builder is now botching the French EPR prototype (dunno about the Finnish EPR, but I've not heard very good news either).

until about 10 years ago, cryptographic devices were classified as ammunition/weapons in France (one of those rules from the '40s fascist puppet state that was REinstated right after the war, this time (again) against the commies).
It took the need to protect credit cards over teh intarnet to compel the government into allowing first 128-bit (up from 40-bit) encryption, then just lift the ban/classification. SSF was just this: a legally compliant, meaning crippled (to 40bit IIRC, but OP seems more familiar than me with that), implementation of SSH.

try, as I have, to use OpenJDK to do plain dumb Play! Framework development under Eclipse (either vanilla Squeeze or Lucid Lynx)

My experience at it is that it will crash about every time you change a bit in a page. Went back to the Sun JDK in no time.

Hopefully one day that gets better (and I'm sure it will, eventually)

As a matter of fact, the existence of religions, basic tenets of the primary religions (from the French point of view: Christianity, Judaism and Islam), and keys to their influence insofar as they shaped Europe's history *are* subjects brushed on in French schools (starting at 3rd grade, then repeatedly until the end of High School). [**]

Enough to have the clues to "read" our world, but never taught as "stuff one ought to believe in" (though anyone is welcome to borrow books from municipal libraries or free to step into any legal sect's building)

[**] with a significant exception with Alsace/Moselle (near the German border) which still apply the 1801 (French) Concordate, which specifically (re)bound the French Government with the Holy See. That law has been repealed in 1904 (the State-Church separation law), except in areas which were part of Germany at the time. Yes, I found crosses on pre-school class walls definitely spooky while I lived there with my kids.

[citreq]
Most people can't read the default fonts ("damn too small") when the DPI increases, and most Win32 applications fail to scale properly. Plus, most people don't know how to enlarge the font sizes to enjoy better drawn text at sizes their eyes can read.
[/citreq]

Blame Steve/Bill on this one, but manufacturers probably don't want to bother with higher return rates, higher defect rates, when the dominant OS will not let their product shine.

Watch Apple for a way out. OSX can (and does) scale very well with the resolution (so can X&FreeType -- eclipse on a 9" netbook? Bring it on)

<quote>leading to greater susceptibility to widespread disease and genetic problems in the generations to come. </quote>

not sure increasing the odds of humanity to decimation (or worse) is actually a <i>very bad</i> thing, from the general ecosystem's point of view...

(of course, as individuals we would certainly disagree with "someone" intentionally spreading something that kills half of our loved ones, but we might have gone a bit too far already on the "conquer and submit" part...

Yeah, but it could VERY easily do a small boot partition on FAT16 (enough to stuff their nice autorun.inf and their installer downloader), and let the rest be handled on ext2 (maybe going far enough down the road to license or otherwise provide some fine mecenate to the author of ext2fs.sys)