The AI-generated cat pictures thread
#2327
Boost Pope
iTrader: (8)
Join Date: Sep 2005
Location: Chicago. (The less-murder part.)
Posts: 33,019
Total Cats: 6,587
It broke.
Link to larger version?
elesjuan wins the thread.
Those pics from ISS and Endeavour are ******* gorgeous. It's a shame we'll not see anything like that for quite a long time, now.
Also, the chick holding the NOS bottle above her bike has a package.
Link to larger version?
elesjuan wins the thread.
Those pics from ISS and Endeavour are ******* gorgeous. It's a shame we'll not see anything like that for quite a long time, now.
Also, the chick holding the NOS bottle above her bike has a package.
#2328
Tour de Franzia
iTrader: (6)
Join Date: Jun 2006
Location: Republic of Dallas
Posts: 29,085
Total Cats: 375
Dan Neil is a horesepower ***-hog who doesn't know how to drive a car, doesn't c are to learn, and disregards reliability while ******* the ******** of German cars.
He also writes like a 16-year old collegiate credit fledgling:
"The steering ratio is tighter than a young cabernet, "
Seriously, come on. Write the article, stfu, you are not Jeremy Clarkson or James May. Write like an adult, please. If he'd ever seen a racetrack, he'd apprecaite 200whp+ in a Miata...but I'll at least concede that the Miata's underpowered novelty is endearing.
He also writes like a 16-year old collegiate credit fledgling:
"The steering ratio is tighter than a young cabernet, "
Seriously, come on. Write the article, stfu, you are not Jeremy Clarkson or James May. Write like an adult, please. If he'd ever seen a racetrack, he'd apprecaite 200whp+ in a Miata...but I'll at least concede that the Miata's underpowered novelty is endearing.
#2329
Boost Pope
iTrader: (8)
Join Date: Sep 2005
Location: Chicago. (The less-murder part.)
Posts: 33,019
Total Cats: 6,587
Looks like the control room of a VVER-440 reactor. Old Soviet design dating back to the late 60s / early 70s, very similar to a western-style PWR. Lots of 'em in Russia and the former Soviet territories, and the newer models of that family are still being built today.
Another view, this one from the (presently idle) Units 3 & 4 at Kozloduy, Bulgaria. Judging by the guy in the sweater, this may be the same control room as the one you pictured.
Another view, this one from the (presently idle) Units 3 & 4 at Kozloduy, Bulgaria. Judging by the guy in the sweater, this may be the same control room as the one you pictured.
#2330
Elite Member
iTrader: (10)
Join Date: Jun 2006
Location: Athens, Greece
Posts: 5,976
Total Cats: 355
#2331
Looks like the control room of a VVER-440 reactor. Old Soviet design dating back to the late 60s / early 70s, very similar to a western-style PWR. Lots of 'em in Russia and the former Soviet territories, and the newer models of that family are still being built today.
Another view, this one from the (presently idle) Units 3 & 4 at Kozloduy, Bulgaria. Judging by the guy in the sweater, this may be the same control room as the one you pictured.
Another view, this one from the (presently idle) Units 3 & 4 at Kozloduy, Bulgaria. Judging by the guy in the sweater, this may be the same control room as the one you pictured.
#2332
Boost Pope
iTrader: (8)
Join Date: Sep 2005
Location: Chicago. (The less-murder part.)
Posts: 33,019
Total Cats: 6,587
Funny thing is, this is pretty much what even the newest reactor control rooms look like. They do use modern PCs for ancillary functions, like logging, analyzing core flux and planning fuel rotations, basically all of the "offline" functions. But the mission-critical stuff is still mostly discrete electronics, much of it analog. The feeling is that simple, discrete control systems are easier to design safely, with little potential for the sort of unexpected interactions or timing snafus that can happen with software-based systems.
#2333
The new PC stuff is easier to program (make changes) and look at datalogs. The old stuff never broke BUT it is getting next to impossible to find replacements. We had a monitor go bad the other day and the maint tech told me it was a $10K part. All of our datalogging printers are pretty much trash and most have been turned off (we are flying blind).
#2334
Boost Pope
iTrader: (8)
Join Date: Sep 2005
Location: Chicago. (The less-murder part.)
Posts: 33,019
Total Cats: 6,587
Of course, nuke plants don't require frequent changes to their fundamental processes (they only do one thing), and datalogging is already computerized, albeit with pen-recorder backups. And most of the control systems are custom-built and endlessly supported anyway, so it's not like you have to worry about a certain assembly being obsoleted by the manufacturer. If some circuit board or indicator at Crystal River needs replacing, Babcock & Wilcox is going to provide one, regardless of how old it is.
On the one hand, I think that nuke plants are probably subjected to a much higher level of scrutiny than is "fair", if you consider other, comparable facilities as a baseline. I'm not sure what they make at the plant where you work, but I'm guessing that if things went sufficiently wrong, it could probably kill a fair number of people.
On the other hand, I work with embedded software every day, and I know first-hand how even in relatively simple, isolated systems it's never perfect. Even if the kernel, the microcode and the SDK are just 100% absolutely faultless, there's always some potential race condition you didn't predict, or some bizarre combination of inputs that results in a resource conflict, or some other damn thing that it never even occurred to anybody to think about.
With discrete systems, it's actually possible to achieve the sort of atomicity that the software reliability experts have been circle-jerking about for decades. An analog comparator doesn't really give a **** about concurrency; it has only one job to do, and it will do it continuously and reliably regardless of whatever else is happening around it.
The events at TMI-2 in 1979 were a pretty good example of this. When things started to go soft and brown in a hurry, things were happening so quickly that the Plant Process Computer couldn't keep up, and was several hours behind realtime in processing and displaying logs and alarms by the end of the event. Despite everything that went wrong, however, there was not a single failure or anomalous operation in the critical process controls (save for the defective valves that started the whole thing). For certain, some of the manual controls and indicators were poorly arranged, ambiguous, and just plain stupidly designed; but that's Human Factors Engineering, which is a whole different animal.
But the actual control systems themselves functioned as designed, operated completely without failure, and in fact saved the plant from an almost total and immediate catastrophe, despite the fact that the operators kept overriding the automatic systems (incorrectly believing that they were malfunctioning and over-filling an already full core.)
Last edited by Joe Perez; 07-05-2011 at 05:45 PM. Reason: added fun picture.