You're not alone in your despair. Categorizing the new discussion system as a clusterfuck doesn't begin to describe how badly broken it is. The slashdot "editors" must never read any of the stories, because, as you point out, it's been *months*, and yet nothing much seems to have changed.

Of course, the whole hierarchy viewing mechanism is also totally fubarred, so you'll probably never even be able to view this response.

I see it as a positive. I'm now wasting much less time on slashdot.

The info in the App store is a little misleading. If you search for the NYT app it says April 1, 2010. But if you actually click for more details, you see that it was updated to version 2.0.4 on December 23, 2010. So it's not as out of date as it appears to be at first glance.

Not sure if you'll ever see this, the newest /. discussion system is a total disaster. Anyway, here goes ...

Pilots are getting "dumber" because planes do so much for them now. But, in order to save lives, I want pilots to be dumber still. I want them to have an "Easy" button, like Staples advertises.

When the situation is totally fucked up, and the pilot can't stabilize it, the pilot hits the Easy button. Then the plane does what it has to in order to at least stay in the air. E.g. in the case of pitot tube failure causing loss of airspeed information, software sets IIRC 85% thrust at a specified AOA. Then at least the plane doesn't drop from the sky. Software can use info from sources that are still available. E.g. if altimeter isn't trustworthy, maybe GPS is valid, so use altitude from that (in case thrust setting varies with altitude). Or if no GPS, then use the radar altimeter. Or maybe the inertial nav system has an idea of the current altitude. Etc. In other words, do whatever you can to keep the plane from crashing. Also, if the plane is already in a bad configuration or even in a stall, then do whatever is possible to recover. A computer should be better at that then a "data entry operator".

This is an incredible can of worms to open. But, given how so many pilots are now "data entry operators" (see a comment below, and see lamentations on various pilot forums), it might actually save lives.

I predict, that ten or twenty years from now, most pilots will be next to useless in emergencies, and something like what I suggest will be de rigueur. Yes, the autopilot can disengage when it can't handle a dynamic situation. But then, after the pilots also can't handle it, there needs to be a "last ditch" computer on board. Given my druthers, I'd rather have Sully fly the plane. But I'd trust a computer over a "data entry operator" any day.

C did not even exist when the first two volumes of TAOCP were published. I'm not at all crazy about MIX, but Knuth can't be criticized for not "simplifying" a language that hadn't even been invented at the time.

Plus, MIX isn't completely "imaginary". Emulators actually exist for it.

You're exactly right, if society valued engineers they would be paid better. It's an imperfect method, but it's the best we've found in many years of trying.

Too bad you're only at +2, many people won't even see your comment.

And you probably won't even see my comment, /. is totally fucked up since the recent change to the discussion system.
 

Back in the old days when Pud ran FuckedCompany.com, he would love to post stories about Accenture. He'd always say: Accenture (pronounced ass-enter).

I currently have mod points, but I'm responding anyway. Two reasons
1) it's already +5 funny
and even if it wasn't, then
2) the new discussion system is so borked that I don't want to moderate: I can't in good conscience do it, because I simply can't follow all the hidden posts and confusing nesting to make sure I'm doing the right thing.

I haven't had mod points in about 6 months. Once my karma went from positive to good they just disappeared. Before that I had them pretty much every week. Things must be desperate, maybe nobody wants to moderate. Currently there are very few highly rated comments in the posts. So now slashdot is scraping the dregs (e.g. I have mod points again)!

I agree in principle; I'm sure there are many difficult implementation details.

Heck, even charging $1 per IP address per year would free up vast numbers of addresses.

E.g. here's a thought example: MIT gets a bill for $16,277,216 for the use of their IP block for the next month. Think they'd pay? How about $16,277,216 for the next year? I don't think they'd pay that either.

But there are far too many sacred cows out there for something like you suggest to be practical. As someone else said just below: "Internet does not work like that. Not even close". So we'll instead spend hundreds of billions of dollars in the switchover.

Thanks for writing that. History was never my strong suite, and they don't usually teach real history in high school anyway.

I'd like to rephrase your analogy a little:

I'd say it's the computer equivalent of encountering a random whore, checking a list of names of infected whores that you carry with you, and then deciding to have unprotected sex with this whore. After all, her name isn't on your list.

Unfortunately I didn't work on my rephrasing for long enough to completely maintain the spirit in which your original was written. E.g. I didn't include the colorfully descriptive phrase "disease-ridden cheap whore". But you get the idea.

Tang. Don't forget Tang.

OK so far.

That's where you are wrong.

And you're wrong again. Waddell & Reed, the "fundamentals trader", was not selling stock. They were selling E-minis. Why? Because it's much more "productive" for them (and for the market in general) to sell a single contract than to sell the corresponding 500 stocks. E-minis, or similar instruments like the SPY (for smaller investors who can't play in "futures"), are very "productive" as part of a well executed investment strategy. It's much more cost effective for a smaller investor to buy and sell SPY than to buy and sell 20 or 30 stocks.

Clearly, things went wrong. But demonizing E-minis (and the corresponding stock hedging that goes along with them) is the wrong takeaway.

I think you have the right idea, but what you propose would have also hurt some innocent investors. These investors entrusted their money to professional advisors to manage (these were small advisors, not the big fund companies like Fidelity). The moron advisors had "stop loss" orders in place on several of the stocks that had wild swings. So what happened was something like this (hypothetical stock, too lazy to Google for actual names involved):

XYZ trades at $60
flash crash begins, XYZ falls to $50
lazy advisor has set a "stop loss" order (which becomes a market order) at $50
flash crash continues; there is no one willing to make a market, so the "market order" in XYZ trades at $0.01 (because many institutions leave token tiny bids in place); the advisor has lost 99+% of his clients money!!!
a few minutes later, XYZ recovers to $60, which is where it should be

Fortunately for the individuals, the trades were "busted".

I think it's gross negligence for an advisor to have these stop loss orders on the books. IMO its better not to have them in place at all (advisors are paid to watch the markets!), but at the least they should be stop limit not stop loss. An advisor who doesn't understand the difference should be fined 100% of his net worth and should be permanently barred from managing other people's money. It would be beneficial to the gene pool to put bullets into the back of these moron advisor's skulls, but most people would consider that to be too harsh a punishment.

All US banks are very happy to offer many years worth of transaction history to any Federal agency that desires this information. Too bad they won't do the same for their putative "customers".

Those bulbs are very good, I use lots of them. The one thing I don't like is they are more squarish and less pear-shaped than standard incandescent bulbs; there are some fixtures where this different shape doesn't provide enough clearance.