Go to court and see if you can stop the thing on that argument..
Why should I want to STOP it?
I just want the vote count in elections to accuartely (enough) reflect the actual votes of the qualified voters. (By "enough" I mean so that the outcome is the same as if it were perfect unless the election is nearly a tie.)
IMHO the electoral system is SO corrupted - and has been since AT LEAST the introduction of electronic tabulation - that the resluts represent the choices of the political machine rather than the will of the people.
This is VERY dangerous. The purpose of elections is not to "be fair". It is to predict the outcome of civil wars well enough that the losers will not be tempted to stage them, thus stabilizing the government.
If corruption of election results becomes large enough to swing the balance of power, AND this becomes known, it is no longer convincing. Then it becomes solely a matter of time until some losing side decides to give the alternative a try.
So... this brings up a question. If I obtain a document indicating that a company broke the law, can that document be suppressed by saying it's copy righted? If so, that's a BIG problem.
That's the claim Scientology used to sue Keith Henson, and how he ended up in Canadian "exile".
Keith was a vocal critic of Scientology, claiming their internal manuals were a manual for criminal activity and medical fraud, and their activities had already cost a number of lives, sometimes under conditions that could be interpreted as premeditated murder. He tried to bring this to the authorities to prosecute them for their actions, partly by a publicity campaign about their activities, which included calling attention to excerpts of leaked copies of the (copyrighted) documents in question.
At one point he participated in a netnews discussion of one of the excerpts - in which another poster had quoted such an excerpt. Keith's reply included the snippet in question (about six steps into quoting previous postings which included it in the postings THEY quoted).
The Scientologists used this to seek, and obtain, a large damage award for copyright violation and an injunction against further "infringement" on their copyright. Keith violated that injunction and was cited for contempt, and is currently on the run (in Canada last I heard) as a result of his continued refusal to knuckle under.
Note that this was just under the basic copyright laws, without any recourse to the DMCA enhancements. Courts are very hard on those who ignore their rulings rather than obeying them unless/until they are recinded or overturned.
I wonder how many precincts in CA plan to use the Diebold system, with its well-known cracks, in the upcoming Gubernatorial Recall election.
With a broad field of candidates splitting the vote, and the field-leader taking the race, small margins could easily swing the election - which means a small number of compromised precincts could swing the election.
And with no human-readable audit trail, if you thought the stink over the Florida Presidential results was bad you ain't seen NOTHING yet.
The TWX used the same infrastructure as the AT&T telephone network. But it was regulated differently.... In particular: Every call (even local) was a toll call. Bummer!
Oops: Turns out there were two versions. Here is a writeup from the Phrack archives (1989!) on the history of the TWX network.
Wasn't TWX called AOL until a few days ago? Was the TWX network a remote ancestor of Q-Link, the service that became America Online?
Nope. No relation to Time Warner (whatever the X was for).
TWX was the TypeWriter eXchange. AT&T's answer to Western Union's Telex exchange back in the days of monopoly regulation on wired communication.
Both were mechanical Teletype (r) machines with built-in 110-baud FSK modems. (And they were the same FSK modem standard that is still used for 110 and 300 baud compatability modes today.)
The TWX used the same infrastructure as the AT&T telephone network. (Even the dials/touchtone pads/card autodialers/6-button control selectors installed in the Teletype were manufactured by Western Electric - the Bell System wholly-owned manufacturing arm.) But it was regulated differently.
In particular: Every call (even local) was a toll call. Bummer!
But then the regents of Dartmouth U got a bee in their bonnet: They were a University. A University was SUPPOSED to be in business to teach students. So this resouce should be available to The Students.
That's Dartmouth College! Retaining "College" in the name of the institution reflects the College's focus on undergraduate teaching, despite meeting all definitions of a University.
I stand corrected. B-)
And that focus on giving undergraduates a good education fits right in with both the Dartmouth BASIC timesharing program and the current free-VoIP move. It's great to see such traditions upheld!
(I was unaware of the distinction because I wasn't a Dartmouth student myself. But I have fond memories of Dartmouth Basic - on which I cut my programming teeth {somewhat unofficially} by late-night long distance from Ann Arbor back in '64-'65 - before I could get into a programming class and get access to the local machines.)
The only downside I see is the high costs of a VoIP phone, but once those start getting mass produced that should drop too.
They're not using standalone VoIP phones. They're using a VoIP softphone application on the students' (already required) PCs, with a headset plugged into the sound card.
Buy the "standard" headset for $50 at the campus store or use any old PC headset you've got kicking around for zero added hardware/software cost.
"What will the success/failure of VoIP on this scale have on telecom?"
Um, doesn't the telecom industry own much of the data backbone as well? When they quit making money from local service, they start making money on bandwidth.
They make a LOT more money selling a retail toll-call voice connection to a consumer than they do from selling the equivalent amount of bandwidth wholesale to an ISP or backbone provider.
A LOT.
Like several orders of magnitude.
Think about it: One phone call, WITHOUT compression, is 64k bps. Your 1.5 Mbps download DSL link is the equivalent of one side of 24 simultaneous calls to anywhere in the world, 24 hours a day, seven days a week. Say you've only got 256k uplink: That's still 4 calls 24/7. Imagine what you'd be paying for THAT if it were retail long distance. What are you paying - retail - for your DSL? (And the tellcos are selling that wholesale to your ISP.)
Assume it's only three orders of magnitude: That turns every billion dollars of revenue into a million. Yet it doesn't decrease their costs of operation enough to mention.
Add in the fact that their profit is the DIFFERENCE between cost and revenue, and they aren't running anywhere near 10,000% return on investment, and you can see what a disaster it is for the telcoms.
Then add in the fact that they haven't amortized (paid off) their current infrastructure - which had maybe a 30-year ammortization schedule - and you can see that they're caught between a rock and a hard place. Drop the prices, they can't pay the interest on their bonds. Bankruptcy.
That is why the telecoms cut WAY back on buying new equipment for voice telephony - strangling or bankrupting a bunch of suppliers in the "telecom crash" - which is still underweigh. And why they're frantically trying to convert their networks to VoIP, before the upstarts eat them alive.
Actually it was set to 1.5/128 but they have since upgraded it to 256 up. I am not complaining at all.
Sounds like they have it throttled back to match ADSL speeds.
Since I've had Comcast I have been able to get almost my full advertised rates any time of the day or night. I would have complaints if they started download limits though.
They're probably doing that on purchased systems (i.e. AT&T Cable) where the users are used to uncapped service at some horrendous rate and then get annoyed when they get slowed down to 20 meg by congestion.
Dartmouth has quite a tradition of making hi-tek utilities free to their students. In particular:
Back in the bad old days of computing "a computer" was a room full of million-buck grey boxes attended by white-coated priests with PhDs. Any user who was not a member of the priesthood (and some who were) was billed by the second for its use and had to hand in his job at the window as a deck of punched cards, coming back hours later for the printed and maybe punched results.
An invention was made in these days: "Time Sharing". (A computer running a multitasking OS that in turn runs multiple copies of a command language processor, each copy serving a separate, directly-connected user. Think "dialup shell account".)
At first it was limited to fancy directly-connected terminals. Then a relatively cheap multple-teletype interface was invented to use the relatively-cheap TWX machines as terminals. Mechanical Teletype (r) machines, typically running 110 baud 8-bit ASCII. And a few, expensive, "Dataphone" modems could be used to allow remote teletypes to dial in over the TWX network.
But CPU time was still billed by the second, as was connect time on the expensive dialup lines or the less expensive directly-connected terminals.
But then the regents of Dartmouth U got a bee in their bonnet: They were a University. A University was SUPPOSED to be in business to teach students. So this resouce should be available to The Students.
Not just students taking a computer class. Not just grad students on a special, sponsored, project. ALL the students. ALL the time. NO bills.
So Dartmouth put in a bunch of Teletypes, all over campus. And wired them to the Computing Center. And gave EVERY student an account. Even entering freshmen. All of 'em. CPU time, disk storage, the whole shebang.
And because they couldn't afford the manpower to babysit the entire student body they invented a very easy-to-teach interpreted computer language, with a built-in, simple, text-file editor. And wrote manuals and lessons that could be read (and run) on-line.
You've probably heard of it.
It was called BASIC.
A fellow named Gates got his start in the industry by porting it to the Altair - the first home computer.
So it doesn't surprise me AT ALL, now that voice telephony is becoming a "marginal good" (i.e. "too cheap to meter", like electirc elevators without ticket-takers or coin slots) that Dartmouth should be the first institution to make it available to their people without an extra fee.
You are throttled to 1.5/256 by the capacity of your link to the next hop (just as you are throttled to, say, 2400 KW by the size of your electric feed).
I am throttled by the config file my CM recieves from Comcasts network when it boots up, not by limits of my line or capabilities of the local feed.
Oops. Sorry. I thought you had DSL, not cable. The 1.5/265 sounded similar to a typical ADSL rate - the 256 is a bit low for the uplink, but the 1.5 matches on the downlink - (and also matches a T1, which is no doubt not a coincidence.)
OK, so you're throttled to the rate they set, which is the same as an ADSL line on down and near it on up. Was that what they advertised and what you contracted for? Or was it something they imposed later, as a punishment for "overuse" of their system? (Former is a contract, latter is, IMHO, fraud on their part.)
SOMEBODY has to lose. So what's fair?
No idea but I really have no idea what you are trying to suggest either;)
Just trying to suggest that, since delivering full-rate during traffic peaks is impossible, the ISPs should distribute the degredation equally among the customers, not try to shed those customers who actually believed their advertising and used the system more than the ISPs planned for.
When demand is higher than capacity, throttle back all users in proportion to their current demand.
When the service is maxxed out, everyone will be throttled naturally.
But not necessarily equally.
If I understand it correctly, the throttling is not explicitly forced to be proportional to anything, but is simply the result of packet dropping due to congestion. This is roughly proportional to the current ATTEMPTED throughput of each user.
But first it's a statistical artifact - when things get busy packets get randomly dropped. Users with a bunch of packets coming or going will TEND to get more dropped, but there's a lot of noise.
Second: The current load of existing TCP connections will ALREADY be throttled back by previous packet drops. So if somebody is, say, transferring a file or listening to a stream he'll already be throttled back by the TCP mechanism. He's being well behaved - but more congestion will throttle him back still further. Meanwhile, somebody who suddenly started a bunch of new connections (which haven't yet throttled back) gets a lot more data than the well-behaved long-established TCP connection or smooth stream feed.
Are you saying throttle the person who is doing a sustained downloads and let the burst users have unlimited?
That's not quite true. The throttling is not explicit, but implied by the statistical
What is the difference as the total bandwidth required would be the same overall. I am already throttled, I pay for 1.5/256, why should I be throttled again?
You are throttled to 1.5/256 by the capacity of your link to the next hop (just as you are throttled to, say, 2400 KW by the size of your electric feed). But if EVERYBODY tries to use their full capacity at the same time, the infrastructure BEHIND the first hop will not be able to carry it all. SOMEBODY has to lose. So what's fair?
I think that we can all agree that it's NOT fair to penalize someone who used a bunch of off-peak capacity and isn't using anything exceptional now that there's a crunch. And it's ESPECIALLY not fair to disconnect him for using off-peak capacity habitually.
I claim that what is fair is for everybody's total bandwidth to be scaled back in proportion to their link speed. (An "internet brownout" if you will.) (I phrased my original suggestion badly: Throttling back in proportion to usage would only do this if everybody had the same speed of connection.)
If slowing everybody's effective link speed down by 10% is what's necessary to get the non-throttling packet drops to go away, that's what should be done. During a crunch, divide the bandwidth among the active users in proportion to their subscribed link speed.
And if the ISP has QoS (Quality of Service) implemented, drop the lower QoS packets first (to keep from interrupting net telephony and low-volume real-time streams until all other traffic to that drop has been throttled away.)
Now it might also be argued that it would be fair, in computing the individual's throttling rate, to look at usage in the last couple minutes and throttle back further those who already got a bunch of data through. Complicates things, and might set up an oscilation if not done properly. But it would let up/downloaders get high average throughput without zorching bursty users like web surfers.
But penalizing - or terminating - users on the basis of daily or monthly history or total bytes downloaded is both horrendously unreasonable and displays massive ignorance (and/or panic) on the part of the executives running the ISP.
Does your bandwidth quota get used up if someone launches a DoS attack on you?
Yes.
Note that one of the recommendations in the warning notice from Comcast is to install or update antivirus software.
The ISP could care less who originated the traffic or why. They only care about the amount of traffic.
Imagine if the phone company - which uses a similarly oversubscribed design for unlimited residential phones - cut off the service of people who receive a lot of calls, rather than giving out an "all-trunks busy" signal when the network congests, then treating excessive occurrences of all-trunks busy as a bug and increasing the infrastructure to reduce it to near-zero. (Similarly with slow dialtone for outgoing calls.)
I would normalize the distrobution periodically and set the cap at the higher 10% level.
I'd just throttle based on congestion, not on usage history. (Or just let the congestion take its course, as the internet was designed to do in the first place.)
Who cares if the "power users" max out when the net capacity is otherwise idle? When demand is higher than capacity, throttle back all users in proportion to their current demand. To do otherwise, without clear notice in the advertisements that you intend to do so, is fraud.
Comcast didn't include enough backbone bandwidth in their business model for people to actually USE the service they sold? Then they can buy some more, or let somebody who actually delivers what they promise take over when the users get fed up with substandard service (AND pay off their former customers for the service failure if said customers decide to press the issue).
It loos like we said the same thing a few minutes away from eachoter.:)
Yep.
I did not know that dipole fields drop off with the cube of the distance until just now and still don't really fully understand it.
- Inverse square because the area the wave is spread out across is increasing with the square of the distance from the emitter, times:
- Inverse first power because as you get farther away from the dipole (a + and a - pole near each other) the two opposite electric charges or magnetic poles appear closer together and cancel better.
Multiply the two effects and you get inverse cube.
Similarly, an infinite line source only falls off with inverse FIRST power because it's effectively only spreading out in one dimension, like an expanding cylinder. The weakening of radiaton from any point due to spreading along the axis dimension is canceled by the increasing significance of the contribution from point emitters located farther along the line source. VERY close to the conductor of an antenna element you get this effect, because the nearby region is a close approximation to a line source.
do you know of a good site that explains this sort of thing?
Sorry. I'm recalling this from college physics.
does anything fall off with the 1/r^4 of the distance for example?
Yes: Quadrapole moment. Imagine a square with + in the upper-right and lower-left corners, - in the upper-left and lower-right. Electric field direction is an X shape, with one crossbar having arrowheads and the other arrowtails, and it falls off with inverse-fourth-power, i.e. very, very fast. It's like the extra inverse-linear factor in dipole moment, except you have that twice: Once for the dimension of each pair of charges as a dipole, again for the other dimension as the angular separation of the dipoles also shrinks with distance.
Correct me if i'm wrong but dont most Radio transmission technologies use some form of magnetic induction in order to achieve their goal. Last i heard passing electricity through a coil produces a magnetic field. Whats new here?
What's new is that they've goofed.
At any given frequency you can launch an electromagnetic wave by using:
- And electric dipole. (Essentially impossible at anything above DC due to the current from the moving charges.)
- A permanent magnet or a current loop (producing a virtual magnetic dipole).
- A combination of the two, to produce the electric and magnetic fields simultaneously.
With a current loop the field very near the loop is essentially pure magnetic and falls off as the first power of distance (as more of the wire's length becomes signficant to the observer).
Moving out a bit more, in the first two the field moderately NEAR the antenna is essentially pure electric or magnetic (respectively) and falls off as a dipole field - with the cube of the distance. (Inverse square for each "pole" of the dipole, times inverse first-power for the smaller angular separation of the poles as viewed by the distant observer.) In the third you get the same effect with both the electric and magnetic field, typical of ordinary antennas.
But the changes to the electric field produce a magnetic field, and vice-versa. By the time you're a wavelength or so away from a simple driven element an electromagnetic field - a "radio wave" - has peeled off. This weakens at inverse-square rate (once you're far enough from the emitter that local additions and cancelations from different parts of it don't confound the issue.)
For signals in the tens of kilohertz and less (audio, for instance), a wavelength is very long. So a coil acts like more like a dipole than an antenna for a long way. Inverse-cube attenuates the signal rapidly with distance (though a strong amplifier can pull it back up - along with any competing noise).
But for computers you'll probably want this to operate at high speed - for images, disk access, etc. Now you're talking megahertz - with coding schemes that end up putting essentially all the informaiton at high frequencies. So the radio-wave effect takes over quickly, and the signal propagates without serious attenuation, regardless of whether the emitter is a magnetic loop (B-field) or electric dipole (E-field) emitter.
The guys operating the TEMPEST equipment will LOVE this system, thanks to the unjustified feeling of security it will give the user.
Home to school to home was 400 miles (Fort Wayne, Indiana, to Terre Haute and back) while I was in college. Having a vehicle with a range of only 300 miles would have been sh*t for the parents who took me to school and back.
The only downside to electric cars of this sort is the charging time. But for long trips there's a simple way around that.
An automobile needs high horsepower for accelleration. But for cruise it requires very little. On the straight-and-level it depends mainly on air and rolling friction, and 18 horses might suffice for even a moderately large sedan, significantly less for a streamlined sports car.
The electric car described has stored battery power enough to both accellerate from stop repeatedly and climb mountains, and uses regenerative braking to salvage much of the gravitational potential and momentum when going downhill and stopping. So the problem for range extension is just to replace the average straight-and-level cruising losses.
Refueling mid-trip is out, due to charging time. But (if I recall the article correctly) the charging circuit is capable of charging the car at about twice the average consumption on a long trip.
So one solution for longer trips is obvious: A small trailer, about the size of a motorcycle sidecar, containing a gasoline generator and a fuel tank. The generator tops off the batteries while the batteries provide the accelleration and climbing power (so you don't hold up traffic in the mountains, as "eco-friendly" gasoline cars do).
The article talks of charging the car from an oven-style home outlet. Let's be pessimistic and say it's not a standard 30-amp oven but more like 50 amps (at 240 volts). That's 12 KW, about 16.09 HP (plus generator losses). So figure about a 20-25 horse engine (so 17 horses is near the peak of its efficiency curve and you can run it at that long-term), a bit less (since the efficiency of the SYSTEM might be better if you reduce the weight of the engine a bit and run it a tad fast), or even QUITE a bit less (since you don't have to charge at a rate that lets you run 70 MPH for 24/7, but can start out charged and finish up mostly drained).
It's not just a speculation: Follow some of the links you see when googling "tzero charging time" and you'll see such a trailer hanging behind another model of electric car.
If the car is set up for it you can reduce the weight of the trailer by leaving off the starting battery and starting the engine from the battery in the car. (Leave off the starter motor, too, using the generator as a motor for startup.) The car's computer can direct the trailer engine to only run when required, eliminating the idling losses and running it at the peak of its efficiency curve, and arranging proper warmup after start before putting load on the engine.
If they could just give this car six times the range at one eleventh the cost... then it would be competive with my new Honda Civic Hybrid for commuting to work.
The car goes 300 miles on a charge. You have an 1,800 mile round-trip commute???
A nut grows in a single season. The carbon in the nut can only come from CO2 in the atmosphere. Therefore burning nuts is carbon neutral over a single nut-growing season.
That depends on what would have happened to the nut if you DIDN'T burn it.
If some of it would have been, say, dumped in a landfill to rot, that carbon would be sequestered from the atmosphere for decades-to-millenia, depending on circumstances. Nutshells are HARD, and woody material a couple feet underground can easily last for archeological, occasionally geological, time. (That's where coal came from.)
On the other hand, if you would have burned it ANYHOW, making power from it may let you avoid burning some other carbon-containing fossil fuel in addition. The fossil-carbon you avoided burning is your gain.
I'm not sure that "insecure by design" is quite fair to the hard-working folks who developed this near-ubiquitous MTA.
Actually, it was.
When Eric Allman first wrote it, it was to be installed on some large number of machines at UCB. And of course as a work in progress it needed a bunch of tweaking. So for his own convenience he included a "wizard mode" backdoor to give himself a remote root exploit on the machines in question. When you're publishing the source (so readers can discover the backdoor) you really can't get more "insecure by design" than that. B-)
Unfortunately, the code got cloned into general use with the wizard mode backdoor still in place. B-( So that was one of the first exploits to get patched out.
= = = =
But all kidding aside...
The original code was written back in the dark ages, when buffer overflows were a "bug" rather than a "security hole". Buffer overflow exploits were almost unheard of and a wizard-level stunt, rather than a newspaper topic and a script-kiddie classic. With gets(3S) in the standard library and heavily used, it's hardly surprising that sendmail had a bunch of buffer-overflow vulnerabilities, and one of 'em has escaped detection until now.
Sendmail was a very important piece of work. And its continued large market share today (despite arguably more secure, cleaner, and easier-to-use replacements) is a testimony to its utility and its author's contribution to the net.
Hey, I'm sorry for the flame, and particularly the name calling. I'd mail you personally to apologize if I could. I saw red after reading this thread and lost my cool.
Consider it delivered and accepted. B-)
This topic interests me though. I'm an H-1B myself. I work at a research lab that stopped hiring non-citizens after 9/11 and we have a staffing problem. I'm leaving what has become an awkwardly restrictive environment,
Sorry to hear that.
(Being "collateral damage" is no fun, even if you do get to live through it.)
[...] and am having a really hard time finding someone to hire into my post.
If you look around the top CS departments in the US, you'll find a real scarcity of talented homegrown graduate students.
And there you, or your site's administrators, are making a classic mistake.
Why are you seeking talend ONLY among:
- Recent Graduates or Grad Students of
- CS departments of top universities?
You're limiting yourself to an artificially constrained market. Consider:
For starters the couple universities with the "top CS departments" limit their enrollment - largely to the children of the idle rich. (I could show you how this is done while still looking "open" and "diversified" on paper.) Rich people have no monopoly on smarts. So try looking at the top people from the better state universities, or from a variety of schools with competent (if not top-in-the-nation) programs. You should easily find somebody with more ability and savvy than even the best of the preppie crowd. The competition for him or her will be less. Such graduates will often display more open-mindedness and/or common sense than a top-tier grad. ("It wasn't taught at my school so it can't be important!") And if your co-workers can handle working as peers with a smart woman or minority member, you're more likely to find one at a lower-tier school.
I can't tell if you're talking about a university research setting or a corporate R&D operation. Universities have a vested interest in dealing only with other academics - maintaining the actual and perceived value of degrees and using funds to support their students rather than random people. So while they can get some opportunity improvement by looking at students of other than the top couple schools, they're still limited to current students and recent grads - which is a drastic squeeze. But if it's corporate, the potenital for hires is FAR greater.
Why a recent grad? Why not a somebody with some experience, that you don't have to break in? Desipte their reputation for "creativity", new grads are constantly be making newbie mistakes, while an old-hand will have workable solutions at his fingertips - along with a host of knowlege about things that LOOK like they should work but DON'T - so you can avoid spending months or years down a plausible dead-end that he's seen kill several companies in the past. Old hands don't fossilize once they get out of school. (Research shows intelligence actually INCREASES somewhat with age.) Instead they accumulate experience, canned solutions to common problems (rather than reinventing the wheel new-grad style), and a collection of new ideas that they'd love to try out - if they just had the right environment. Meanwhile they keep up with the academic advances in their field AND with the stuff that's only in the trade journals now, that the academics will finally teach maybe 10 years later - if ever.
While you're at it, don't limit yourself to graduates. Some of the top people in the industry are old hands who came up through the ranks or dropped out of school short of even a bachelor's degree for any of a number of reasons that may be POSITIVELY correlated with talent. For instance: Focus on their specialty to the exclusion of distribution-requirement sidetracks. "Working their way through college" and stopping when their outside job became a lucrative carreer. Heck: Even some of Minsky's original MIT h
In case you didn't notice, the ruling does NOTHING to create a media monopoly. What it does is increase, to 45%, the potential audience that may be reached by stations owned by a single network entity.
This does NOT say that a nework can own all the stations in a market. It does not even say they can own ONE station in every market. They can't even own ONE station per area where the aggregate of all the stations they own reach even HALF the people in the country!
This is simple a change in the balance of power and money flow between the networks and the privately owned local stations affiliated with the networks, allowing the networks to directly own and operate a somewhat larger fraction (though still less than half of the potential audicnce worth) of the stations, rather than licencing their feed to a station run as a separate business by a separate owner.
Can you IMAGINE the government applying a similar rule to newspapers, or printers? "One entity can only own enough newspapers to reach 45% of the potential newspaper readers (i.e. voters)!" "One can only own enough printing presses to print manefestos for 45% of the population!"
Or how about this: "One entity can have network connectivity to reach no more than 45% of the population!"
Free speech advocates would be screaming about censorship!
(And then there's the tie-in with newspapers - where the same entity can't own a paper and a news station in the same city.)
THIS is why your networks are run by conglomerates, (all spitting the establishment / politically-correct party line) while alternative views are restricted to cable channels and syndicated talk-shows (when they appear at all): Nobody can buy up little radio or TV stations and set up a national-reach alternative.
I really want to see the Supreme Court rule on whether the FCC should be able to impose such a rule at all, or if the whole limitation must be struck down as incompatabile with the First Amendment.
Disaster - How the 'Look-Say" method got started
on
Can You Raed Tihs?
·
· Score: 1
Note that the research was done on experienced readers.
The disastrous "Look-Say" method of teaching English reading was started by a similar piece of research.
The researchers studied the better readers in elementary school and found that they were reading words as chunks.
The goal was to get all the students to be good readers. So they designed a program where they TAUGHT words as chunks, discarding teaching the language as a phonetic code. This method was called the "Look-Say" method, as opposed to the earlier "Phonics" method.
And we ended up with a generation of illiterates who couldn't handle any word they hadn't explicitly encountered in class. They'd graduate with a written vocabulary in the hundreds of words, if that, with strange tendencies to misspell, and with no.
The falacy is that the good students went through a progression of learning stages. First they learned letters and letter-groups as a phonetic code. Then as they gained proficiency they'd rote-memorize shortcuts for sylables, word fragments, entire common words, then progressively less common words. But when they encountered an unfamilar word they'd drop back as many levels as necessary to crack it, to the phonetic level if necessary, and do it automatically. (That's why you probably didn't even stumble the first time you encountered "Lite Beer" {or "Lite [whatever]} rather than "Light Beer".)
Victims of the Look-Say method were taught the language, not as an alphabetic language with a phonetic code, but as a hieroglyphic language, where each word is a separate picture and must be learned separately.
= = = =
The observation that internally-scrambled words are easily read seems to me to be one applicable only to advanced readers. Try it on a third-grader and I bet you'll get different results.
This limits its usefulness, and must be taken into account if you intend to do anything with it.
(Of course it might be useful to hide adult subjects from children - at least until they are very good readers. B-) )
I guess your family sprouted from the ground in North America, rather than coming here looking for work?
I've traced a couple branches back to the revolution. Before that they tended to end up here either because they were thrown out of there or fled from there to avoid persecution.
That's why they HAD a revolution - and why certain "dead white males" laid the foundation that eventually freed the slaves - both here and from under a stack of tyrannies back in Europe and elsewhere.
As to my wife (who tends to agree with me on thses subjects), one branch of her family came over on the Mayflower while another met the boat.
I take it, from your lack of understanding that some of us may be descended from people who came here for reasons other than moving to money (like for instance avoiding glaciers while working on exterminating the terror birds and mammoths, and grooming the rest of the continent into a low maintainence farm) that your predecessors are part of the "Ellis Island Crowd"?
Not that it matters.
What DOES matter is whether you personally have the integrity to let others live their lives, or insist on molding them into your own politically correct image and flaming them when they fail to conform.
But immigrants (who go through channels) are still welcome. The issue with the H1s is not the immigrants themselves, but the managers who break the law in order to obtain H1s for some other country's huddled masses, in order to make them into their company's serfs while laying off, or refusing to hire, perfectly qualified native-born or naturalized US citizens.
Why do you think there are so few US-born blacks, hispanics, or women employed in hi-tek - as other than janitors, secretaries, and food-service workers? It's not that there aren't any who are qualified, you know.
One of my former colleagues, when we were visiting Doug one day, had the bright idea of having him autograph his mouse.
Doug duly autographed it - and mentioned that this was the first time anybody had asked him. (This was in the late '80s or early '90s, so it wasn't like nobody had had the opportunity.)
So at that point he had the only Engelbart-autographed mouse. (And even if somebody else has asked since - which the rest of us didn't to avoid me-too ism and maintain the value of HIS mouse - he still has the first.)
Slashdot Mirror
Go to court and see if you can stop the thing on that argument..
Why should I want to STOP it?
I just want the vote count in elections to accuartely (enough) reflect the actual votes of the qualified voters. (By "enough" I mean so that the outcome is the same as if it were perfect unless the election is nearly a tie.)
IMHO the electoral system is SO corrupted - and has been since AT LEAST the introduction of electronic tabulation - that the resluts represent the choices of the political machine rather than the will of the people.
This is VERY dangerous. The purpose of elections is not to "be fair". It is to predict the outcome of civil wars well enough that the losers will not be tempted to stage them, thus stabilizing the government.
If corruption of election results becomes large enough to swing the balance of power, AND this becomes known, it is no longer convincing. Then it becomes solely a matter of time until some losing side decides to give the alternative a try.
So... this brings up a question. If I obtain a document indicating that a company broke the law, can that document be suppressed by saying it's copy righted? If so, that's a BIG problem.
That's the claim Scientology used to sue Keith Henson, and how he ended up in Canadian "exile".
Keith was a vocal critic of Scientology, claiming their internal manuals were a manual for criminal activity and medical fraud, and their activities had already cost a number of lives, sometimes under conditions that could be interpreted as premeditated murder. He tried to bring this to the authorities to prosecute them for their actions, partly by a publicity campaign about their activities, which included calling attention to excerpts of leaked copies of the (copyrighted) documents in question.
At one point he participated in a netnews discussion of one of the excerpts - in which another poster had quoted such an excerpt. Keith's reply included the snippet in question (about six steps into quoting previous postings which included it in the postings THEY quoted).
The Scientologists used this to seek, and obtain, a large damage award for copyright violation and an injunction against further "infringement" on their copyright. Keith violated that injunction and was cited for contempt, and is currently on the run (in Canada last I heard) as a result of his continued refusal to knuckle under.
Note that this was just under the basic copyright laws, without any recourse to the DMCA enhancements. Courts are very hard on those who ignore their rulings rather than obeying them unless/until they are recinded or overturned.
I wonder how many precincts in CA plan to use the Diebold system, with its well-known cracks, in the upcoming Gubernatorial Recall election.
With a broad field of candidates splitting the vote, and the field-leader taking the race, small margins could easily swing the election - which means a small number of compromised precincts could swing the election.
And with no human-readable audit trail, if you thought the stink over the Florida Presidential results was bad you ain't seen NOTHING yet.
The TWX used the same infrastructure as the AT&T telephone network. But it was regulated differently. ... In particular: Every call (even local) was a toll call. Bummer!
Oops: Turns out there were two versions. Here is a writeup from the Phrack archives (1989!) on the history of the TWX network.
dial in over the TWX network.
Wasn't TWX called AOL until a few days ago? Was the TWX network a remote ancestor of Q-Link, the service that became America Online?
Nope. No relation to Time Warner (whatever the X was for).
TWX was the TypeWriter eXchange. AT&T's answer to Western Union's Telex exchange back in the days of monopoly regulation on wired communication.
Both were mechanical Teletype (r) machines with built-in 110-baud FSK modems. (And they were the same FSK modem standard that is still used for 110 and 300 baud compatability modes today.)
The TWX used the same infrastructure as the AT&T telephone network. (Even the dials/touchtone pads/card autodialers/6-button control selectors installed in the Teletype were manufactured by Western Electric - the Bell System wholly-owned manufacturing arm.) But it was regulated differently.
In particular: Every call (even local) was a toll call. Bummer!
But then the regents of Dartmouth U got a bee in their bonnet: They were a University. A University was SUPPOSED to be in business to teach students. So this resouce should be available to The Students.
That's Dartmouth College! Retaining "College" in the name of the institution reflects the College's focus on undergraduate teaching, despite meeting all definitions of a University.
I stand corrected. B-)
And that focus on giving undergraduates a good education fits right in with both the Dartmouth BASIC timesharing program and the current free-VoIP move. It's great to see such traditions upheld!
(I was unaware of the distinction because I wasn't a Dartmouth student myself. But I have fond memories of Dartmouth Basic - on which I cut my programming teeth {somewhat unofficially} by late-night long distance from Ann Arbor back in '64-'65 - before I could get into a programming class and get access to the local machines.)
The only downside I see is the high costs of a VoIP phone, but once those start getting mass produced that should drop too.
They're not using standalone VoIP phones. They're using a VoIP softphone application on the students' (already required) PCs, with a headset plugged into the sound card.
Buy the "standard" headset for $50 at the campus store or use any old PC headset you've got kicking around for zero added hardware/software cost.
"What will the success/failure of VoIP on this scale have on telecom?"
Um, doesn't the telecom industry own much of the data backbone as well? When they quit making money from local service, they start making money on bandwidth.
They make a LOT more money selling a retail toll-call voice connection to a consumer than they do from selling the equivalent amount of bandwidth wholesale to an ISP or backbone provider.
A LOT.
Like several orders of magnitude.
Think about it: One phone call, WITHOUT compression, is 64k bps. Your 1.5 Mbps download DSL link is the equivalent of one side of 24 simultaneous calls to anywhere in the world, 24 hours a day, seven days a week. Say you've only got 256k uplink: That's still 4 calls 24/7. Imagine what you'd be paying for THAT if it were retail long distance. What are you paying - retail - for your DSL? (And the tellcos are selling that wholesale to your ISP.)
Assume it's only three orders of magnitude: That turns every billion dollars of revenue into a million. Yet it doesn't decrease their costs of operation enough to mention.
Add in the fact that their profit is the DIFFERENCE between cost and revenue, and they aren't running anywhere near 10,000% return on investment, and you can see what a disaster it is for the telcoms.
Then add in the fact that they haven't amortized (paid off) their current infrastructure - which had maybe a 30-year ammortization schedule - and you can see that they're caught between a rock and a hard place. Drop the prices, they can't pay the interest on their bonds. Bankruptcy.
That is why the telecoms cut WAY back on buying new equipment for voice telephony - strangling or bankrupting a bunch of suppliers in the "telecom crash" - which is still underweigh. And why they're frantically trying to convert their networks to VoIP, before the upstarts eat them alive.
Actually it was set to 1.5/128 but they have since upgraded it to 256 up. I am not complaining at all.
Sounds like they have it throttled back to match ADSL speeds.
Since I've had Comcast I have been able to get almost my full advertised rates any time of the day or night. I would have complaints if they started download limits though.
They're probably doing that on purchased systems (i.e. AT&T Cable) where the users are used to uncapped service at some horrendous rate and then get annoyed when they get slowed down to 20 meg by congestion.
But why Dartmouth?
Because Dartmouth students talk a lot.
But seriously...
Dartmouth has quite a tradition of making hi-tek utilities free to their students. In particular:
Back in the bad old days of computing "a computer" was a room full of million-buck grey boxes attended by white-coated priests with PhDs. Any user who was not a member of the priesthood (and some who were) was billed by the second for its use and had to hand in his job at the window as a deck of punched cards, coming back hours later for the printed and maybe punched results.
An invention was made in these days: "Time Sharing". (A computer running a multitasking OS that in turn runs multiple copies of a command language processor, each copy serving a separate, directly-connected user. Think "dialup shell account".)
At first it was limited to fancy directly-connected terminals. Then a relatively cheap multple-teletype interface was invented to use the relatively-cheap TWX machines as terminals. Mechanical Teletype (r) machines, typically running 110 baud 8-bit ASCII. And a few, expensive, "Dataphone" modems could be used to allow remote teletypes to dial in over the TWX network.
But CPU time was still billed by the second, as was connect time on the expensive dialup lines or the less expensive directly-connected terminals.
But then the regents of Dartmouth U got a bee in their bonnet: They were a University. A University was SUPPOSED to be in business to teach students. So this resouce should be available to The Students.
Not just students taking a computer class. Not just grad students on a special, sponsored, project. ALL the students. ALL the time. NO bills.
So Dartmouth put in a bunch of Teletypes, all over campus. And wired them to the Computing Center. And gave EVERY student an account. Even entering freshmen. All of 'em. CPU time, disk storage, the whole shebang.
And because they couldn't afford the manpower to babysit the entire student body they invented a very easy-to-teach interpreted computer language, with a built-in, simple, text-file editor. And wrote manuals and lessons that could be read (and run) on-line.
You've probably heard of it.
It was called BASIC.
A fellow named Gates got his start in the industry by porting it to the Altair - the first home computer.
So it doesn't surprise me AT ALL, now that voice telephony is becoming a "marginal good" (i.e. "too cheap to meter", like electirc elevators without ticket-takers or coin slots) that Dartmouth should be the first institution to make it available to their people without an extra fee.
You are throttled to 1.5/256 by the capacity of your link to the next hop (just as you are throttled to, say, 2400 KW by the size of your electric feed).
;)
I am throttled by the config file my CM recieves from Comcasts network when it boots up, not by limits of my line or capabilities of the local feed.
Oops. Sorry. I thought you had DSL, not cable. The 1.5/265 sounded similar to a typical ADSL rate - the 256 is a bit low for the uplink, but the 1.5 matches on the downlink - (and also matches a T1, which is no doubt not a coincidence.)
OK, so you're throttled to the rate they set, which is the same as an ADSL line on down and near it on up. Was that what they advertised and what you contracted for? Or was it something they imposed later, as a punishment for "overuse" of their system? (Former is a contract, latter is, IMHO, fraud on their part.)
SOMEBODY has to lose. So what's fair?
No idea but I really have no idea what you are trying to suggest either
Just trying to suggest that, since delivering full-rate during traffic peaks is impossible, the ISPs should distribute the degredation equally among the customers, not try to shed those customers who actually believed their advertising and used the system more than the ISPs planned for.
When demand is higher than capacity, throttle back all users in proportion to their current demand.
When the service is maxxed out, everyone will be throttled naturally.
But not necessarily equally.
If I understand it correctly, the throttling is not explicitly forced to be proportional to anything, but is simply the result of packet dropping due to congestion. This is roughly proportional to the current ATTEMPTED throughput of each user.
But first it's a statistical artifact - when things get busy packets get randomly dropped. Users with a bunch of packets coming or going will TEND to get more dropped, but there's a lot of noise.
Second: The current load of existing TCP connections will ALREADY be throttled back by previous packet drops. So if somebody is, say, transferring a file or listening to a stream he'll already be throttled back by the TCP mechanism. He's being well behaved - but more congestion will throttle him back still further. Meanwhile, somebody who suddenly started a bunch of new connections (which haven't yet throttled back) gets a lot more data than the well-behaved long-established TCP connection or smooth stream feed.
Are you saying throttle the person who is doing a sustained downloads and let the burst users have unlimited?
That's not quite true. The throttling is not explicit, but implied by the statistical
What is the difference as the total bandwidth required would be the same overall. I am already throttled, I pay for 1.5/256, why should I be throttled again?
You are throttled to 1.5/256 by the capacity of your link to the next hop (just as you are throttled to, say, 2400 KW by the size of your electric feed). But if EVERYBODY tries to use their full capacity at the same time, the infrastructure BEHIND the first hop will not be able to carry it all. SOMEBODY has to lose. So what's fair?
I think that we can all agree that it's NOT fair to penalize someone who used a bunch of off-peak capacity and isn't using anything exceptional now that there's a crunch. And it's ESPECIALLY not fair to disconnect him for using off-peak capacity habitually.
I claim that what is fair is for everybody's total bandwidth to be scaled back in proportion to their link speed. (An "internet brownout" if you will.) (I phrased my original suggestion badly: Throttling back in proportion to usage would only do this if everybody had the same speed of connection.)
If slowing everybody's effective link speed down by 10% is what's necessary to get the non-throttling packet drops to go away, that's what should be done. During a crunch, divide the bandwidth among the active users in proportion to their subscribed link speed.
And if the ISP has QoS (Quality of Service) implemented, drop the lower QoS packets first (to keep from interrupting net telephony and low-volume real-time streams until all other traffic to that drop has been throttled away.)
Now it might also be argued that it would be fair, in computing the individual's throttling rate, to look at usage in the last couple minutes and throttle back further those who already got a bunch of data through. Complicates things, and might set up an oscilation if not done properly. But it would let up/downloaders get high average throughput without zorching bursty users like web surfers.
But penalizing - or terminating - users on the basis of daily or monthly history or total bytes downloaded is both horrendously unreasonable and displays massive ignorance (and/or panic) on the part of the executives running the ISP.
Does your bandwidth quota get used up if someone launches a DoS attack on you?
Yes.
Note that one of the recommendations in the warning notice from Comcast is to install or update antivirus software.
The ISP could care less who originated the traffic or why. They only care about the amount of traffic.
Imagine if the phone company - which uses a similarly oversubscribed design for unlimited residential phones - cut off the service of people who receive a lot of calls, rather than giving out an "all-trunks busy" signal when the network congests, then treating excessive occurrences of all-trunks busy as a bug and increasing the infrastructure to reduce it to near-zero. (Similarly with slow dialtone for outgoing calls.)
I would normalize the distrobution periodically and set the cap at the higher 10% level.
I'd just throttle based on congestion, not on usage history. (Or just let the congestion take its course, as the internet was designed to do in the first place.)
Who cares if the "power users" max out when the net capacity is otherwise idle? When demand is higher than capacity, throttle back all users in proportion to their current demand. To do otherwise, without clear notice in the advertisements that you intend to do so, is fraud.
Comcast didn't include enough backbone bandwidth in their business model for people to actually USE the service they sold? Then they can buy some more, or let somebody who actually delivers what they promise take over when the users get fed up with substandard service (AND pay off their former customers for the service failure if said customers decide to press the issue).
It loos like we said the same thing a few minutes away from eachoter. :)
Yep.
I did not know that dipole fields drop off with the cube of the distance until just now and still don't really fully understand it.
- Inverse square because the area the wave is spread out across is increasing with the square of the distance from the emitter, times:
- Inverse first power because as you get farther away from the dipole (a + and a - pole near each other) the two opposite electric charges or magnetic poles appear closer together and cancel better.
Multiply the two effects and you get inverse cube.
Similarly, an infinite line source only falls off with inverse FIRST power because it's effectively only spreading out in one dimension, like an expanding cylinder. The weakening of radiaton from any point due to spreading along the axis dimension is canceled by the increasing significance of the contribution from point emitters located farther along the line source. VERY close to the conductor of an antenna element you get this effect, because the nearby region is a close approximation to a line source.
do you know of a good site that explains this sort of thing?
Sorry. I'm recalling this from college physics.
does anything fall off with the 1/r^4 of the distance for example?
Yes: Quadrapole moment. Imagine a square with + in the upper-right and lower-left corners, - in the upper-left and lower-right. Electric field direction is an X shape, with one crossbar having arrowheads and the other arrowtails, and it falls off with inverse-fourth-power, i.e. very, very fast. It's like the extra inverse-linear factor in dipole moment, except you have that twice: Once for the dimension of each pair of charges as a dipole, again for the other dimension as the angular separation of the dipoles also shrinks with distance.
Correct me if i'm wrong but dont most Radio transmission technologies use some form of magnetic induction in order to achieve their goal. Last i heard passing electricity through a coil produces a magnetic field. Whats new here?
What's new is that they've goofed.
At any given frequency you can launch an electromagnetic wave by using:
- And electric dipole. (Essentially impossible at anything above DC due to the current from the moving charges.)
- A permanent magnet or a current loop (producing a virtual magnetic dipole).
- A combination of the two, to produce the electric and magnetic fields simultaneously.
With a current loop the field very near the loop is essentially pure magnetic and falls off as the first power of distance (as more of the wire's length becomes signficant to the observer).
Moving out a bit more, in the first two the field moderately NEAR the antenna is essentially pure electric or magnetic (respectively) and falls off as a dipole field - with the cube of the distance. (Inverse square for each "pole" of the dipole, times inverse first-power for the smaller angular separation of the poles as viewed by the distant observer.) In the third you get the same effect with both the electric and magnetic field, typical of ordinary antennas.
But the changes to the electric field produce a magnetic field, and vice-versa. By the time you're a wavelength or so away from a simple driven element an electromagnetic field - a "radio wave" - has peeled off. This weakens at inverse-square rate (once you're far enough from the emitter that local additions and cancelations from different parts of it don't confound the issue.)
For signals in the tens of kilohertz and less (audio, for instance), a wavelength is very long. So a coil acts like more like a dipole than an antenna for a long way. Inverse-cube attenuates the signal rapidly with distance (though a strong amplifier can pull it back up - along with any competing noise).
But for computers you'll probably want this to operate at high speed - for images, disk access, etc. Now you're talking megahertz - with coding schemes that end up putting essentially all the informaiton at high frequencies. So the radio-wave effect takes over quickly, and the signal propagates without serious attenuation, regardless of whether the emitter is a magnetic loop (B-field) or electric dipole (E-field) emitter.
The guys operating the TEMPEST equipment will LOVE this system, thanks to the unjustified feeling of security it will give the user.
Home to school to home was 400 miles (Fort Wayne, Indiana, to Terre Haute and back) while I was in college. Having a vehicle with a range of only 300 miles would have been sh*t for the parents who took me to school and back.
The only downside to electric cars of this sort is the charging time. But for long trips there's a simple way around that.
An automobile needs high horsepower for accelleration. But for cruise it requires very little. On the straight-and-level it depends mainly on air and rolling friction, and 18 horses might suffice for even a moderately large sedan, significantly less for a streamlined sports car.
The electric car described has stored battery power enough to both accellerate from stop repeatedly and climb mountains, and uses regenerative braking to salvage much of the gravitational potential and momentum when going downhill and stopping. So the problem for range extension is just to replace the average straight-and-level cruising losses.
Refueling mid-trip is out, due to charging time. But (if I recall the article correctly) the charging circuit is capable of charging the car at about twice the average consumption on a long trip.
So one solution for longer trips is obvious: A small trailer, about the size of a motorcycle sidecar, containing a gasoline generator and a fuel tank. The generator tops off the batteries while the batteries provide the accelleration and climbing power (so you don't hold up traffic in the mountains, as "eco-friendly" gasoline cars do).
The article talks of charging the car from an oven-style home outlet. Let's be pessimistic and say it's not a standard 30-amp oven but more like 50 amps (at 240 volts). That's 12 KW, about 16.09 HP (plus generator losses). So figure about a 20-25 horse engine (so 17 horses is near the peak of its efficiency curve and you can run it at that long-term), a bit less (since the efficiency of the SYSTEM might be better if you reduce the weight of the engine a bit and run it a tad fast), or even QUITE a bit less (since you don't have to charge at a rate that lets you run 70 MPH for 24/7, but can start out charged and finish up mostly drained).
It's not just a speculation: Follow some of the links you see when googling "tzero charging time" and you'll see such a trailer hanging behind another model of electric car.
If the car is set up for it you can reduce the weight of the trailer by leaving off the starting battery and starting the engine from the battery in the car. (Leave off the starter motor, too, using the generator as a motor for startup.) The car's computer can direct the trailer engine to only run when required, eliminating the idling losses and running it at the peak of its efficiency curve, and arranging proper warmup after start before putting load on the engine.
If they could just give this car six times the range at one eleventh the cost... then it would be competive with my new Honda Civic Hybrid for commuting to work.
The car goes 300 miles on a charge. You have an 1,800 mile round-trip commute???
A nut grows in a single season. The carbon in the nut can only come from CO2 in the atmosphere. Therefore burning nuts is carbon neutral over a single nut-growing season.
That depends on what would have happened to the nut if you DIDN'T burn it.
If some of it would have been, say, dumped in a landfill to rot, that carbon would be sequestered from the atmosphere for decades-to-millenia, depending on circumstances. Nutshells are HARD, and woody material a couple feet underground can easily last for archeological, occasionally geological, time. (That's where coal came from.)
On the other hand, if you would have burned it ANYHOW, making power from it may let you avoid burning some other carbon-containing fossil fuel in addition. The fossil-carbon you avoided burning is your gain.
I'm not sure that "insecure by design" is quite fair to the hard-working folks who developed this near-ubiquitous MTA.
Actually, it was.
When Eric Allman first wrote it, it was to be installed on some large number of machines at UCB. And of course as a work in progress it needed a bunch of tweaking. So for his own convenience he included a "wizard mode" backdoor to give himself a remote root exploit on the machines in question. When you're publishing the source (so readers can discover the backdoor) you really can't get more "insecure by design" than that. B-)
Unfortunately, the code got cloned into general use with the wizard mode backdoor still in place. B-( So that was one of the first exploits to get patched out.
= = = =
But all kidding aside...
The original code was written back in the dark ages, when buffer overflows were a "bug" rather than a "security hole". Buffer overflow exploits were almost unheard of and a wizard-level stunt, rather than a newspaper topic and a script-kiddie classic. With gets(3S) in the standard library and heavily used, it's hardly surprising that sendmail had a bunch of buffer-overflow vulnerabilities, and one of 'em has escaped detection until now.
Sendmail was a very important piece of work. And its continued large market share today (despite arguably more secure, cleaner, and easier-to-use replacements) is a testimony to its utility and its author's contribution to the net.
Hey, I'm sorry for the flame, and particularly the name calling. I'd mail you personally to apologize if I could. I saw red after reading this thread and lost my cool.
Consider it delivered and accepted. B-)
This topic interests me though. I'm an H-1B myself. I work at a research lab that stopped hiring non-citizens after 9/11 and we have a staffing problem. I'm leaving what has become an awkwardly restrictive environment,
Sorry to hear that.
(Being "collateral damage" is no fun, even if you do get to live through it.)
[...] and am having a really hard time finding someone to hire into my post.
If you look around the top CS departments in the US, you'll find a real scarcity of talented homegrown graduate students.
And there you, or your site's administrators, are making a classic mistake.
Why are you seeking talend ONLY among:
- Recent Graduates or Grad Students of
- CS departments of top universities?
You're limiting yourself to an artificially constrained market. Consider:
For starters the couple universities with the "top CS departments" limit their enrollment - largely to the children of the idle rich. (I could show you how this is done while still looking "open" and "diversified" on paper.) Rich people have no monopoly on smarts. So try looking at the top people from the better state universities, or from a variety of schools with competent (if not top-in-the-nation) programs. You should easily find somebody with more ability and savvy than even the best of the preppie crowd. The competition for him or her will be less. Such graduates will often display more open-mindedness and/or common sense than a top-tier grad. ("It wasn't taught at my school so it can't be important!") And if your co-workers can handle working as peers with a smart woman or minority member, you're more likely to find one at a lower-tier school.
I can't tell if you're talking about a university research setting or a corporate R&D operation. Universities have a vested interest in dealing only with other academics - maintaining the actual and perceived value of degrees and using funds to support their students rather than random people. So while they can get some opportunity improvement by looking at students of other than the top couple schools, they're still limited to current students and recent grads - which is a drastic squeeze. But if it's corporate, the potenital for hires is FAR greater.
Why a recent grad? Why not a somebody with some experience, that you don't have to break in? Desipte their reputation for "creativity", new grads are constantly be making newbie mistakes, while an old-hand will have workable solutions at his fingertips - along with a host of knowlege about things that LOOK like they should work but DON'T - so you can avoid spending months or years down a plausible dead-end that he's seen kill several companies in the past. Old hands don't fossilize once they get out of school. (Research shows intelligence actually INCREASES somewhat with age.) Instead they accumulate experience, canned solutions to common problems (rather than reinventing the wheel new-grad style), and a collection of new ideas that they'd love to try out - if they just had the right environment. Meanwhile they keep up with the academic advances in their field AND with the stuff that's only in the trade journals now, that the academics will finally teach maybe 10 years later - if ever.
While you're at it, don't limit yourself to graduates. Some of the top people in the industry are old hands who came up through the ranks or dropped out of school short of even a bachelor's degree for any of a number of reasons that may be POSITIVELY correlated with talent. For instance: Focus on their specialty to the exclusion of distribution-requirement sidetracks. "Working their way through college" and stopping when their outside job became a lucrative carreer. Heck: Even some of Minsky's original MIT h
there goes my media monopoly
Say WHAT?
In case you didn't notice, the ruling does NOTHING to create a media monopoly. What it does is increase, to 45%, the potential audience that may be reached by stations owned by a single network entity.
This does NOT say that a nework can own all the stations in a market. It does not even say they can own ONE station in every market. They can't even own ONE station per area where the aggregate of all the stations they own reach even HALF the people in the country!
This is simple a change in the balance of power and money flow between the networks and the privately owned local stations affiliated with the networks, allowing the networks to directly own and operate a somewhat larger fraction (though still less than half of the potential audicnce worth) of the stations, rather than licencing their feed to a station run as a separate business by a separate owner.
Can you IMAGINE the government applying a similar rule to newspapers, or printers? "One entity can only own enough newspapers to reach 45% of the potential newspaper readers (i.e. voters)!" "One can only own enough printing presses to print manefestos for 45% of the population!"
Or how about this: "One entity can have network connectivity to reach no more than 45% of the population!"
Free speech advocates would be screaming about censorship!
(And then there's the tie-in with newspapers - where the same entity can't own a paper and a news station in the same city.)
THIS is why your networks are run by conglomerates, (all spitting the establishment / politically-correct party line) while alternative views are restricted to cable channels and syndicated talk-shows (when they appear at all): Nobody can buy up little radio or TV stations and set up a national-reach alternative.
I really want to see the Supreme Court rule on whether the FCC should be able to impose such a rule at all, or if the whole limitation must be struck down as incompatabile with the First Amendment.
Note that the research was done on experienced readers.
.
The disastrous "Look-Say" method of teaching English reading was started by a similar piece of research.
The researchers studied the better readers in elementary school and found that they were reading words as chunks.
The goal was to get all the students to be good readers. So they designed a program where they TAUGHT words as chunks, discarding teaching the language as a phonetic code. This method was called the "Look-Say" method, as opposed to the earlier "Phonics" method.
And we ended up with a generation of illiterates who couldn't handle any word they hadn't explicitly encountered in class. They'd graduate with a written vocabulary in the hundreds of words, if that, with strange tendencies to misspell, and with no
The falacy is that the good students went through a progression of learning stages. First they learned letters and letter-groups as a phonetic code. Then as they gained proficiency they'd rote-memorize shortcuts for sylables, word fragments, entire common words, then progressively less common words. But when they encountered an unfamilar word they'd drop back as many levels as necessary to crack it, to the phonetic level if necessary, and do it automatically. (That's why you probably didn't even stumble the first time you encountered "Lite Beer" {or "Lite [whatever]} rather than "Light Beer".)
Victims of the Look-Say method were taught the language, not as an alphabetic language with a phonetic code, but as a hieroglyphic language, where each word is a separate picture and must be learned separately.
= = = =
The observation that internally-scrambled words are easily read seems to me to be one applicable only to advanced readers. Try it on a third-grader and I bet you'll get different results.
This limits its usefulness, and must be taken into account if you intend to do anything with it.
(Of course it might be useful to hide adult subjects from children - at least until they are very good readers. B-) )
I guess your family sprouted from the ground in North America, rather than coming here looking for work?
I've traced a couple branches back to the revolution. Before that they tended to end up here either because they were thrown out of there or fled from there to avoid persecution.
That's why they HAD a revolution - and why certain "dead white males" laid the foundation that eventually freed the slaves - both here and from under a stack of tyrannies back in Europe and elsewhere.
As to my wife (who tends to agree with me on thses subjects), one branch of her family came over on the Mayflower while another met the boat.
I take it, from your lack of understanding that some of us may be descended from people who came here for reasons other than moving to money (like for instance avoiding glaciers while working on exterminating the terror birds and mammoths, and grooming the rest of the continent into a low maintainence farm) that your predecessors are part of the "Ellis Island Crowd"?
Not that it matters.
What DOES matter is whether you personally have the integrity to let others live their lives, or insist on molding them into your own politically correct image and flaming them when they fail to conform.
But immigrants (who go through channels) are still welcome. The issue with the H1s is not the immigrants themselves, but the managers who break the law in order to obtain H1s for some other country's huddled masses, in order to make them into their company's serfs while laying off, or refusing to hire, perfectly qualified native-born or naturalized US citizens.
Why do you think there are so few US-born blacks, hispanics, or women employed in hi-tek - as other than janitors, secretaries, and food-service workers? It's not that there aren't any who are qualified, you know.
One of my former colleagues, when we were visiting Doug one day, had the bright idea of having him autograph his mouse.
Doug duly autographed it - and mentioned that this was the first time anybody had asked him. (This was in the late '80s or early '90s, so it wasn't like nobody had had the opportunity.)
So at that point he had the only Engelbart-autographed mouse. (And even if somebody else has asked since - which the rest of us didn't to avoid me-too ism and maintain the value of HIS mouse - he still has the first.)