> the Space Shuttle's systems are run on proprietary > software written by a company that basically just > writes the shuttle software.
Specifically, of the five CPUs in the primary Shuttle management system, four run identical copies of the management/control software. This version was originally written by IBM's then Federal Systems Division, which was later sold (to Loral I believe). The fifth CPU runs code written to the same specifictions by the shuttle's prime contractor (then Rockwell, now Boeing). The two groups were (and I believe are) only allowed to communicate through formal written specifications and are never allowed to speak directly or to see one another's code. Whenever the software is changed, both versions must independently pass the same functional tests and then the entire cluster of five CPUs must pass the functional tests as a unit.
I haven't seen any detailed write-ups since they upgraded the Shuttle cockpit using what was essentially the Boeing 767 avionics, but I assume similar procedures still apply.
According to one history of the 1991 Gulf War that I read, a British planning officer in London lost his portable computer (they weren't laptops then) with quite a bit of critical information on it. The London police let it be known among their contacts that it would _really_ be best if it were to be returned no-questions-asked, and it was dropped off at a police station within a day.
In a similar case in one city I was living in, 4 people in two years tried to get their spouse murdered by hanging out at a bar known to be frequented by hardened criminals and striking up a bargain with a willing thug (don't ask me why we had so many of those cases in that burg!). In all 4 cases the thug went right to the police and got fitted out for a wire. As one of them said in an interview, "I am a professional burgler but that doesn't mean I don't have standards".
So maybe the guy who stole it decided it was best not to have the entire FBI and US Army on his tail and turned it back in.
Not necessarily, as in both the open outcry and sealed-bid 1st price auctions there are various incentives for participants not to bid their maximum personal valuation (most often the "loser's curse", but there are other reasons as well).
Robert Weber of Northwestern University has written about this extensively. The solution is the 2nd-price auction. In this system, everyone puts in one bid which is the maximum they would be willing to pay. The highest bidder gets the item, but pays what the _2nd_ highest bidder bid. That way there is no incentive to try to outguess your opponents' strategies or "snipe".
The concept is hard to explain, though, so it doesn't get used much.
Dvorak writes a lot of different stuff, including some real journalism and technology analysis. But his best known work, and that which I am sure earns him his bread-and-butter, is technology gossip. Like every gossip column ever written since the first traveling minstrel appeared on the scene 30,000 years ago, Dvoark's gossip columns consist of a mixture of truth, exaggeration, spin (whether planted by the technology companies or generated by Dvorak himself), trolling, and some totally made-up stuff.
For example, Dvorak has been trying to force the monitor companies to bring new technologies to market for at least the last 20 years. That is why he hypes-Hypes-HYPES any rumour of a new display technoloy (seen that 300 dpi Texas Instruments display he reported "almost ready for production" in 1995 yet?). 40% truth, 40% exaggeration, 20% Dvorak-generated spin.
But as I said, that is how gossip columns of any kind work. Don't like it, don't read it.
> Nobody would have predicted that most people would now > take pictures by holding the camera out in front of them > and look at the preview screen to frame a shot.
As opposed to the Rolliflex-type camera which you held (and still hold) at waist level, looking down into the viewfinder (just don't step forward when shooting the Grand Canyon). Or the view camera, where you put the black cloth over your head and stare into a piece of ground glass?
I prefer mechanical SLRs to digital, myself, but there were more types of cameras in 1920 than there are today.
> Its all fine and great, but this fine > DRM alternative called 'Freedom' doesn't > pay salaries
Have you ever dealt with anyone in the music Industry (LA/Nashville/New York)? The amount of money they skim from the artists for "salaries" is staggering and sickening. One music company I encountered had a policy that every employee of the company always flew first class and stayed at the Four Seasons or the closest local equivalent. _Everyone_, all the time. And by "everyone" I mean every single person, including the junior clerk going to SAP to take an A/P data entry class. Where do you think that money comes from? Why do you think the music industry is so hot to use the awesome coercive power of the State to continue the flow of that money into their coffers?
> There are at present only two methods for sifting > uranium atoms, or isotopes, to create the right mix.
AVLIS has been around since the 1970s, and there is also the South African cyclonic process. There are also hints in the public literature that there are other methods that were examined by the Manhatten Project and not pursued for various reasons.
In the end it all comes down to data. Languages come and go. Code comes and goes. But it is quite common to find data which has survived essentially intact in structure and content for 20, 30, even 40 years.
So don't start with any language. Start with building a deep understanding of data: what it is, how it is modeled, the techniques used to manage it. Since most data today is stored in relational databases, next proceed to understand how RDBMSs work in theory and in practice, and how to write good performant SQL. If at the end of the day you understand why "multiplatform code" is a bad idea, You Will Have Learned(tm).
Now my question for the language mavens: most people of my generation learned procedural languages, and never really "got" the whole object-oriented thing. Most people of the 1990s and later generations seem to have started with object-oriented, but they in turn don't "get" pure the pure procedural approach (and for sure very few of either camp understand SQL). Is there any good learning language that would help a newbie put a foot in both worlds?
I just don't see how in the long run anti-virus software can be free. An AV supplier needs a team of researchers, preferably with at least some number of analysts available 24/7/365.25, and if it is going to have a realtime Windows scanner it needs programmers with a good knowledge of Windows realtime filesystem internals, and they normally don't come cheap.
Personally, I find F-Prot quite good. Home users get 5 PCs per subscription, and for corporate users their volume licensing discount is very reasonable. YMMV.
I don't think there is any question that those three sequences will produce a different nominee from the same set, and yet I have essentially zero voice in which sequence is used. (sequences are proposed by party insiders and ratified by state legislators, and unless you have $ millions to throw around neither will listen to you).
Similarly, when it comes to the Presidental election my vote is worth less than a vote from Wyoming or Montana (due to the minimum number of electoral votes per state).
So, two open, democratic voting processes freely and openly agreed upon by all parties where each individual citizen goes to the voting booth and punches one ticket. And yet some votes count more than others.
sPh
PS The funny thing is, I haven't really disagreed with anything you have said. Yet you are arguing quite strongly against my position. This too is a problem for the wikil model I think.
> You can only pay so many people money to > work on something. But people who believe > in something are a multitude, and will > generally work tireless at an effort > even if not paid...
As a person who has been a member and an officer of many volunteer-based community organizations over the last 25 years, I must respectfully disagree. There are are few organizations out there there have remained in being on a volunteer or close-to-volunteer basis for many years. Just as there are a few open source projects (notably Linux) that have maintained momentum on a volunteer or semi-volunteer basis over a reasonable period.
But the vast majority of volunteer organizations run down and vanish after a year or two. The ones that appear to be "volunteer" orgs, but which stay in being, such as the Boy Scouts, Red Cross, etc all have large paid staffs "coordinating" (=recruiting, flogging, and shaming) the volunteers, and almost all have a very large nest egg hidden away behind the scenes.
I love community organizations, don't get me wrong. But I have seldom observed what you describe.
One thing I forgot: whenever discussing mass voting one should at least review Kenneth Arrow's work on voting and the mathematics of decidibilty. It gives one pause to think about the meaning of phrases such as "your vote matters".
> And this is what the mass mind will look > like; a whole lot of people arguing and > coming to very rough consensus. It's > never going to converge on a set of > opinions that exactly match your own.
I agree, and that is a very good thing to keep in mind.
But consider the case where I hold a party in a very large ballroom and invite a reprentative sample of humanity; 100,000 people or so. After an hour those people will be scattered around the room with some probability distribution that could be considered the "normal human interaction factor".
If you then bring [insert worldwide celebrity here] in a side door, a very large percentage of the attendees are going to leave their position in the distribution and move toward the celebrity (and some will move away). The result would no longer be the "normal interaction factor". It would be correct for the circumstances, but not the true picture overall.
That is what I am saying is happening with wiki-type systems: deliberate attempts to push the results in one extreme direction.
In theory the wikipedia idea (many minds, many eyes, perhaps a voting mechanism) should work and result in articles which are fairly close to the state of human (knowledge * belief). And it did seem to be working for a while.
But in reality, people who are paid money to do something can spend far more time and effort than those who cotribute out of ego or community spirit. So it is not surprising to me that big entities are throwing a few bucks to their marketing firms to influence the web information flow. And marketing interns don't cost all that much, either: they are typically paid $15/hour and billed at $75. Peanuts compared to real marketing and advertising expenses.
I strongly suspect we are seeing the same thing on the political blogs. Except for those few that have a very large readership that takes self-policing seriouisly (e.g. DailyKos), I suspect that 20-30% of the comments on the key political blogs are being posted by paid agents. And of those comments, many flame-starters and most thread-redirectors are coming from those agents.
I think the "mass mind of humanity" idea ain't gonna work.
I taught my first "how to user corporate e-mail" class in 1991. I stressed to that class, and every one since, that every e-mail message whether deleted or not is potentially findable and discoverable in court, and that you should not put anything in e-mail you don't want to see on the front page of the New York Times, the National Enquirer, or both. Nothing has changed here.
=== If there is work to be done, then I'd like to dial up the local expert/employee and know that the problem will be fixed *quickly* and efficiently. ===
Only the fire department and the Marine Corps keep enough people on standby to handle any problem presented to them immediately (and even the Marines are a bit tied up at the moment). Every other entity queues, prioritizes, and triages. Your IT department *could* maintain enough knowledgable experts to answer your difficult questions in depth whenever you picked up the phone - I once worked with one that did. That department lasted a little less than 2 years; once senior management figured out how much it was costing they terminated it and replaced it with an outsourcer at 1/3 the cost. 1/5 the level of service, but that was not senior management's concern (and perhaps rightly so).
There are lots of reasons to support underdogs, and for better or for worse most of the candidates directly support by Kos and me in the last election cycle were not especially ideologically liberal. They were generally underdogs running against incumbents (most people running against incumbents are underdogs), and in many cases the small amount of funds provided to them helped force their opponents to spend huge amounts of money in return, diverting money from other places.
If my goal in life was to support people who were "winners" I'd be writing checks to Joe Biden, Hilllary Clinton, and Ted Kennedy. All 3 of those candidates will win their next election. All 3 of those candidates have far more money than they need to win their next election. All 3 of those candidates still have no problem getting people to line up to give them even more money for their campaigns.
> This is great. The New York Times along with most of the > press don't like blogs. So they write about bloggers that > post positive material about Walmart. Walmart which is one > of the current targets of dislike by many in the online > community. And what are these evil bloggers doing?
Indeed. The NYT and even moreso the Washington Post are horrified by the free exchange of political information and fact checking that blogs represent. The "NYT blogger ethics" kerfluffel has become a complete joke in the political blogging world. If those two traditional media outlets would just go ahead and hire some people who really understand the Internet (after what - 11 years?) they would be better off. But since that would mean reducing or even eliminating some existing fiefs, they can't do it - they will ride their traditional media model all the way down.
Unless I'm missing something this New York Times article is just another stab at holding bloggers to ethical standards and practices which don't apply anywhere else in the universe.
The public relations industry existed long before bloggers came along and they had reporters' phone numbers long before they had the email addresses of bloggers. Barely edited press releases have long been published, especially at smaller newspapers. I get press releases and information from all over the place all the time. Obviously disclosure is a nice idea if there are any financial relationships, a practice not always followed by our hallowed 4th estate, but if people want to devote their blogs to throwing up Wal Mart press releases they're free.
The main reason stories like this are even written is that contrary to popular opinnion the internet often provides a lot more transparency even when there are efforts to hide it. Astroturfing operations of various kinds through all media are nothing new, they're just usually harder to track. If Wal Mart pays 50 people to call talk radio all day and extol its virtues would anyone know?
I'm not defending all astro turfing practices or its practitioners, and there are certainly ethical issues that can be raised. But "Wal Mart PR guy reaches out to bloggers" just isn't much of a story. PR people reach out to me all the time. So what.
=== Email is not a place to store documents. Many others have said the same thing. ===
That would be fine, if there were actually any usable way to store documents. But the various document management products of the 1990-1995 timeframe died out, and of course never made the transition to the web. Yes, there are some super-duper document management "solutions" left, but no one actually uses them.
E-mail is not a good correspondence/document storage system, but it works for most ordinary human beings. So they use it for that. And taking away that functionality is counteproductive to the needs of the actual system users.
Slashdot Mirror
> the Space Shuttle's systems are run on proprietary
> software written by a company that basically just
> writes the shuttle software.
Specifically, of the five CPUs in the primary Shuttle management system, four run identical copies of the management/control software. This version was originally written by IBM's then Federal Systems Division, which was later sold (to Loral I believe). The fifth CPU runs code written to the same specifictions by the shuttle's prime contractor (then Rockwell, now Boeing). The two groups were (and I believe are) only allowed to communicate through formal written specifications and are never allowed to speak directly or to see one another's code. Whenever the software is changed, both versions must independently pass the same functional tests and then the entire cluster of five CPUs must pass the functional tests as a unit.
I haven't seen any detailed write-ups since they upgraded the Shuttle cockpit using what was essentially the Boeing 767 avionics, but I assume similar procedures still apply.
sPh
> after Gillette charges me, what,
> three million dollars for this sort
> of thing by the time it hits the market.
And you just _know_ they are already working on a version with NOT one but THREE points to follow on after that, at $11 million/cartridge.
sPh
> "FreeDOS-32 is along that direction."
> Hall says that software will include
> features like multitasking and flat memory
Throw in windowing, and you would have FreeDesqView-386.
sPh
According to one history of the 1991 Gulf War that I read, a British planning officer in London lost his portable computer (they weren't laptops then) with quite a bit of critical information on it. The London police let it be known among their contacts that it would _really_ be best if it were to be returned no-questions-asked, and it was dropped off at a police station within a day.
In a similar case in one city I was living in, 4 people in two years tried to get their spouse murdered by hanging out at a bar known to be frequented by hardened criminals and striking up a bargain with a willing thug (don't ask me why we had so many of those cases in that burg!). In all 4 cases the thug went right to the police and got fitted out for a wire. As one of them said in an interview, "I am a professional burgler but that doesn't mean I don't have standards".
So maybe the guy who stole it decided it was best not to have the entire FBI and US Army on his tail and turned it back in.
sPh
Not necessarily, as in both the open outcry and sealed-bid 1st price auctions there are various incentives for participants not to bid their maximum personal valuation (most often the "loser's curse", but there are other reasons as well).
sPh
Robert Weber of Northwestern University has written about this extensively. The solution is the 2nd-price auction. In this system, everyone puts in one bid which is the maximum they would be willing to pay. The highest bidder gets the item, but pays what the _2nd_ highest bidder bid. That way there is no incentive to try to outguess your opponents' strategies or "snipe".
The concept is hard to explain, though, so it doesn't get used much.
sPh
Dvorak writes a lot of different stuff, including some real journalism and technology analysis. But his best known work, and that which I am sure earns him his bread-and-butter, is technology gossip. Like every gossip column ever written since the first traveling minstrel appeared on the scene 30,000 years ago, Dvoark's gossip columns consist of a mixture of truth, exaggeration, spin (whether planted by the technology companies or generated by Dvorak himself), trolling, and some totally made-up stuff.
For example, Dvorak has been trying to force the monitor companies to bring new technologies to market for at least the last 20 years. That is why he hypes-Hypes-HYPES any rumour of a new display technoloy (seen that 300 dpi Texas Instruments display he reported "almost ready for production" in 1995 yet?). 40% truth, 40% exaggeration, 20% Dvorak-generated spin.
But as I said, that is how gossip columns of any kind work. Don't like it, don't read it.
sPh
> Nobody would have predicted that most people would now
> take pictures by holding the camera out in front of them
> and look at the preview screen to frame a shot.
As opposed to the Rolliflex-type camera which you held (and still hold) at waist level, looking down into the viewfinder (just don't step forward when shooting the Grand Canyon). Or the view camera, where you put the black cloth over your head and stare into a piece of ground glass?
I prefer mechanical SLRs to digital, myself, but there were more types of cameras in 1920 than there are today.
sPh
> Its all fine and great, but this fine
> DRM alternative called 'Freedom' doesn't
> pay salaries
Have you ever dealt with anyone in the music Industry (LA/Nashville/New York)? The amount of money they skim from the artists for "salaries" is staggering and sickening. One music company I encountered had a policy that every employee of the company always flew first class and stayed at the Four Seasons or the closest local equivalent. _Everyone_, all the time. And by "everyone" I mean every single person, including the junior clerk going to SAP to take an A/P data entry class. Where do you think that money comes from? Why do you think the music industry is so hot to use the awesome coercive power of the State to continue the flow of that money into their coffers?
sPh
> There are at present only two methods for sifting
> uranium atoms, or isotopes, to create the right mix.
AVLIS has been around since the 1970s, and there is also the South African cyclonic process. There are also hints in the public literature that there are other methods that were examined by the Manhatten Project and not pursued for various reasons.
sPh
In the end it all comes down to data. Languages come and go. Code comes and goes. But it is quite common to find data which has survived essentially intact in structure and content for 20, 30, even 40 years.
So don't start with any language. Start with building a deep understanding of data: what it is, how it is modeled, the techniques used to manage it. Since most data today is stored in relational databases, next proceed to understand how RDBMSs work in theory and in practice, and how to write good performant SQL. If at the end of the day you understand why "multiplatform code" is a bad idea, You Will Have Learned(tm).
Now my question for the language mavens: most people of my generation learned procedural languages, and never really "got" the whole object-oriented thing. Most people of the 1990s and later generations seem to have started with object-oriented, but they in turn don't "get" pure the pure procedural approach (and for sure very few of either camp understand SQL). Is there any good learning language that would help a newbie put a foot in both worlds?
sPh
I just don't see how in the long run anti-virus software can be free. An AV supplier needs a team of researchers, preferably with at least some number of analysts available 24/7/365.25, and if it is going to have a realtime Windows scanner it needs programmers with a good knowledge of Windows realtime filesystem internals, and they normally don't come cheap.
Personally, I find F-Prot quite good. Home users get 5 PCs per subscription, and for corporate users their volume licensing discount is very reasonable. YMMV.
sPh
Consider the following US presidential primary sequences:
Sequence 1
==============
Iowa
New Hampshire
- Arizona
- Deleware
- Missouri
- New Mexico
Michigan
Washington
Sequence 2
==============
California
Illinois
New York
Mass.
Oregon
Vermont
Sequence 3
==============
Mississippi
Alabama
Georgia
Alabama
Oklahoma
South Carolina
Utah
I don't think there is any question that those three sequences will produce a different nominee from the same set, and yet I have essentially zero voice in which sequence is used. (sequences are proposed by party insiders and ratified by state legislators, and unless you have $ millions to throw around neither will listen to you).
Similarly, when it comes to the Presidental election my vote is worth less than a vote from Wyoming or Montana (due to the minimum number of electoral votes per state).
So, two open, democratic voting processes freely and openly agreed upon by all parties where each individual citizen goes to the voting booth and punches one ticket. And yet some votes count more than others.
sPh
PS The funny thing is, I haven't really disagreed with anything you have said. Yet you are arguing quite strongly against my position. This too is a problem for the wikil model I think.
> You can only pay so many people money to
> work on something. But people who believe
> in something are a multitude, and will
> generally work tireless at an effort
> even if not paid...
As a person who has been a member and an officer of many volunteer-based community organizations over the last 25 years, I must respectfully disagree. There are are few organizations out there there have remained in being on a volunteer or close-to-volunteer basis for many years. Just as there are a few open source projects (notably Linux) that have maintained momentum on a volunteer or semi-volunteer basis over a reasonable period.
But the vast majority of volunteer organizations run down and vanish after a year or two. The ones that appear to be "volunteer" orgs, but which stay in being, such as the Boy Scouts, Red Cross, etc all have large paid staffs "coordinating" (=recruiting, flogging, and shaming) the volunteers, and almost all have a very large nest egg hidden away behind the scenes.
I love community organizations, don't get me wrong. But I have seldom observed what you describe.
sPh
One thing I forgot: whenever discussing mass voting one should at least review Kenneth Arrow's work on voting and the mathematics of decidibilty. It gives one pause to think about the meaning of phrases such as "your vote matters".
sPh
> And this is what the mass mind will look
> like; a whole lot of people arguing and
> coming to very rough consensus. It's
> never going to converge on a set of
> opinions that exactly match your own.
I agree, and that is a very good thing to keep in mind.
But consider the case where I hold a party in a very large ballroom and invite a reprentative sample of humanity; 100,000 people or so. After an hour those people will be scattered around the room with some probability distribution that could be considered the "normal human interaction factor".
If you then bring [insert worldwide celebrity here] in a side door, a very large percentage of the attendees are going to leave their position in the distribution and move toward the celebrity (and some will move away). The result would no longer be the "normal interaction factor". It would be correct for the circumstances, but not the true picture overall.
That is what I am saying is happening with wiki-type systems: deliberate attempts to push the results in one extreme direction.
sPh
In theory the wikipedia idea (many minds, many eyes, perhaps a voting mechanism) should work and result in articles which are fairly close to the state of human (knowledge * belief). And it did seem to be working for a while.
But in reality, people who are paid money to do something can spend far more time and effort than those who cotribute out of ego or community spirit. So it is not surprising to me that big entities are throwing a few bucks to their marketing firms to influence the web information flow. And marketing interns don't cost all that much, either: they are typically paid $15/hour and billed at $75. Peanuts compared to real marketing and advertising expenses.
I strongly suspect we are seeing the same thing on the political blogs. Except for those few that have a very large readership that takes self-policing seriouisly (e.g. DailyKos), I suspect that 20-30% of the comments on the key political blogs are being posted by paid agents. And of those comments, many flame-starters and most thread-redirectors are coming from those agents.
I think the "mass mind of humanity" idea ain't gonna work.
sPh
> Considering the funky keyboard they named
> after him, is this any surprise?
I remember JD writing a while ago (1995?) that he was surprised as anyone to find out the keyboard Dvorak actually was a relative of his.
sPh
Microsoft's recent spate of acquiring large-scale business software vendors seems vaguely familiar....
Oh yeah, that it what Larry Ellison is doing too. Except he started sooner.
sPh
I taught my first "how to user corporate e-mail" class in 1991. I stressed to that class, and every one since, that every e-mail message whether deleted or not is potentially findable and discoverable in court, and that you should not put anything in e-mail you don't want to see on the front page of the New York Times, the National Enquirer, or both. Nothing has changed here.
sPh
sPh
> This is great. The New York Times along with most of the
> press don't like blogs. So they write about bloggers that
> post positive material about Walmart. Walmart which is one
> of the current targets of dislike by many in the online
> community. And what are these evil bloggers doing?
Indeed. The NYT and even moreso the Washington Post are horrified by the free exchange of political information and fact checking that blogs represent. The "NYT blogger ethics" kerfluffel has become a complete joke in the political blogging world. If those two traditional media outlets would just go ahead and hire some people who really understand the Internet (after what - 11 years?) they would be better off. But since that would mean reducing or even eliminating some existing fiefs, they can't do it - they will ride their traditional media model all the way down.
sPh
E-mail is not a good correspondence/document storage system, but it works for most ordinary human beings. So they use it for that. And taking away that functionality is counteproductive to the needs of the actual system users.
sPh