although it's not open source and not free for purchase without advertisements, Opera 7 does everything the administrator says IE does but Firefox doesn't.
The nature of the Bible is an interesting subject. There are many ways to approach it.
For a middle-of-the-road view from modern archaeology, see The Bible and Interpretation, as well as the IMO excellent book The Bible Unearthed: Archaeology's New Vision of Ancient Israel and the Origin of Its Sacred Texts, Israel Finkelstein, Neil Asher Silberman.
The thing about some archaeologists is, are they looking to see what the evidence out there says? Or are they looking to find evidence for a Divine hand-me-down? The traps in the latter approach are many, including confusing absence of evidence for evidence of absence.
Taking this comment "... without an evil empire dictating to them" as a starting point, I'd love to hear people's views on the relationship between creativity/innovation and corporate capitalism. I quite deliberately distinguish between corporate, shareholder-oriented capitalism and other forms of capitalism and do it while understanding, I believe, the powerful ability corporate capitalism has to raise, well, capital.
Why is it that originals and innovators so often leave the entity they help found? It can't be easy. Is it that they are all snobs who "really don't want to work an honest buck for a living"? Is it that the corporation undervalues or devalues them as it succeeds? Is it that corporations, at least big ones, are incompatible with innovation or, at least, seeing it implemented. Or is it that innovation and creativity are themselves enumerated in some kind of different value unit which is interconverted with capital only with difficulty?
I have heard a related argument saying that corporations never actually pay taxes because those costs are passed on to customers. It sounds reasonable, but it is actually wrong -- and it is slippery -- so it deserves some kind of response.
The problem with that argument is it is choosing to make the corporation/non-corporation economic boundary be fluid while at the same time implying that the economic boundary between a person and that person's universe is not fluid. If corporations never pay taxes using this argument, then by extending it we can conclude noone ever pays taxes. I am assessed taxes which, actually, my employer pays, not me. If I cannot afford pay these taxes I find an employer who can pay the taxes. If that economic boundary is assumed rigid for a person -- that is, the person is an entity capable of actually paying something out of some repository of value that is theirs -- then to be consistent it must be assumed to be rigid for a corporation.
If corporations "cannot" in some fundamental way pay taxes then they cannot pay fines either. Consequently none of them should care when they get fined.
This is an interesting line of questioning for a couple of reasons.
First, when Selective Availability was active, people did work up means of getting military-level accuracy and better by using things like differential GPS. (See also a project by NASA. There are many other references from the geophysical sciences community.) It was safe, at least then, because the time needed to get a good position fix was on the order of hours. That rate is fine for geophysics.
Second, part of the political and strategic thinking about GPS was to put assets in orbit so they might serve as a target instead of ground-based systems. That is, the mindset at the time was very much one of fighting a nuclear war. The problem of that was in part seen as one of releasing a cataclysm if the nuclear option was exercised. So, it was thought, if juicy enough targets were put in space, an adversary could use a nuclear weapon to destroy those to press the point of their seriousness home without committing to a direct attack and its devastating retaliation. I imagine that was also true of communications satellites.
The USA actually did conduct some atmospheric tests to measure effects of high altitude nuclear blasts. While the results are classified and these were conducted before satellites were widely used, the possibility of nuclear attack is taken seriously enough that designs for military satellites, including GPS, undergo testing for nuclear hardening.
Indeed, regarding the ISS lookover, it did not occur to anyone, as you implied. It's interesting to consider what that implies about the mindset of folks involved. A checkout procedure at each takeoff is standard-standard for any aircraft. But NASA's mentality saw launch from Earth as being the "start", rather than the journey from space as being one.
Practically speaking, however, Columbia was in an orbit too significantly different from the ISS to be able to maneuver there for such a lookover. Of course, it was planned that way, and perhaps such a constraint should be imposed. If it is, it will be a huge blow to the space program because there are many orbital inclinations which would then be out of reach, and many of them are very useful, like the polar-circling ones.
Of course, having more space stations aloft would help, but we've been there with that....
The goal of quantum computing has always been claimed to be greatly increased computation speeds by applying massive, exponential parallelism. There is no limitation claimed for the kind of computation that can be sped up in this manner, merely that it is sped up.
If the problem is a matter of searching through a massive computational space for one item, however that item is characterized, and computation stops--or might as well stop--as soon as it is found, then I agree, quantum computing can result in a great speedup. But most computing is not of that kind.
Most computing is above rearranging memory into some kind of preferred state. The computing we do today which we characterize as requiring a lot of MIPS or GIPS is the kind that involves large amounts of memory to be massaged or filled, e.g., image manipulation, weather models.
The problem with trying to apply "general purpose parallelism" to the average computation as expressed today is that it rapidly devolves into a case where only a few of the concurrent units, however small and fast, end up computing the execution sequence leading to the answer. In that case the net speed of the computation is whatever the speed is of the individual units, for concurrency has been forgone during this process.
Consider what happens when modern processors pre-fetch contents of memory on both sides of conditional instructions ahead of the computation deciding the conditional being completed. There is some gain in speed because whichever path on the conditional the plan is finally decided, the processor does not have to await the fetch of results from memory. However, if there is yet another conditional instruction on each branch, while those can be pre-fetched as well, only 1/4 of them will be used along the actual execution path. That means, yes, the cost of pre-fetch doesn't need to be incurred, but 3/4 of the effort to do the pre-fetch is thrown away to gain that. As more conditional instructions are encountered to save the cost of the pre-fetch, only 2**(-N) where N is the number of conditionals will actually apply to the execution path. This quantity gets small very quickly as N increases and 1 - 2**(-N) approaches one. This means that in any computation involving conditional choices, speedup rapidly decreases so the speed of the single unit is all that matters.
Of course this only applies to problems and programs which have been designed for sequential computers that are attempted on such concurrent machines. If the program is specially structured for the machine and to take advantage of the characteristics of the problem, as is done for today's processors which feature large amounts of concurrency, large speedup is possible. But the arrangement and architecture of the program doesn't survive a change in problem. It has to be designed all over again.
This is quite different than simply giving your C program to a compiler and having a reasonable expectation that it will compile, execute, and run. Because that expectation doesn't survive the move to any massively parallel environment, I am skeptical about its prospects, be it realized through quantum computing or anything else.
--jtg
P.S. Oh, regarding credentials, consider: If even the Devil states that 2 x 2 = 4, I am going to believe him. (P.P.Waldenström)
I just introduced my two teen kids today to the original "Spacewar" as re-created in a Java applet by MIT's Media Lab. They are video game and PC game experts. They were amazed at how much function and fun there was in the thing. 1962. PDP-1. What, 8K of memory?
The inefficient use of memory and space is, IMO, not a technical problem. It's that some folks want to start raking in cash on schedules which are incompatible with the basic technology of their products.
They can think they have their day. IMO, the age of the Patient Undercutters is coming. Remember when Hitachi cloned the IBM MVS OS without having access to its internal documentation? There are huge financial incentives for bright, less expensive Asian companies to give Microsoft a real run for their money.
Think of Microsoft as the Ford, Chrysler, and GM of the 1970s.
To play the devilry advocate for a moment, what makes anyone think the programming of quantum monstrosities will be any easier than it is today?
Think of it, the model of computing we use today is one silly little processor chugging its way through memory, fetching data and instructions, doing one thing at a time. The "frontier" thusfar has been learning to do those things faster and faster, albeit one thing at a time. In limited cases and situations, we can throw concurrency at these problems, but they are still apportioned at the problem level, not at the compiler or hardware level.
So, with quantum, you can explore a billion alternative pathways at a time. Great. How do you ever make a decision? And, once made, how do you keep from making (optimally) half of those pathways irrelevant at each step? Suppose you explore a thousand trillion pathways at a time? How many decisions = powers of two are needed to reduce that to essentially a single pathway? Something like 55.
In his weekly spiel, Bob Cringely commented on the Compaq computers that were being removed from government buildings, wondering in comment where they came from since apparently Iraq had none in 1991 and they were supposed to be under an embargo.
I don't know if those facts are right, but Cringely usually checks things out.
About ten years ago, I was doing some statistical investigations of the incidence of homicides in urban centers. I came across some discrepancies between the reports from the Bureau of Justice Statistics ("BJS"), the database referred to in the parent of this post, and the National Center for Health Statistics ("NCHS"). Homicides are reported both by administrative arms of the courts and law enforcement, and by county and state pathologists, although using different reporting channels. In theory, they should agree. There are really no different standards used in one versus another, since pathologists determine cause of death.
I was startled to see significant discrepancies. Indeed, when I looked into the matter further, I found there had been entire research papers written on the subject. (Don't have time to find those now, but those interested can e-mail me.) Further study indicated I needed to know something about how BJS obtained their numbers, so I called BJS.
I spoke to a representative there for a long time. Now, this charcterization is, as I said, ten years old, and perhaps it has improved. But the rep said contribution of numbers and classification of crime incidents were voluntary programs from state and county authorities. In contrast, y'see, reporting from pathologists is mandatory and they are audited for accuracy. No such auditing is done for BJS or states or counties. When I asked why, the rep said that they needed to cooperate with local law authorities on a wide range of programs involving enforcement and otherwise, and they did not want to alienate them.
So, at least as of ten years ago, if a local sherriff or state wanted to make their official track record look better, they could "administratively misclassify" a set of crimes and there is no mechanism for finding them out. It is only detectable in the case of homicide because both BJS and NCHS have the same events recorded in their databases.
And it took SciFi's "Children of Dune" to make you aware of this?
Sorry, I wouldn't care since life is too short to worry about things other than in what one passionately believes, but even OUR industry is subject to these kinds of "faddish" forces, whether these are moved by corporate advertising, such as.Net, or common cause, such as enthusiasm for technologies tied into Java.
Okay. Since this is a group which prides itself upon technological erudition and logic, please justify your assessment that "the test results are su[r]prisingly accurate". How do you know? Using "gut feel"? If your company uses these tests all the time, how can you know if you've not hired someone who failed the test?
How can you know that? How do they know that? Are you saying something as complicated as a personality or workstyle or someone's compatibility with a team can be characterized completely by a scalar value?
This is very true, chaeron. I think it is also a side effect of another phenomenon, something I realized painfully in 2002 while out of work.
The fact is that software development and information technology is an industry undergoing severe deflation. It used to be such deflation affected only hardware, a byproduct of it becoming a commodity industry. But with interest rates so low and growth not so good, most businesses are really cutting costs and anything which doesn't translate directly and can be seen to translate directly into improved sales or profits is a likely candidate.
In my opinion, which was discussed extensively in a forum at realrates.com, demand for programmers has been constant since 1990 and, in fact, slightly decreasing. There has been a lot of demand for other computing specialists, including the database development crowd to which I now belong. The reason for this is that businesses have moved to buying shrinkwrap solutions rather than keeping a stable of expensive programmers. For most companies who cannot sell a product to a zillion customers and thereby fund an expensive development group, doing software is just too expensive and consumes resources not related directly to sales or support. Heck, I saw an article in one of those business mags for the suits that recommended rethinking corporate philanthropy so that it is more oriented towards sales and so the business gets something back for it. Buying shrinkwrap is the cheapest form of outsourcing there is.
If tools for special effects in movies are any indication, large sectors of what were hot domains for programmers are becoming standardized and usable directly by functionals. Naturally, the company that makes these tools needs programmers, but once a market has a dozen competitors offering products that everyone else buys that caps the limit on the number of coders. And, of course, if the sector sees a shakeout, as most eventually do, that trims the demand more.
I think this phenomenon is behind the rise of one-size-fits-all enterprise packages like PeopleSoft, SAP, and the Oracle Financials entrants. Suits seem to have decided that software is so much a drain they win bigger if they distort their policies and procedures to fit the software rather than using custom. I think this phenomenon is also behind the push for software that lets one DBA administer hundreds of databases.
Not all software application areas seem to be sucuumbing to this. Games development still looks like it can afford building independent engines and the like. But it'll probably happen there, too. I mean, realistically, what's the point of repeating a creation of a good renderer? Most games today seem to be built using tools that let artist-designers develop scenes and having extensive parameter files. How much longer before a half dozen companies standardize that and make up for any efficiencies by simply exploiting faster hardware?
Frankly, I think the reason why scripts are becoming so popular and widely used is that they can do a job quickly. That means it's cheaper because it draws fewer programming hours. And, with reason, I think the suits don't see a cost to impossible-to-maintain software, at least not one that's comparable to the price of paying the programmer hours. If there are components of their process that must be maintained carefully, they buy that. Anything else by definition doesn't require long term maintenance. It does require speed, flexibility, and low cost.
I, like many people here, love to code and it's what I envisioned doing indefinitely. But the demand is decreasing. So, I get my kicks doing ETL, database, and business intelligence stuff, using gawk to do it whenever I can. And with gawk I can knock out applications really fast. Clients often don't even know or care as long as the job gets done cheaply and quickly.
Hey, we have always known, struggled, and argued with software being too expensive to develop and maintain. Much of the training professional programmers get concerns how to build better software. We've always felt a tension between quality and speed in the sense of rapid development. The suits have decided that we took too long to solve the problem and found their own way. Oh well, the computer giveth and the computer taketh away.
Re:Conservative or Liberal, Salon Was/Is Great
on
Salon Asks for Help
·
· Score: 1
Yes, I agree. Killing off publications for political reasons is like refusing to answer questions of reporters in a news conference because a body doesn't like the political slant of the media they represent.
Re:Economizing, deflation, or whatever, it's too b
on
Salon Asks for Help
·
· Score: 1
Is there such a thing as a pro-American socialist, in your opinion?
Economizing, deflation, or whatever, it's too bad
on
Salon Asks for Help
·
· Score: 2, Insightful
I've been a member of Salon for quite a while and did what I could to support them. I would be sorry to see it go. I think it's funny that people think Salon is "left". I mean, if it is left, what is the Boston Phoenix or the Village Voice?
I also would be sorry, and more pertinently to Slashdot, because I think their design for semi-automated publishing was kind of neat, and it is one of the last examples of a house doing their own development work I know of. That is a dwindling group.
While I cannot address the questions of rent for their offices -- which if true, I agree seem excessive -- I think "the end of the dot-com bubble" means more than the crashing of way-out business models, excessive spending, and such. I mean, when MoTown was starting up, they were excessive in parties, liquor, etc
To me, these companies are failing as much because of deflation in the information technology industry as anything else. That deflation is caused:
partly because of low interest rates in the economy
partly because the hardware component of the industry is now commodity-based and people have an expectation that prices should drop, for those and telecommunications costs
partly because programming labor is cheaper and more widely available
partly because non-IT businesses are fiercely cutting costs, including moving to shrinkwrap solutions for their IT needs, even if they are not a good match
partly because the Internet marketplace has long had expectations that things there should be free or available at modest charges.
The last effect is a subtle, I think. Since good news coverage and similar entertainment is now available on the Internet and cheaply, any premium or brick-and-mortar company has to deal with not so much with e-business competition but with the expectation that new can be had for much less. Why subscribe to the New York Times paper when most of what's good about it is available online for zip?
I think whatever happens to Salon is part of a trend, because what we earn for doing information technology is diminishing and will continue to diminish.
Yeah, I pretty much agree. The only serious limitation I see continuing with the current Shuttle computers is the reported difficulty on finding replacement parts for them.
I used to work for Federal Systems in Owego, NY, near here, where the Shuttle's computers are built. Indeed, I used to work for the department that developed the 'support software' (HAL/S compilers, assemblers, link-editors, and simulators) used to program them. I was still working there during the 1986 Challenger disaster. (I left in 1995 after Gerstner sold us to Loral as a "reward" for being so successful.) The computers did very well during Challenger, although there's no way they could have held together during Columbia's breakup.
The computers and core memories were shipped to Owego in vats of de-ionized water. The memories were powered up and their information slowly copied onto external media. When decoded what they revealed, apart from standard guidance information and status which was provided to the Challenger investigation board, was that the computers sensed power had dropped off and with their remaining power proceeded to do an orderly shutdown of the computer systems, writing system state into their core memories in anticipation of a restart.
It was a horrible tragedy for all of us, as was Columbia, but it's gratifying some things worked as designed.
Slashdot Mirror
although it's not open source and not free for purchase without advertisements, Opera 7 does everything the administrator says IE does but Firefox doesn't.
better, IE implements CSS2 more completely.
and it's available for Linux as well as Win32.
WHAT?! A suit as "most powerful man in technology journalism!? **NEVER**!!!!
It's gotta be Bob Cringely.
The nature of the Bible is an interesting subject. There are many ways to approach it.
For a middle-of-the-road view from modern archaeology, see The Bible and Interpretation, as well as the IMO excellent book The Bible Unearthed: Archaeology's New Vision of Ancient Israel and the Origin of Its Sacred Texts, Israel Finkelstein, Neil Asher Silberman.
It is available at Powell's among Other Places.
The thing about some archaeologists is, are they looking to see what the evidence out there says? Or are they looking to find evidence for a Divine hand-me-down? The traps in the latter approach are many, including confusing absence of evidence for evidence of absence.
Taking this comment "... without an evil empire dictating to them" as a starting point, I'd love to hear people's views on the relationship between creativity/innovation and corporate capitalism. I quite deliberately distinguish between corporate, shareholder-oriented capitalism and other forms of capitalism and do it while understanding, I believe, the powerful ability corporate capitalism has to raise, well, capital.
Why is it that originals and innovators so often leave the entity they help found? It can't be easy. Is it that they are all snobs who "really don't want to work an honest buck for a living"? Is it that the corporation undervalues or devalues them as it succeeds? Is it that corporations, at least big ones, are incompatible with innovation or, at least, seeing it implemented. Or is it that innovation and creativity are themselves enumerated in some kind of different value unit which is interconverted with capital only with difficulty?
I have heard a related argument saying that corporations never actually pay taxes because those costs are passed on to customers. It sounds reasonable, but it is actually wrong -- and it is slippery -- so it deserves some kind of response.
The problem with that argument is it is choosing to make the corporation/non-corporation economic boundary be fluid while at the same time implying that the economic boundary between a person and that person's universe is not fluid. If corporations never pay taxes using this argument, then by extending it we can conclude noone ever pays taxes. I am assessed taxes which, actually, my employer pays, not me. If I cannot afford pay these taxes I find an employer who can pay the taxes. If that economic boundary is assumed rigid for a person -- that is, the person is an entity capable of actually paying something out of some repository of value that is theirs -- then to be consistent it must be assumed to be rigid for a corporation.
If corporations "cannot" in some fundamental way pay taxes then they cannot pay fines either. Consequently none of them should care when they get fined.
This is an interesting line of questioning for a couple of reasons.
First, when Selective Availability was active, people did work up means of getting military-level accuracy and better by using things like differential GPS. (See also a project by NASA. There are many other references from the geophysical sciences community.) It was safe, at least then, because the time needed to get a good position fix was on the order of hours. That rate is fine for geophysics.
Second, part of the political and strategic thinking about GPS was to put assets in orbit so they might serve as a target instead of ground-based systems. That is, the mindset at the time was very much one of fighting a nuclear war. The problem of that was in part seen as one of releasing a cataclysm if the nuclear option was exercised. So, it was thought, if juicy enough targets were put in space, an adversary could use a nuclear weapon to destroy those to press the point of their seriousness home without committing to a direct attack and its devastating retaliation. I imagine that was also true of communications satellites.
The USA actually did conduct some atmospheric tests to measure effects of high altitude nuclear blasts. While the results are classified and these were conducted before satellites were widely used, the possibility of nuclear attack is taken seriously enough that designs for military satellites, including GPS, undergo testing for nuclear hardening.
Indeed, regarding the ISS lookover, it did not occur to anyone, as you implied. It's interesting to consider what that implies about the mindset of folks involved. A checkout procedure at each takeoff is standard-standard for any aircraft. But NASA's mentality saw launch from Earth as being the "start", rather than the journey from space as being one.
Practically speaking, however, Columbia was in an orbit too significantly different from the ISS to be able to maneuver there for such a lookover. Of course, it was planned that way, and perhaps such a constraint should be imposed. If it is, it will be a huge blow to the space program because there are many orbital inclinations which would then be out of reach, and many of them are very useful, like the polar-circling ones.
Of course, having more space stations aloft would help, but we've been there with that ....
The goal of quantum computing has always been claimed to be greatly increased computation speeds by applying massive, exponential parallelism. There is no limitation claimed for the kind of computation that can be sped up in this manner, merely that it is sped up.
If the problem is a matter of searching through a massive computational space for one item, however that item is characterized, and computation stops--or might as well stop--as soon as it is found, then I agree, quantum computing can result in a great speedup. But most computing is not of that kind.
Most computing is above rearranging memory into some kind of preferred state. The computing we do today which we characterize as requiring a lot of MIPS or GIPS is the kind that involves large amounts of memory to be massaged or filled, e.g., image manipulation, weather models.
The problem with trying to apply "general purpose parallelism" to the average computation as expressed today is that it rapidly devolves into a case where only a few of the concurrent units, however small and fast, end up computing the execution sequence leading to the answer. In that case the net speed of the computation is whatever the speed is of the individual units, for concurrency has been forgone during this process.
Consider what happens when modern processors pre-fetch contents of memory on both sides of conditional instructions ahead of the computation deciding the conditional being completed. There is some gain in speed because whichever path on the conditional the plan is finally decided, the processor does not have to await the fetch of results from memory. However, if there is yet another conditional instruction on each branch, while those can be pre-fetched as well, only 1/4 of them will be used along the actual execution path. That means, yes, the cost of pre-fetch doesn't need to be incurred, but 3/4 of the effort to do the pre-fetch is thrown away to gain that. As more conditional instructions are encountered to save the cost of the pre-fetch, only 2**(-N) where N is the number of conditionals will actually apply to the execution path. This quantity gets small very quickly as N increases and 1 - 2**(-N) approaches one. This means that in any computation involving conditional choices, speedup rapidly decreases so the speed of the single unit is all that matters.
Of course this only applies to problems and programs which have been designed for sequential computers that are attempted on such concurrent machines. If the program is specially structured for the machine and to take advantage of the characteristics of the problem, as is done for today's processors which feature large amounts of concurrency, large speedup is possible. But the arrangement and architecture of the program doesn't survive a change in problem. It has to be designed all over again.
This is quite different than simply giving your C program to a compiler and having a reasonable expectation that it will compile, execute, and run. Because that expectation doesn't survive the move to any massively parallel environment, I am skeptical about its prospects, be it realized through quantum computing or anything else.
--jtg
P.S. Oh, regarding credentials, consider: If even the Devil states that 2 x 2 = 4, I am going to believe him. (P.P.Waldenström)
I agree with this in a big way.
I just introduced my two teen kids today to the original "Spacewar" as re-created in a Java applet by MIT's Media Lab. They are video game and PC game experts. They were amazed at how much function and fun there was in the thing. 1962. PDP-1. What, 8K of memory?
The inefficient use of memory and space is, IMO, not a technical problem. It's that some folks want to start raking in cash on schedules which are incompatible with the basic technology of their products.
They can think they have their day. IMO, the age of the Patient Undercutters is coming. Remember when Hitachi cloned the IBM MVS OS without having access to its internal documentation? There are huge financial incentives for bright, less expensive Asian companies to give Microsoft a real run for their money.
Think of Microsoft as the Ford, Chrysler, and GM of the 1970s.
To play the devilry advocate for a moment, what makes anyone think the programming of quantum monstrosities will be any easier than it is today?
Think of it, the model of computing we use today is one silly little processor chugging its way through memory, fetching data and instructions, doing one thing at a time. The "frontier" thusfar has been learning to do those things faster and faster, albeit one thing at a time. In limited cases and situations, we can throw concurrency at these problems, but they are still apportioned at the problem level, not at the compiler or hardware level.
So, with quantum, you can explore a billion alternative pathways at a time. Great. How do you ever make a decision? And, once made, how do you keep from making (optimally) half of those pathways irrelevant at each step? Suppose you explore a thousand trillion pathways at a time? How many decisions = powers of two are needed to reduce that to essentially a single pathway? Something like 55.
Yeah, right. Big breakthrough.
In his weekly spiel, Bob Cringely commented on the Compaq computers that were being removed from government buildings, wondering in comment where they came from since apparently Iraq had none in 1991 and they were supposed to be under an embargo.
I don't know if those facts are right, but Cringely usually checks things out.
About ten years ago, I was doing some statistical investigations of the incidence of homicides in urban centers. I came across some discrepancies between the reports from the Bureau of Justice Statistics ("BJS"), the database referred to in the parent of this post, and the National Center for Health Statistics ("NCHS"). Homicides are reported both by administrative arms of the courts and law enforcement, and by county and state pathologists, although using different reporting channels. In theory, they should agree. There are really no different standards used in one versus another, since pathologists determine cause of death.
I was startled to see significant discrepancies. Indeed, when I looked into the matter further, I found there had been entire research papers written on the subject. (Don't have time to find those now, but those interested can e-mail me.) Further study indicated I needed to know something about how BJS obtained their numbers, so I called BJS.
I spoke to a representative there for a long time. Now, this charcterization is, as I said, ten years old, and perhaps it has improved. But the rep said contribution of numbers and classification of crime incidents were voluntary programs from state and county authorities. In contrast, y'see, reporting from pathologists is mandatory and they are audited for accuracy. No such auditing is done for BJS or states or counties. When I asked why, the rep said that they needed to cooperate with local law authorities on a wide range of programs involving enforcement and otherwise, and they did not want to alienate them.
So, at least as of ten years ago, if a local sherriff or state wanted to make their official track record look better, they could "administratively misclassify" a set of crimes and there is no mechanism for finding them out. It is only detectable in the case of homicide because both BJS and NCHS have the same events recorded in their databases.
*sigh* for sure.(:-(}
And it took SciFi's "Children of Dune" to make you aware of this?
.Net, or common cause, such as enthusiasm for technologies tied into Java.
Sorry, I wouldn't care since life is too short to worry about things other than in what one passionately believes, but even OUR industry is subject to these kinds of "faddish" forces, whether these are moved by corporate advertising, such as
A sample of Cornell's usage pattern is available and, yes, it's primarily, KaZaA.
Ah, sorry for being dense.
Okay. Since this is a group which prides itself upon technological erudition and logic, please justify your assessment that "the test results are su[r]prisingly accurate". How do you know? Using "gut feel"? If your company uses these tests all the time, how can you know if you've not hired someone who failed the test?
How can you know that? How do they know that? Are you saying something as complicated as a personality or workstyle or someone's compatibility with a team can be characterized completely by a scalar value?
Yeah, business decisions are often made for the strangest reasons, like what pot of money can be used to pay for it.
Dr Dobbs has a section devoted to scripting languages as well as one devoted to lightweight languages.
Yes, and we know the reason why VB was introduced was to give Jolt a market to sell their caffeine-free products to.
Sorry, couldn't resist.
This is very true, chaeron. I think it is also a side effect of another phenomenon, something I realized painfully in 2002 while out of work.
The fact is that software development and information technology is an industry undergoing severe deflation. It used to be such deflation affected only hardware, a byproduct of it becoming a commodity industry. But with interest rates so low and growth not so good, most businesses are really cutting costs and anything which doesn't translate directly and can be seen to translate directly into improved sales or profits is a likely candidate.
In my opinion, which was discussed extensively in a forum at realrates.com, demand for programmers has been constant since 1990 and, in fact, slightly decreasing. There has been a lot of demand for other computing specialists, including the database development crowd to which I now belong. The reason for this is that businesses have moved to buying shrinkwrap solutions rather than keeping a stable of expensive programmers. For most companies who cannot sell a product to a zillion customers and thereby fund an expensive development group, doing software is just too expensive and consumes resources not related directly to sales or support. Heck, I saw an article in one of those business mags for the suits that recommended rethinking corporate philanthropy so that it is more oriented towards sales and so the business gets something back for it. Buying shrinkwrap is the cheapest form of outsourcing there is.
If tools for special effects in movies are any indication, large sectors of what were hot domains for programmers are becoming standardized and usable directly by functionals. Naturally, the company that makes these tools needs programmers, but once a market has a dozen competitors offering products that everyone else buys that caps the limit on the number of coders. And, of course, if the sector sees a shakeout, as most eventually do, that trims the demand more.
I think this phenomenon is behind the rise of one-size-fits-all enterprise packages like PeopleSoft, SAP, and the Oracle Financials entrants. Suits seem to have decided that software is so much a drain they win bigger if they distort their policies and procedures to fit the software rather than using custom. I think this phenomenon is also behind the push for software that lets one DBA administer hundreds of databases.
Not all software application areas seem to be sucuumbing to this. Games development still looks like it can afford building independent engines and the like. But it'll probably happen there, too. I mean, realistically, what's the point of repeating a creation of a good renderer? Most games today seem to be built using tools that let artist-designers develop scenes and having extensive parameter files. How much longer before a half dozen companies standardize that and make up for any efficiencies by simply exploiting faster hardware?
Frankly, I think the reason why scripts are becoming so popular and widely used is that they can do a job quickly. That means it's cheaper because it draws fewer programming hours. And, with reason, I think the suits don't see a cost to impossible-to-maintain software, at least not one that's comparable to the price of paying the programmer hours. If there are components of their process that must be maintained carefully, they buy that. Anything else by definition doesn't require long term maintenance. It does require speed, flexibility, and low cost.
I, like many people here, love to code and it's what I envisioned doing indefinitely. But the demand is decreasing. So, I get my kicks doing ETL, database, and business intelligence stuff, using gawk to do it whenever I can. And with gawk I can knock out applications really fast. Clients often don't even know or care as long as the job gets done cheaply and quickly.
Hey, we have always known, struggled, and argued with software being too expensive to develop and maintain. Much of the training professional programmers get concerns how to build better software. We've always felt a tension between quality and speed in the sense of rapid development. The suits have decided that we took too long to solve the problem and found their own way. Oh well, the computer giveth and the computer taketh away.
Yes, I agree. Killing off publications for political reasons is like refusing to answer questions of reporters in a news conference because a body doesn't like the political slant of the media they represent.
Is there such a thing as a pro-American socialist, in your opinion?
I've been a member of Salon for quite a while and did what I could to support them. I would be sorry to see it go. I think it's funny that people think Salon is "left". I mean, if it is left, what is the Boston Phoenix or the Village Voice?
I also would be sorry, and more pertinently to Slashdot, because I think their design for semi-automated publishing was kind of neat, and it is one of the last examples of a house doing their own development work I know of. That is a dwindling group.
While I cannot address the questions of rent for their offices -- which if true, I agree seem excessive -- I think "the end of the dot-com bubble" means more than the crashing of way-out business models, excessive spending, and such. I mean, when MoTown was starting up, they were excessive in parties, liquor, etc
To me, these companies are failing as much because of deflation in the information technology industry as anything else. That deflation is caused:
The last effect is a subtle, I think. Since good news coverage and similar entertainment is now available on the Internet and cheaply, any premium or brick-and-mortar company has to deal with not so much with e-business competition but with the expectation that new can be had for much less. Why subscribe to the New York Times paper when most of what's good about it is available online for zip?
I think whatever happens to Salon is part of a trend, because what we earn for doing information technology is diminishing and will continue to diminish.
Yeah, I pretty much agree. The only serious limitation I see continuing with the current Shuttle computers is the reported difficulty on finding replacement parts for them.
I used to work for Federal Systems in Owego, NY, near here, where the Shuttle's computers are built. Indeed, I used to work for the department that developed the 'support software' (HAL/S compilers, assemblers, link-editors, and simulators) used to program them. I was still working there during the 1986 Challenger disaster. (I left in 1995 after Gerstner sold us to Loral as a "reward" for being so successful.) The computers did very well during Challenger, although there's no way they could have held together during Columbia's breakup.
The computers and core memories were shipped to Owego in vats of de-ionized water. The memories were powered up and their information slowly copied onto external media. When decoded what they revealed, apart from standard guidance information and status which was provided to the Challenger investigation board, was that the computers sensed power had dropped off and with their remaining power proceeded to do an orderly shutdown of the computer systems, writing system state into their core memories in anticipation of a restart.
It was a horrible tragedy for all of us, as was Columbia, but it's gratifying some things worked as designed.