No! I'm simply saying it depends on compiler design such that one cannot tell by the algorithm written in a high-level language ALONE which high-level algorithm is faster or more efficient.
The compiler design may or may NOT affect other sort algorithms, but that issue depends on the specifics of compiler design. A high level computer language does NOT specify exactly "how" to carry it out on the chip or machine level. That's ultimately up to the compiler and/or chip designer, and they have a lot of choices, and potentially future choices that nobody's discovered/invented yet.
Even with existing compilers, certain high-level constructs are more optimize-able than others at the compiler/machine level. You write the high-level code one way and you get fast, but slow another way, even though it may be the same general algorithm. One has to know something about the compiler to know that's the result. It's NOT an inherent factor of the algorithm choice used in the app language.
The app-level algorithm coding is one factor among multiple (such as compiler and chip design), and that's why a proof based on the app-level code ALONE is probably not possible (other than constraining the results to various degrees).
Just look a good lil liberal, you see everything in a vacuum
Such insults are unhelpful. I believe conservatives have patterns of defective cognitive processes, but find it doesn't change anything to tell them that, other than start flame-wars.
Well okay, can you prove it's "objectively slower" for say 5,000 items?
If that's still too open-ended, then select some typical/realistic/common limits or constraints to do it with so we are not dealing with edge cases or unlikely situations.
Hey, I patented this kind of aggressive trolling. Pay up or my attorneys will confiscate your mother's basement, including your extensive hockey-mask collection.
Let's look at this judgement thing in more detail. Email is relatively new to many gov't agencies. The laws and practices regarding it were still being worked out and evaluated. The "paper-era" laws didn't directly apply.
The "regular" office server was not intended for classified info either. The home/office thing is Chevy-vs-Chevy, NOT Chevy-vs-Lexus. Thus, the home server is NOT clearly a more risky choice. Arguments about which is technically less risky will be full of technical and operational nuances beyond H's background, and probably still subject to diff professional opinions. (Gov't infrastructure support is often shoddy, I would note, in part due to lowest bidding and lack of monitoring by gov't staff.)
And, there was no (clear) law against home servers at the time.
And, she's not an IT expert nor expected to be.
Jeb and C. Powell also used an outside server/service for gov't work emails such that it's clear "best practices" had not been established. "Do what every one else is doing" does not give the answer.
Thus, saying it's an "obviously" bad decision doesn't fly.
What mental computation steps do you imagine she should have taken in her position at that time that would clearly lead to a "no" answer?
I agree it's a poor choice, but not "obviously poor", especially to a non-IT person. It's a nuanced "poor".
I didn't say "simply remove" or anything close, that I can see. The worker she was communicating with in the message of topic is allegedly the very person whose job it is to "unclassified [it] by the appropriate process", or at least initiate the necessary process.
Not properly. It died and investigators had a hard time getting copies of old emails from it. I don't know what specifically went wrong, but I have read they had difficulty extracting its data.
Our laws as written are "crap code": vague, contradictory, and outdated because they're written and selected by politicians and lobbyists, NOT Vulcans or somebody comparable.
There is no perfect candidate. The other candidates have either done shady/suspicious things themselves, are too extreme to get elected, or have insufficient experience.
I'm not saying being "extreme" is inherently bad, only that it's pointless to back a candidate who's unlikely to get elected.
Plus, I believe one has to have a bit of a dark side to deal with shady world leaders. A lot of the job is out-jerking jerks. It's why Kirk mad a better captain than Spock.
Well, that's still a focus on hardware, which I already pointed out, but perhaps I should have elaborated on.
Further, a sufficiently smart compiler could perhaps optimize a bubble sort to be as fast as the more commonly-used algorithms such to say bubble sort is "objectively slower" than the usual algorithms is probably either wrong, contentious, or very complicated to fully prove.
One has to make assumptions about the brand or nature of the compilers and CPU(s) being used before making broad efficiency/speed claims.
As far as what's practically more efficient giving existing/common compilers and tools, I'd still be hesitant to call that "science". Is for example testing boat speed on the ocean "science" in the strict sense? I'd call that "engineering" more so than "science". Science discovered the laws of physics. Testing specific devices (boats) against existing models of physics is not really part of that.
And, most of the tricky or controversial debates I've seen about software are not related to execution speed.
For example, nobody has proven that nested blocks are "objectively better" than go-to's, even though in practice most (now) agree nested blocks are usually the better choice for most algorithms. But the answer to such a proof is probably tied to the human brain more than any external laws of physics.
That's where the interesting/puzzling/sticky parts of the field lay, NOT in machine speed (unless you are a speed fan).
...to me, she was blatantly trying to skirt those laws by setting up her own server.
How can it be "blatant" if it's "to me"? If it's a guess or interpretation of her motivations on your part, then "blatant" probably shouldn't apply.
Google give synonyms for "blatant" as: flagrant, glaring, obvious, undisguised, unconcealed, and open.
In other words, obvious to (external) observers. If it's obvious, then "to me" wouldn't apply. Therefore, it seems a contradiction. I seek clarification on what you intended to state.
As far as guessing her motives, she claims it was for the convenience of carrying fewer devices. Until mind-reading machines are invented, we'll probably never know if that is true. It's often better to focus on external realities rather than motivation guessing.
They have emails from her telling people to take classification markings off.
Hold on, Tex, she said in an interview that's simply short-hand for cleaning it up for non-classified release, such as removing the classified parts and rephrasing. She said the office worker she sent it to knew what she really (fully) meant because he had done it many times before.
If all my internal emails were interpreted by the public/press verbatim, I'd probably be on trial also.
Think about your own internal work emails being read and interpreted by bloggers, pundits, and trolls.
I don't know if her claim is true yet, but until that's determined from the investigation, it should be "innocent until proven guilty".
Further, note that if she had used the "regular" office server instead of her home server, the classified/problematic emails would probably still end up on a server NOT designed for classified materials. The risk/breach would still exist. (There was a separate transmission system for classified stuff, but it wasn't email as we know it.)
The regular office server was no more special or vetted than her home server. Pundits keep implying it is.
I'm not defending her actions, only saying many are jumping to conclusions prematurely. Using a home server is probably not illegal (although the laws are subject to interpretation*), just poor judgement, which she admitted to. Members of the other party made similar errors of judgement.
* Wealthy people are more likely to afford top lawyers who can successfully argue their side of such vague laws, and in that sense, her "privilege" may indeed just get her out of it. But that's life in an unequal society. OJ would probably lose if he were poor.
Actually, there is NOT a lot of science in computing studies outside of the guts of hardware. Science is about modelling the physical world, but computers run virtual worlds (OS's, programming environments, database models), and that is where most students will spend their time.
If you are doing "science" on a virtual world, you're really doing math: proving things about a model, NOT about the actual world. Science is about modelling and predicting the real world, while math is about modelling and predicting fake/virtual worlds.
Science could be done on the software engineering side, such as getting the best code (maintainable) and best customer satisfaction with the fewest resources, but this is closer to economics, finance, and psychology: "soft" sciences. They don't call it "economic science" or "psychology science".
Further, such "efficiency" studies are expensive in practice and there's a very limited amount of actual data on them. It's like field-science without a budget to do field science such that it turns into a big argument over pet theories (i.e. fancy sounding guesses).
The subject perhaps should be called "computer engineering" or "computer technology" and/or "software technology" or "computer studies" or the like.
But titles like "computer studies" and "software technology" sounds too fluffy. You can't charge fat tuition on fluffy-sounding subjects. Thus, there's probably a marketing angle to the current usage of "science" in "computer science".
Oh well, perfect and sufficiently-compact labels for subjects are hard to come by. Tradition may trump better titles. If I were king, I'd make them use "computer technology" or "information technology" in their course/subject titles. It's vague enough to not trigger purity fights, and vague enough to cover lots of sub-topics, which a good IT course set would do.
While I agree that programming is decent entry-level job into other IT fields, I'm not sure the emphasis of IT teaching should be on coding. Structural factoring (basic normalization, redundancy identification), set theory, logic, general architecture (clients, servers, databases, networks, security) and so forth should also be part of such courses.
Even if you never code on the job, understanding the relationships between data elements and system parts is important for most office workers, from clerks to managers and most cubicle dwellers in-between.
As far as the programming part, one of the best intro teaching projects I've seen is to have students code up their resume and cover letter, and then use a list or database to populate the resume and cover letter template with destination information (company name, company contact, specialty, address, etc.)
Have a condition whereby if the company specialty is not known (blank), then a sentence(s) about interest in the specialty is bypassed from the result. Thus, they learn looping over data, and conditionals. While not intended to actually be used for job-hunting, it gives students a feel for the practical side of programming. (This is kind of like MS's Mail-Merge.)
Plus, OOP is not the right solution for some parts of systems. The "OO everywhere" mentality of the last decade is mostly dead. We've learned, yet again, "use the right tool for the job".
The FP (functional) fans are making the same mistake lately, I believe: "FP everywhere". In part because JavaScript (ironically) has a lousy OOP model such that people stick anonymous functions all over the place. (Oh oh, I didn't mean to start a Paradigm War.)
For example, JS's "setTimout" should be part of a "timer" class instead of rely on anonymous functions. That would allow more parameters and options (possibly for future expansion):
// pseudo-code t = new(timer); t.milliseconds = 4000; t.overlap = false;// don't start a new thread if existing not done t.maxWait = 12000; function t.onTick {// over-ride (define) event
doSomethingPeriodically(); } function t.exceedWait() {
threadTookTooLongCustomHandling(
this.errMessage); }
But a bigger-picture problem is that there is more to IT than coding, and coding jobs may even shrink, as offshoring picks up. I have more to say about this in another message.
Examples? I see the worse customer service in oligopolies, bailouts or not. In oligopolies you can suck because because your other 1 to 2 competitors also suck, and you wave to each other as you pass by them with a wink-wink and give them the Golden Parachute solute.
W signed the exit agreement with their government, not O (with an exit timeline). Some argue O should have "done more" to continue the occupation, but the US would probably be blamed for "attracting" the outside attackers by being a prized target if we had done so.
Anybody who claims their Mid. E. crystal ball is accurate should be bludgeoned over the head with it.
That "wand" is called "negotiate better trade deals" instead of the lopsided deals we keep ending up with. I don't believe in him for many things, BUT negotiating better trade deals is right up his alley. Put some pressure on other countries to be consumers instead of just exporters. Whoever wins prez hopefully will find a way to leverage his negotiating skills.
Companies like Apple, BMW, Rolex, pre-Carly HP, etc. have a reputation for quality and customer service, and thus are more likely to survive and thrive in the longer term.
A consulting company doesn't have to be a "luxury brand", but at least strive to stand out from the average, and value reputation.
It may take a while to build reputation, but good things often take patience. Think of consulting quality as a sparse niche ready to be filled. There are too many consulting co's taking the quick-profit niche; it's full.
That's what IBM did wrong: set short term money/earning goals that were so aggressive that they burned long-term good-will to reach them.
Technology is such that one can often sacrifice the long-term to get short-term gains/features/improvements.
If you want to succeed in consulting for the longer term, then view yourself as a reputation company instead of a product/deliverable company. Measure your success by how happy your customers are at least as much as by current profits. If you make them happy, they'll go to you again for other projects.
You can use your good-will as a selling point in that you invite potential customers to interview current and past customers having similar projects. If your competitor(s) is a jerk, then the potential customer will find that out either when the competitor cannot provide sufficient references, or when their references tell the truth (Oracle, cough cough).
Slashdot Mirror
Sinister elements will read about this and breed giant birds, or feed beans to a thousand trained pidgins.
No! I'm simply saying it depends on compiler design such that one cannot tell by the algorithm written in a high-level language ALONE which high-level algorithm is faster or more efficient.
The compiler design may or may NOT affect other sort algorithms, but that issue depends on the specifics of compiler design. A high level computer language does NOT specify exactly "how" to carry it out on the chip or machine level. That's ultimately up to the compiler and/or chip designer, and they have a lot of choices, and potentially future choices that nobody's discovered/invented yet.
Even with existing compilers, certain high-level constructs are more optimize-able than others at the compiler/machine level. You write the high-level code one way and you get fast, but slow another way, even though it may be the same general algorithm. One has to know something about the compiler to know that's the result. It's NOT an inherent factor of the algorithm choice used in the app language.
The app-level algorithm coding is one factor among multiple (such as compiler and chip design), and that's why a proof based on the app-level code ALONE is probably not possible (other than constraining the results to various degrees).
Such insults are unhelpful. I believe conservatives have patterns of defective cognitive processes, but find it doesn't change anything to tell them that, other than start flame-wars.
That's math, not science. The main difference between the two is that science has to be somewhere tied to reality while math doesn't.
Well okay, can you prove it's "objectively slower" for say 5,000 items?
If that's still too open-ended, then select some typical/realistic/common limits or constraints to do it with so we are not dealing with edge cases or unlikely situations.
Hey, I patented this kind of aggressive trolling. Pay up or my attorneys will confiscate your mother's basement, including your extensive hockey-mask collection.
Let's look at this judgement thing in more detail. Email is relatively new to many gov't agencies. The laws and practices regarding it were still being worked out and evaluated. The "paper-era" laws didn't directly apply.
The "regular" office server was not intended for classified info either. The home/office thing is Chevy-vs-Chevy, NOT Chevy-vs-Lexus. Thus, the home server is NOT clearly a more risky choice. Arguments about which is technically less risky will be full of technical and operational nuances beyond H's background, and probably still subject to diff professional opinions. (Gov't infrastructure support is often shoddy, I would note, in part due to lowest bidding and lack of monitoring by gov't staff.)
And, there was no (clear) law against home servers at the time.
And, she's not an IT expert nor expected to be.
Jeb and C. Powell also used an outside server/service for gov't work emails such that it's clear "best practices" had not been established. "Do what every one else is doing" does not give the answer.
Thus, saying it's an "obviously" bad decision doesn't fly.
What mental computation steps do you imagine she should have taken in her position at that time that would clearly lead to a "no" answer?
I agree it's a poor choice, but not "obviously poor", especially to a non-IT person. It's a nuanced "poor".
I didn't say "simply remove" or anything close, that I can see. The worker she was communicating with in the message of topic is allegedly the very person whose job it is to "unclassified [it] by the appropriate process", or at least initiate the necessary process.
Not properly. It died and investigators had a hard time getting copies of old emails from it. I don't know what specifically went wrong, but I have read they had difficulty extracting its data.
That's been the hard part.
Our laws as written are "crap code": vague, contradictory, and outdated because they're written and selected by politicians and lobbyists, NOT Vulcans or somebody comparable.
That's a Freudian Typo. Should be "made".
There is no perfect candidate. The other candidates have either done shady/suspicious things themselves, are too extreme to get elected, or have insufficient experience.
I'm not saying being "extreme" is inherently bad, only that it's pointless to back a candidate who's unlikely to get elected.
Plus, I believe one has to have a bit of a dark side to deal with shady world leaders. A lot of the job is out-jerking jerks. It's why Kirk mad a better captain than Spock.
Well, that's still a focus on hardware, which I already pointed out, but perhaps I should have elaborated on.
Further, a sufficiently smart compiler could perhaps optimize a bubble sort to be as fast as the more commonly-used algorithms such to say bubble sort is "objectively slower" than the usual algorithms is probably either wrong, contentious, or very complicated to fully prove.
One has to make assumptions about the brand or nature of the compilers and CPU(s) being used before making broad efficiency/speed claims.
As far as what's practically more efficient giving existing/common compilers and tools, I'd still be hesitant to call that "science". Is for example testing boat speed on the ocean "science" in the strict sense? I'd call that "engineering" more so than "science". Science discovered the laws of physics. Testing specific devices (boats) against existing models of physics is not really part of that.
And, most of the tricky or controversial debates I've seen about software are not related to execution speed.
For example, nobody has proven that nested blocks are "objectively better" than go-to's, even though in practice most (now) agree nested blocks are usually the better choice for most algorithms. But the answer to such a proof is probably tied to the human brain more than any external laws of physics.
That's where the interesting/puzzling/sticky parts of the field lay, NOT in machine speed (unless you are a speed fan).
How can it be "blatant" if it's "to me"? If it's a guess or interpretation of her motivations on your part, then "blatant" probably shouldn't apply.
Google give synonyms for "blatant" as: flagrant, glaring, obvious, undisguised, unconcealed, and open.
In other words, obvious to (external) observers. If it's obvious, then "to me" wouldn't apply. Therefore, it seems a contradiction. I seek clarification on what you intended to state.
As far as guessing her motives, she claims it was for the convenience of carrying fewer devices. Until mind-reading machines are invented, we'll probably never know if that is true. It's often better to focus on external realities rather than motivation guessing.
Hold on, Tex, she said in an interview that's simply short-hand for cleaning it up for non-classified release, such as removing the classified parts and rephrasing. She said the office worker she sent it to knew what she really (fully) meant because he had done it many times before.
http://hotair.com/archives/201...
If all my internal emails were interpreted by the public/press verbatim, I'd probably be on trial also.
Think about your own internal work emails being read and interpreted by bloggers, pundits, and trolls.
I don't know if her claim is true yet, but until that's determined from the investigation, it should be "innocent until proven guilty".
Further, note that if she had used the "regular" office server instead of her home server, the classified/problematic emails would probably still end up on a server NOT designed for classified materials. The risk/breach would still exist. (There was a separate transmission system for classified stuff, but it wasn't email as we know it.)
The regular office server was no more special or vetted than her home server. Pundits keep implying it is.
I'm not defending her actions, only saying many are jumping to conclusions prematurely. Using a home server is probably not illegal (although the laws are subject to interpretation*), just poor judgement, which she admitted to. Members of the other party made similar errors of judgement.
* Wealthy people are more likely to afford top lawyers who can successfully argue their side of such vague laws, and in that sense, her "privilege" may indeed just get her out of it. But that's life in an unequal society. OJ would probably lose if he were poor.
Actually, there is NOT a lot of science in computing studies outside of the guts of hardware. Science is about modelling the physical world, but computers run virtual worlds (OS's, programming environments, database models), and that is where most students will spend their time.
If you are doing "science" on a virtual world, you're really doing math: proving things about a model, NOT about the actual world. Science is about modelling and predicting the real world, while math is about modelling and predicting fake/virtual worlds.
Science could be done on the software engineering side, such as getting the best code (maintainable) and best customer satisfaction with the fewest resources, but this is closer to economics, finance, and psychology: "soft" sciences. They don't call it "economic science" or "psychology science".
Further, such "efficiency" studies are expensive in practice and there's a very limited amount of actual data on them. It's like field-science without a budget to do field science such that it turns into a big argument over pet theories (i.e. fancy sounding guesses).
The subject perhaps should be called "computer engineering" or "computer technology" and/or "software technology" or "computer studies" or the like.
But titles like "computer studies" and "software technology" sounds too fluffy. You can't charge fat tuition on fluffy-sounding subjects. Thus, there's probably a marketing angle to the current usage of "science" in "computer science".
Oh well, perfect and sufficiently-compact labels for subjects are hard to come by. Tradition may trump better titles. If I were king, I'd make them use "computer technology" or "information technology" in their course/subject titles. It's vague enough to not trigger purity fights, and vague enough to cover lots of sub-topics, which a good IT course set would do.
Why emphasize just coding when programming jobs may shrink?
While I agree that programming is decent entry-level job into other IT fields, I'm not sure the emphasis of IT teaching should be on coding. Structural factoring (basic normalization, redundancy identification), set theory, logic, general architecture (clients, servers, databases, networks, security) and so forth should also be part of such courses.
Even if you never code on the job, understanding the relationships between data elements and system parts is important for most office workers, from clerks to managers and most cubicle dwellers in-between.
As far as the programming part, one of the best intro teaching projects I've seen is to have students code up their resume and cover letter, and then use a list or database to populate the resume and cover letter template with destination information (company name, company contact, specialty, address, etc.)
Have a condition whereby if the company specialty is not known (blank), then a sentence(s) about interest in the specialty is bypassed from the result. Thus, they learn looping over data, and conditionals. While not intended to actually be used for job-hunting, it gives students a feel for the practical side of programming. (This is kind of like MS's Mail-Merge.)
Plus, OOP is not the right solution for some parts of systems. The "OO everywhere" mentality of the last decade is mostly dead. We've learned, yet again, "use the right tool for the job".
The FP (functional) fans are making the same mistake lately, I believe: "FP everywhere". In part because JavaScript (ironically) has a lousy OOP model such that people stick anonymous functions all over the place. (Oh oh, I didn't mean to start a Paradigm War.)
For example, JS's "setTimout" should be part of a "timer" class instead of rely on anonymous functions. That would allow more parameters and options (possibly for future expansion):
But a bigger-picture problem is that there is more to IT than coding, and coding jobs may even shrink, as offshoring picks up. I have more to say about this in another message.
If another brand had the same problem, it would likely never make news. Everyone would expect it to suck.
Examples? I see the worse customer service in oligopolies, bailouts or not. In oligopolies you can suck because because your other 1 to 2 competitors also suck, and you wave to each other as you pass by them with a wink-wink and give them the Golden Parachute solute.
How about give a few options rather than make a single subject be the bottleneck. For example, give a 3-path choice:
1. Algebra (probably I and II)
2. Statistics and probability
3. Discrete math: logic/set-theory, etc.
W signed the exit agreement with their government, not O (with an exit timeline). Some argue O should have "done more" to continue the occupation, but the US would probably be blamed for "attracting" the outside attackers by being a prized target if we had done so.
Anybody who claims their Mid. E. crystal ball is accurate should be bludgeoned over the head with it.
That "wand" is called "negotiate better trade deals" instead of the lopsided deals we keep ending up with. I don't believe in him for many things, BUT negotiating better trade deals is right up his alley. Put some pressure on other countries to be consumers instead of just exporters. Whoever wins prez hopefully will find a way to leverage his negotiating skills.
Companies like Apple, BMW, Rolex, pre-Carly HP, etc. have a reputation for quality and customer service, and thus are more likely to survive and thrive in the longer term.
A consulting company doesn't have to be a "luxury brand", but at least strive to stand out from the average, and value reputation.
It may take a while to build reputation, but good things often take patience. Think of consulting quality as a sparse niche ready to be filled. There are too many consulting co's taking the quick-profit niche; it's full.
That's what IBM did wrong: set short term money/earning goals that were so aggressive that they burned long-term good-will to reach them.
Technology is such that one can often sacrifice the long-term to get short-term gains/features/improvements.
If you want to succeed in consulting for the longer term, then view yourself as a reputation company instead of a product/deliverable company. Measure your success by how happy your customers are at least as much as by current profits. If you make them happy, they'll go to you again for other projects.
You can use your good-will as a selling point in that you invite potential customers to interview current and past customers having similar projects. If your competitor(s) is a jerk, then the potential customer will find that out either when the competitor cannot provide sufficient references, or when their references tell the truth (Oracle, cough cough).
Analog voting can be tampered with also. A system of checks and balances is needed regardless of the technology used.