Has anyone ever seen any evidence about the difference between Quadro and GeForce?
I’ve read through all of the white papers describing the “differences” between a Quadro and a GeForce card [1]. Approaching it as a skeptic, this paper seems written by the Quadro marketing team. Most of the claims made about the quadro are also true about the geforce - for example, just the first one says that the quadro supports anti-aliasing in hardware, but then so does geforce.
Consider the Quadro P4000, retailing for $799 vs. the 1080ti at $749
Architecture: both are pascal chips
Clock rates: titan is 1536MHz, pascal is 1202MHz
Memory width: titan is 352-bit, pascal is 256-bit
Memory Bw: titan is 484MBps, pascal is 243MBps
CUDA cores: titan 3584 cores, pascal is 1792 cores
Both support G-sync, directX 12, Vulkan API, OpenGL 4.5,.
Both support up to 4 displays, both support 4K and 5K displays.
I've even looked at rendered images side by side and I cannot see any difference... the only thing I can tell is that its a different profit center for nVidia.
[1] www.nvidia.com/object/quadro_geforce.html
From the article:... bring outsiders with no experience onto teams to keep creativity and innovation on track. When experts have to slow down and go back to basics to bring an outsider up to speed, "it forces them to look at their world differently and, as a result, they come up with new solutions to old problems."
Wow. Isn't this what Universities have been doing for hundreds of years? Its almost like academics understand this concept or something.
Thank you Ingo for replying! I am very interested in your work on scheduling both from an academic and pragmatic point of view. In an earlier post, someone commented on the anecdotal differences in the Solaris scheduler to the Linux scheduler w.r.t. heavily loaded systems.
I was wondering if you have any comment on this, specifically with regards to the increasing use of virtualization, specifically VMWare products. Will/can CFS improve vm performance? Will CFS improve the throughput processing capabilities of Linux, or is it chiefly designed to improve interactive processing w/out degrading throughput?
Finally, I am curious, has there been any consideration to allowing scheduling hints to be passed in from the programmer, ala cache hint directives in some of intel's compilers.
You are exactly right. The Commonwealth of Pennsylvania placed high standards on our teachers, including requiring them to maintain a certain degree of virtue and decorum in their private lives. In fact, any teacher found committing even a misdemeanor even in a non-public forum can be stripped of their credentials at anytime.
Further, the entire text of the article explained that there were some concerns raised by her practicum advisers. The implication being that she wasn't really up to par after all.
So, while Millersville University took a harsh approach to this one candidate, I actually applaud their efforts. As little as 20 years ago in Pennsylvania and across the country, Universities maintained the right of "in loco parentis" (sp?). Now, Universities are petrified of ticking a student off and incurring the inevitable litigation.
Maybe the pendulum will begin to swing back toward some serious ass-kicking on students. Professors are NOT there to baby sit. Professors do NOT have to "give A's" to everyone. Students are NOT entitled to a degree. Students who work full-time so they can afford a nice car, cheap booze, and loads of pot are going to have serious difficulty maintaining any academic standing. So, go MU, I think you sent a strong message to students. I just hope it doesn't go too far.
P.S. Teachers aren't the only individuals expected to maintain social decorum - so are commonwealth employees.
First, the current Web 2.0={Facebook, Blogs, Tagging, Mashups,... } is NOT the end of the Semantic Web. What tiny bit of SW technology that has leaked into the infrastructure of these technologies is a tiny fraction of the capabilities of this technology.
In fact, I am amazed and appalled at the general reaction of the slashdot community. The general consensus is that this couldn't possibly work because it will require "corporations" to come together (ironic given the etnymology of the word corporation). The same could be said of the early Linux community. That Linus Torvalds is never going to amount to anything because his new OS will require people all over the place to agree on standards (the kernel) and to cooperate in a huge development process that has never been tried on this scale before. I'm quite certain that Linux is just a passing fancy.
Seriously though, today's computing paradigm sucks. I spend most of my time working on stupid technical problems instead of actually working on the hard research issues I am trying to focus on. Case in point, I spent the weekend trying to get JDBC and/or PERL to insert a long string into a CLOB of XMLType with Oracle. Why? Because the agent I developed was performing tests and producing result sets in XML, and it would be handy to store in Oracle. In the end, I spent more time solving the "how do I do this stupid task" than working on the actual research issues of the project.
In other cases, we (as a community) still spend our time worrying about little bits of a the technical minutiae. One of the things that draws my interested toward the Semantic Web is that it won't work on today's computing paradigm. It isn't going to be successful so long as we are worried about connecting this SQL query to that CGI script, and discussing how we screen scrape that HTML page and extract that information from a CDF file. These are the things we have been doing since the first magentic tapes were sent in the mail to be read by some other system.
I think it is time that we start moving on and looking forward to what the next generation of computing CAN do. Yes, there are significant technical challenges ahead. Yes, there will be false starts and probably a high infant mortality rate as we move forward. Already there have been numerous Semantic Web languages (e.g. DAML+OIL) that are being replaced by others (e.g. OWL), and already the research community is pushing a new version of OWL to include in its definition things which we cannot do today.
I guess if anything, I am encouraged by the stalwarts of today's technology calling for the impending implosion of the Semantic Web. It means that the research community has pushed so far to the edge of what We (as a species) are currently capable of that only visionaries (not including myself in this set) can see the whole picture. This is good. This represents the first honest See-change in the future of computing that I've seen in years.
So, if anything, keep calling for its demise. Keep predicting its death. One day, when the research moves it from science to technology, you will be excited about how cool this new technology is, and Slashdot will be filled with discussions of how Microsoft's implementation of the Semantic Web is terrible, and how the OSS version of a Semantic Web database is 5% faster than Oracle's. And there will be some comfort in knowing that all is right in the world after all.
And how many people sat around and said, "when someone can deliver information over the internet without requiring special readers and a PhD in CS to use, I'll be interested." And of that set of people, how many were successful in the explosive growth of the internet vs. being steamrolled or back-shelved. This is the attitude that I see so prevalent on Slashdot regarding this topics. Why not take some time to at least consider the alternatives.
Love the imagary your post conjures; where is that bleach to pour into my mind. I must agree with you though, blogging is not inherently part of the SW; it is part of BBS systems. Good lord, what is old is new again. Blog posts are just mesasge forums where you are your own sysop.
As far as the semantic web does go, it is not simply for supporting some new fancy SQL. In fact, the purpose is to support representation, consumption, and reasoning about knowledge. Knowledge != Data. In your "interesting" example, we would use the semantic web to represent exactly what properties make something an image (knowledge representation), allow our agents to collect information about things that directly satisfy those properies (consume knowledge), and to infer that others things which satisfy those properties are also images (reasoning).
You may have a rule that says an image is a file on the web with an MIME type of image/jpg
I may say that an image is a stream of bytes which begins with "GIF87" (or whatever the header is for a GIF file). If our reasoner can infer our two descriptions describe the same things, then it can infer that the set of all images is the union of our two rules. Thus, our agent has created NEW knowledge that we never put into the system.
Have you ever read the original presentation of work by Codd on relational databases? How about the RFC standards on TCP/IP? How about the original presentation and arguments on the inclusion of Interrupts in a processor? Boy, those were so easy to understand and obvious that they were even published at all.
The process of science is to push the state of the art; which by definition is new and novel. This is the job of the computer science researcher. It is left to others to examine the research and reformulate in terms that mere mortals can understand.
If you understand the concepts behind the OSI layers, Lambda expressions, or symmetric multi-processing, thank a computer science educator who abstracted and distilled the hell of the science and research and packaged in such a way that you can understand it and maybe even use it.
To claim that failure is imminent because the current presentation of the Semantic Web is too complex is nonsense.
I've read most of the responses; which can be summed as:
Don't bother, its not useful, take $X instead.
I was a CS major; but it sucked, so now I'm a $Y instead.
We can't hire enough CS majors
We've hired CS majors, but they sucked, so now we only hire $Z
CS is too theoretical
Who needs a degree anyway, I taught myself how to program in $P, and now my life is perfect.
CS != programming, so unless you want to be an academic, don't bother.
While these points are all valid facets of a rather rough diamond, there are some other overlooked sides as well.
When your CS profs were making you study the Traveling Salesman Problem and fifty different ways of sorting in O(n lg n) time, the ultimate goal wasn't to teach you how to code TSP, MergeSort, or Dijkstra's Semaphores. The goal of a good CS program, whether from Stanford or your local community college is to teach the theory of computation. Specifically, how to recognize the common patterns that exist in problems, and how to solve those problems using a variety of well known techniques.
In short, the ultimate goal of a good CS program isn't to create programmers who can write code in some specific language, or use some specific operating system, or use a particular vendors router. The goal is to create individuals who can reach into their experience to identify existing solutions to similar problems, use sound engineering principles to implement those solutions, and to communicate those solutions to others in precise scientific languages. With more training (advanced degrees) or experience one hopes that instead of relying on existing patterns, we can develop new solutions to problems using sound scientific methods for discovery; and analyze the soundness of those results. Computer Science doesn't just allow us (here I mean humanity) to make statements about a particular machine or language; it allows us to make statements about the Universe. That is what makes us a science, a discipline separate from physics, math, or philosophy.
Consider the situation where you are tasked to develop a solution to a problem. Without Computer Science, you can only say "I can't solve this problem." If this is your response to me, as an employer, my response is "Then, you're fired, I'll find someone else who can." With Computer Science, your response can be "I can't solve this problem, and I can prove that no one else can either." That is a very different place to be.
I see many of our freshman walk into our CS 1 (and for the 66% who remain into our CS 2) classes, and are stunned to find out that CS isn't like their Cisco or Windows certification they had in high school. It is not just writing programs. They find themselves lost when they discover that we aren't going to teach them VB and Windows server and Word. So, after a short battle, a trip to their advisor, and a visit to the registrar's office, they typically go off to Business Info Systems (MIS). Not that I want a smack down with Info. Systems people, but their goal is more applied, and technology specific. For some of our students, it is a better fit, and we are all much better for it.
To conclude my wanton rant, I will add that we, as a discipline, do a terrible job at portraying to society what it is that we do an who we are. Even my family has no clue what it is that I do as a computer scientist. They think I'm just the ultimate tech support guy. The fact that more than 50 years after the discipline was created, we are still having arguments about what it is, is telling in and of itself. Is CS worth it? Yes. Without it, computing would never happen. Imagine the performance of computer systems without pipelining and branch prediction; concealed concurrency and multi-processing; O(n lg n) sorts and the TSP; and all of the other innovations that make today's technology not only possible but usable. Is CS for everyone? No.
Ontologies for the Semantic Web are based on description logics (OWL-DL) or first-order logics (Owl-Full). We define classes and their relationships (T-Box definitions), and we define instance assertions (A-Box definitions).
For example, we could define the Apple domain as :
We can assign the domain of makesComputer to be a ComputerMfg, and the range to be a Computer (the inverse would be flipped).
Class rdf:ID="Computer"
Class rdf:ID="Garbage"
Class rdf:ID="ComputerMfg"
ObjectProperty rdf:ID="computerMadeBy, domain rdf:resource="#Computer", range rdf:resource="#ComputerMfg", inverseOf rdf:ID="makesComputer",
ObjectProperty rdf:about="#makesComputer", domain rdf:resource="#ComputerMfg", range rdf:resource="#Computer", inverseOf rdf:resource="#computerMadeBy"
Nothing about Apple yet. So,
We can assert that "APPLE" is a ComputerMfg (not Garbage), and that it is related to the symbol PoweerBook by the makesComputer / computerMadeBy relationship.
So, using the Semantic Web (as it stands) requires crisp description logics, and admits (almost) no ambiguity. For those who want to pick at me, yes, OWA and UNA make things a little strange.
Given that natural language is fraught with uncertainty, this is the root of the automatic ontology generation problem (and the beginning of my research).
The article does a pretty bad job at explaining the situation. The idea behind the Semantic Web is simply to provide a framework for information to be marked up for machines rather than human eyes. The idea is that using an agreed upon frame of reference for the symbols contained in the page (an ontology), agents are able to make use of the information contained there. Further, an agent can collection data from several different ontologies and (hopefully) perform basic reasoning tasks over that data, and (even better) complete some advanced tasks for the agent's user.
The article would have us believe that this is going to expose everyone to massive amounts of privacy invasion. This is not necessarily the case. It is already the case that there are privacy mechanisms to protect information in the SW (e.g. require agents to authenticate to a site to retrieve restricted information). Beyond simple mechanisms, there is a lot of research being conducted on the idea of trust in the semantic web - e.g. how does my agent know to trust a slashdot article as absolute truth and a wikipedia article as outright fabrication (or vice versa).
As for making the content of the internet widely available, some researchers feel this will never happen. As another commenter noted that it is essential that there is agreement in the definition of concepts (ontologies) to enable the SW to work (if my agent believes the symbol "apple" refers to the concept Computer, and your agent believes it refers to "garbage", we may have some interesting but less than useful results). I am researching ontology generation using information extraction / NLP techniques, and it is certainly a difficult problem, and one that isn't likely to have a trivial problem (in some respects, this is goes back to the origins of AI in the 1950's, and we're still hacking at it today).
For some good references on the Semantic Web (beyond Wikipedia), check out some of these links
I am a CS professor (but I don't play one on TV). Our department held this argument several years ago. I believe that pedagogically, the answer depends on the course you are teaching, specifically what are its course objectives.
For example, we teach a CS1/CS2 sequence of courses using an Objects-First curriciulum. In our CS1 course, students are exposed to a wide variety of "theory" (e.g. OO design and modelling, inheritance) and they are also exposed to language / syntax issues (e.g. loops, arrays, conditional execution).
I wrote a lab manual for our students to use, and the very first lab gives them information on the role of the compiler, including translating this stuff called "source code" into this stuff called "byte code" which is executed by this thing called the "virtual machine." In this lab, students write a stupid program (hello world, I think), and they compile it manually and run it. Then they do the same thing with an example that they are told contains errors, and they have to find the error and fix it (the class name doesn't match the class file name - Java *hates* this). This is the only lab where I have them using the command line. It teaches them that they *can* do it if they want; and that there isn't any magic behind the compiler.
The second lab uses the Eclipse IDE, and we redo the first lab (including trying to recreate the error). Then, the lab guides them through writing classes using Eclipse's code generation wizards (which, IMHO are *WAY* cool).
As the semester progresses, several new tips are introduced (e.g. after encapsulation is taught, and they've written 30 or 40 getX and setY methods, I teach them the generate getters / setters wizard).
The final plug for using an IDE isn't in the editor - its in the debugger. There are few things that appeal to the visual learners (that studies show a statistically significant portion of CS majors are) like watching code execute through the debugger. Students who are having a tough time building mental models of how this stuff works can actually see every instruction execute through the debugger - and it really helps.
I've been a CS professor for 6 years (at a "teaching" school), and have taught a wide swath of undergraduate and graduate, lower and upper division courses, and in all cases I've found that introducing students to apropriate tools helps them keep their attention on the important subject material (in the case of intro programming courses that is learning how to use the language and solve problems). Isn't this what its all about?
Slashdot Mirror
Has anyone ever seen any evidence about the difference between Quadro and GeForce? I’ve read through all of the white papers describing the “differences” between a Quadro and a GeForce card [1]. Approaching it as a skeptic, this paper seems written by the Quadro marketing team. Most of the claims made about the quadro are also true about the geforce - for example, just the first one says that the quadro supports anti-aliasing in hardware, but then so does geforce. Consider the Quadro P4000, retailing for $799 vs. the 1080ti at $749 Architecture: both are pascal chips Clock rates: titan is 1536MHz, pascal is 1202MHz Memory width: titan is 352-bit, pascal is 256-bit Memory Bw: titan is 484MBps, pascal is 243MBps CUDA cores: titan 3584 cores, pascal is 1792 cores Both support G-sync, directX 12, Vulkan API, OpenGL 4.5 ,.
Both support up to 4 displays, both support 4K and 5K displays.
I've even looked at rendered images side by side and I cannot see any difference ... the only thing I can tell is that its a different profit center for nVidia.
[1] www.nvidia.com/object/quadro_geforce.html
From the article: ... bring outsiders with no experience onto teams to keep creativity and innovation on track. When experts have to slow down and go back to basics to bring an outsider up to speed, "it forces them to look at their world differently and, as a result, they come up with new solutions to old problems."
Wow. Isn't this what Universities have been doing for hundreds of years? Its almost like academics understand this concept or something.
Thank you Ingo for replying! I am very interested in your work on scheduling both from an academic and pragmatic point of view. In an earlier post, someone commented on the anecdotal differences in the Solaris scheduler to the Linux scheduler w.r.t. heavily loaded systems. I was wondering if you have any comment on this, specifically with regards to the increasing use of virtualization, specifically VMWare products. Will/can CFS improve vm performance? Will CFS improve the throughput processing capabilities of Linux, or is it chiefly designed to improve interactive processing w/out degrading throughput? Finally, I am curious, has there been any consideration to allowing scheduling hints to be passed in from the programmer, ala cache hint directives in some of intel's compilers.
You are exactly right. The Commonwealth of Pennsylvania placed high standards on our teachers, including requiring them to maintain a certain degree of virtue and decorum in their private lives. In fact, any teacher found committing even a misdemeanor even in a non-public forum can be stripped of their credentials at anytime. Further, the entire text of the article explained that there were some concerns raised by her practicum advisers. The implication being that she wasn't really up to par after all. So, while Millersville University took a harsh approach to this one candidate, I actually applaud their efforts. As little as 20 years ago in Pennsylvania and across the country, Universities maintained the right of "in loco parentis" (sp?). Now, Universities are petrified of ticking a student off and incurring the inevitable litigation. Maybe the pendulum will begin to swing back toward some serious ass-kicking on students. Professors are NOT there to baby sit. Professors do NOT have to "give A's" to everyone. Students are NOT entitled to a degree. Students who work full-time so they can afford a nice car, cheap booze, and loads of pot are going to have serious difficulty maintaining any academic standing. So, go MU, I think you sent a strong message to students. I just hope it doesn't go too far. P.S. Teachers aren't the only individuals expected to maintain social decorum - so are commonwealth employees.
First, the current Web 2.0={Facebook, Blogs, Tagging, Mashups, ... } is NOT the end of the Semantic Web. What tiny bit of SW technology that has leaked into the infrastructure of these technologies is a tiny fraction of the capabilities of this technology.
In fact, I am amazed and appalled at the general reaction of the slashdot community. The general consensus is that this couldn't possibly work because it will require "corporations" to come together (ironic given the etnymology of the word corporation). The same could be said of the early Linux community. That Linus Torvalds is never going to amount to anything because his new OS will require people all over the place to agree on standards (the kernel) and to cooperate in a huge development process that has never been tried on this scale before. I'm quite certain that Linux is just a passing fancy.
Seriously though, today's computing paradigm sucks. I spend most of my time working on stupid technical problems instead of actually working on the hard research issues I am trying to focus on. Case in point, I spent the weekend trying to get JDBC and/or PERL to insert a long string into a CLOB of XMLType with Oracle. Why? Because the agent I developed was performing tests and producing result sets in XML, and it would be handy to store in Oracle. In the end, I spent more time solving the "how do I do this stupid task" than working on the actual research issues of the project.
In other cases, we (as a community) still spend our time worrying about little bits of a the technical minutiae. One of the things that draws my interested toward the Semantic Web is that it won't work on today's computing paradigm. It isn't going to be successful so long as we are worried about connecting this SQL query to that CGI script, and discussing how we screen scrape that HTML page and extract that information from a CDF file. These are the things we have been doing since the first magentic tapes were sent in the mail to be read by some other system.
I think it is time that we start moving on and looking forward to what the next generation of computing CAN do. Yes, there are significant technical challenges ahead. Yes, there will be false starts and probably a high infant mortality rate as we move forward. Already there have been numerous Semantic Web languages (e.g. DAML+OIL) that are being replaced by others (e.g. OWL), and already the research community is pushing a new version of OWL to include in its definition things which we cannot do today.
I guess if anything, I am encouraged by the stalwarts of today's technology calling for the impending implosion of the Semantic Web. It means that the research community has pushed so far to the edge of what We (as a species) are currently capable of that only visionaries (not including myself in this set) can see the whole picture. This is good. This represents the first honest See-change in the future of computing that I've seen in years.
So, if anything, keep calling for its demise. Keep predicting its death. One day, when the research moves it from science to technology, you will be excited about how cool this new technology is, and Slashdot will be filled with discussions of how Microsoft's implementation of the Semantic Web is terrible, and how the OSS version of a Semantic Web database is 5% faster than Oracle's. And there will be some comfort in knowing that all is right in the world after all.
And how many people sat around and said, "when someone can deliver information over the internet without requiring special readers and a PhD in CS to use, I'll be interested." And of that set of people, how many were successful in the explosive growth of the internet vs. being steamrolled or back-shelved. This is the attitude that I see so prevalent on Slashdot regarding this topics. Why not take some time to at least consider the alternatives.
Love the imagary your post conjures; where is that bleach to pour into my mind. I must agree with you though, blogging is not inherently part of the SW; it is part of BBS systems. Good lord, what is old is new again. Blog posts are just mesasge forums where you are your own sysop. As far as the semantic web does go, it is not simply for supporting some new fancy SQL. In fact, the purpose is to support representation, consumption, and reasoning about knowledge. Knowledge != Data. In your "interesting" example, we would use the semantic web to represent exactly what properties make something an image (knowledge representation), allow our agents to collect information about things that directly satisfy those properies (consume knowledge), and to infer that others things which satisfy those properties are also images (reasoning). You may have a rule that says an image is a file on the web with an MIME type of image/jpg I may say that an image is a stream of bytes which begins with "GIF87" (or whatever the header is for a GIF file). If our reasoner can infer our two descriptions describe the same things, then it can infer that the set of all images is the union of our two rules. Thus, our agent has created NEW knowledge that we never put into the system.
Have you ever read the original presentation of work by Codd on relational databases? How about the RFC standards on TCP/IP? How about the original presentation and arguments on the inclusion of Interrupts in a processor? Boy, those were so easy to understand and obvious that they were even published at all. The process of science is to push the state of the art; which by definition is new and novel. This is the job of the computer science researcher. It is left to others to examine the research and reformulate in terms that mere mortals can understand. If you understand the concepts behind the OSI layers, Lambda expressions, or symmetric multi-processing, thank a computer science educator who abstracted and distilled the hell of the science and research and packaged in such a way that you can understand it and maybe even use it. To claim that failure is imminent because the current presentation of the Semantic Web is too complex is nonsense.
- Don't bother, its not useful, take $X instead.
- I was a CS major; but it sucked, so now I'm a $Y instead.
- We can't hire enough CS majors
- We've hired CS majors, but they sucked, so now we only hire $Z
- CS is too theoretical
- Who needs a degree anyway, I taught myself how to program in $P, and now my life is perfect.
- CS != programming, so unless you want to be an academic, don't bother.
While these points are all valid facets of a rather rough diamond, there are some other overlooked sides as well. When your CS profs were making you study the Traveling Salesman Problem and fifty different ways of sorting in O(n lg n) time, the ultimate goal wasn't to teach you how to code TSP, MergeSort, or Dijkstra's Semaphores. The goal of a good CS program, whether from Stanford or your local community college is to teach the theory of computation. Specifically, how to recognize the common patterns that exist in problems, and how to solve those problems using a variety of well known techniques. In short, the ultimate goal of a good CS program isn't to create programmers who can write code in some specific language, or use some specific operating system, or use a particular vendors router. The goal is to create individuals who can reach into their experience to identify existing solutions to similar problems, use sound engineering principles to implement those solutions, and to communicate those solutions to others in precise scientific languages. With more training (advanced degrees) or experience one hopes that instead of relying on existing patterns, we can develop new solutions to problems using sound scientific methods for discovery; and analyze the soundness of those results. Computer Science doesn't just allow us (here I mean humanity) to make statements about a particular machine or language; it allows us to make statements about the Universe. That is what makes us a science, a discipline separate from physics, math, or philosophy. Consider the situation where you are tasked to develop a solution to a problem. Without Computer Science, you can only say "I can't solve this problem." If this is your response to me, as an employer, my response is "Then, you're fired, I'll find someone else who can." With Computer Science, your response can be "I can't solve this problem, and I can prove that no one else can either." That is a very different place to be. I see many of our freshman walk into our CS 1 (and for the 66% who remain into our CS 2) classes, and are stunned to find out that CS isn't like their Cisco or Windows certification they had in high school. It is not just writing programs. They find themselves lost when they discover that we aren't going to teach them VB and Windows server and Word. So, after a short battle, a trip to their advisor, and a visit to the registrar's office, they typically go off to Business Info Systems (MIS). Not that I want a smack down with Info. Systems people, but their goal is more applied, and technology specific. For some of our students, it is a better fit, and we are all much better for it. To conclude my wanton rant, I will add that we, as a discipline, do a terrible job at portraying to society what it is that we do an who we are. Even my family has no clue what it is that I do as a computer scientist. They think I'm just the ultimate tech support guy. The fact that more than 50 years after the discipline was created, we are still having arguments about what it is, is telling in and of itself. Is CS worth it? Yes. Without it, computing would never happen. Imagine the performance of computer systems without pipelining and branch prediction; concealed concurrency and multi-processing; O(n lg n) sorts and the TSP; and all of the other innovations that make today's technology not only possible but usable. Is CS for everyone? No.Ontologies for the Semantic Web are based on description logics (OWL-DL) or first-order logics (Owl-Full). We define classes and their relationships (T-Box definitions), and we define instance assertions (A-Box definitions).
For example, we could define the Apple domain as :
Classes: Computer, Garbage, ComputerMfg
Roles: makesComputer computerMadeBy
We can assign the domain of makesComputer to be a ComputerMfg, and the range to be a Computer (the inverse would be flipped).
Class rdf:ID="Computer"
Class rdf:ID="Garbage"
Class rdf:ID="ComputerMfg"
ObjectProperty rdf:ID="computerMadeBy, domain rdf:resource="#Computer", range rdf:resource="#ComputerMfg", inverseOf rdf:ID="makesComputer",
ObjectProperty rdf:about="#makesComputer", domain rdf:resource="#ComputerMfg", range rdf:resource="#Computer", inverseOf rdf:resource="#computerMadeBy"
Nothing about Apple yet. So,
We can assert that "APPLE" is a ComputerMfg (not Garbage), and that it is related to the symbol PoweerBook by the makesComputer / computerMadeBy relationship.
Computer rdf:ID="PowerBook", computerMadeBy ComputerMfg rdf:ID="APPLE"
makesComputer rdf:resource="#PowerBook"
So, using the Semantic Web (as it stands) requires crisp description logics, and admits (almost) no ambiguity. For those who want to pick at me, yes, OWA and UNA make things a little strange.
Given that natural language is fraught with uncertainty, this is the root of the automatic ontology generation problem (and the beginning of my research).
The article would have us believe that this is going to expose everyone to massive amounts of privacy invasion. This is not necessarily the case. It is already the case that there are privacy mechanisms to protect information in the SW (e.g. require agents to authenticate to a site to retrieve restricted information). Beyond simple mechanisms, there is a lot of research being conducted on the idea of trust in the semantic web - e.g. how does my agent know to trust a slashdot article as absolute truth and a wikipedia article as outright fabrication (or vice versa).
As for making the content of the internet widely available, some researchers feel this will never happen. As another commenter noted that it is essential that there is agreement in the definition of concepts (ontologies) to enable the SW to work (if my agent believes the symbol "apple" refers to the concept Computer, and your agent believes it refers to "garbage", we may have some interesting but less than useful results). I am researching ontology generation using information extraction / NLP techniques, and it is certainly a difficult problem, and one that isn't likely to have a trivial problem (in some respects, this is goes back to the origins of AI in the 1950's, and we're still hacking at it today).
For some good references on the Semantic Web (beyond Wikipedia), check out some of these links
For example, we teach a CS1/CS2 sequence of courses using an Objects-First curriciulum. In our CS1 course, students are exposed to a wide variety of "theory" (e.g. OO design and modelling, inheritance) and they are also exposed to language / syntax issues (e.g. loops, arrays, conditional execution).
I wrote a lab manual for our students to use, and the very first lab gives them information on the role of the compiler, including translating this stuff called "source code" into this stuff called "byte code" which is executed by this thing called the "virtual machine." In this lab, students write a stupid program (hello world, I think), and they compile it manually and run it. Then they do the same thing with an example that they are told contains errors, and they have to find the error and fix it (the class name doesn't match the class file name - Java *hates* this). This is the only lab where I have them using the command line. It teaches them that they *can* do it if they want; and that there isn't any magic behind the compiler.
The second lab uses the Eclipse IDE, and we redo the first lab (including trying to recreate the error). Then, the lab guides them through writing classes using Eclipse's code generation wizards (which, IMHO are *WAY* cool).
As the semester progresses, several new tips are introduced (e.g. after encapsulation is taught, and they've written 30 or 40 getX and setY methods, I teach them the generate getters / setters wizard).
The final plug for using an IDE isn't in the editor - its in the debugger. There are few things that appeal to the visual learners (that studies show a statistically significant portion of CS majors are) like watching code execute through the debugger. Students who are having a tough time building mental models of how this stuff works can actually see every instruction execute through the debugger - and it really helps.
I've been a CS professor for 6 years (at a "teaching" school), and have taught a wide swath of undergraduate and graduate, lower and upper division courses, and in all cases I've found that introducing students to apropriate tools helps them keep their attention on the important subject material (in the case of intro programming courses that is learning how to use the language and solve problems). Isn't this what its all about?