Not likely. I would imagine that each item has a unique id, not just each web page, since their needs to be some way to identify what the target of a link is. Just because a link ends in pdf, or jpg, or gif, does not mean that it is of that type. The crawlers undoubtedly record the content-type of fetched resources.
So I would guess that they already use more than 32 bits per item with everything in a single item ID space, or they use 32-bits plus some code indicating the ID-space, or more perhaps a variable length code depending on the item type, e.g. like UTF8. In any case, they should have exceeded 32-bits long ago.
More to the point, why are humans required at all in the manufacturing process. I would expect the entire manufacturing and testing process, from sand to plastic-encased chip, to be automated enough that people in bunny suits should not be needed. Maybe they are needed to replace the robots and fill up the supplies, but other than that, what do they do?
I use google most of the time. But I find that the innovations are coming from elsewhere. Frankly, Google is innovating with baby steps. Spellcheck is a nice feature, but it's not revolutionary or unique. Google labs is bunch of undergraduate level bullshit stuff. It's not the stuff of supposed army-of-PhDs breakthroughs.
I like Google because it is fast, real fast and uncluttered, but the results are not better that Teoma or AllTheWeb. The link analysis that was unique to Google, 6 years ago, was the real quantum leap forward. But now everybody else has caught up. It appears to me that the differentiation is fast, bug-free quality of service and a clean UI.
Short of another breakthrough from Google, I think Microsoft could still clobber Google. Google has got no stickiness.
The real reason, as quoted from a CNet
article on Aug 13, 2002:
The new desktops appear to be a slick interpretation of Microsoft's new licensing terms and a way to navigate customer demand for PCs without an OS installed. The Microsoft licensing terms, which were put in place on Aug. 1, specify that PC makers must ship PCs with an operating system. The new policy exists to prevent piracy and to better track OS shipments.
For what the timstamping service actually does, it costs way, way, way too much. It cost 80 cents per email, 10 cents in bulk. This is super trivial; it should cost 1 or 2 cents, and yahoo mail or hotmail could do it for free. I don't see what Authentidate offers, other than a countersignature with a private key, timeserver, and a hash.
And of course, there is a free PGP timestamping service, but unfortunately, that does not have the backing of the USPS.
Anyone know of something similar that is cheap?
Re:How long before people start gaming the system?
on
IBM vs. Content Chaos
·
· Score: 1
I second this sentiment, that gaming of any system is likely, not merely possible.
This is because humans can be "gamed" in the real world. That is, one can fabricate a "buzz" about things, not simply by overt measures like commercials, but plants in social situations. Sony or some other consumer electronics companies planted people in Times Square and other highly visible situations to pretend to use some cool new gadget. Then people see it and tell their friends and then eventually, they hope, there will be an underground sort of "buzz" generated about the cool product.
If people can't tell that they are being manipulated on the street, I see no hope for machines to be able to figure this out either. Furthermore, when applied to the web, it is not difficult to imagine writing programs that could post good or bad reviews of any product or company to blogs, newsgroups, or whatever.
The posted reviews would have to be varied and still capture the same meaning. That does not seem that hard. Even a simple randomized (CFG) grammar could provide enough variability to make that same statements but with different phrases. It doesn't have to fool a human, just a machine that is looking for grammatical sentences with key adjectives. Perhaps something like...
START:= PRODUCT "is" COOL_PHRASE. "I installed one" TIME_PHRASE HOME_PLACE.
COOL_PHRASE:= "wicked" | "phat" | "fly".
TIME_PHRASE:= "during Christmas" | "after Kwansa" | "last weekend".
HOME_PLACE:= "at my crib" | "in my apartment" | "outside my mansion" | "in my old jetstream" | "in my mom's basement"
Basically, link farms combined with natural language generation techniques could make this type of semantic inference about as hard as deciphering spam.
I imagine that a first pass of a WebFountain approach would try to cluster pages or comments based on the inferred sentiment. Then a human would have to read all the individual comments to filter out the semantic spam. Then the remainder would have to be reranked.
I was astonished with how low the budget was. I think some billionaries could easily be cajoled into funding some probes with their pocket change, to create perhaps, the Virgin MegaProbe, Trump Traveller, Oprah Orbiter, Google Globetrotter (whenever they go public)?
Anyway, they could be guilted into doing some interesting science while engorging their egos.
I, for one, would welcome some new search engine overlord. I also agree that google's results are not so great anymore, so I have become adept at adding all sorts of quotes, pluses, OR, AND, '*'s, site:, etc. But all of these tricks are ways of overcoming the absence of personalization in google, and others. I want personalized categories and personalized filters and personalized anything else.
I already use yahoo mail, and I love MyYahoo.
I would gladly wait a whole second or two, or even wait for email, to get a useful page of results that is biased to my categories. I don't give a flying fuck about what is most popular to the general web audience. Fundamentally, there is no such thing as "most relevant" because that it usually (maybe exclusively) context-dependent and person-dependent. On a Monday, my searches could be linux or programming related, so I want a customizable filter or checkbox that biases everything for that. But on Friday, I may want to find localized pizza joints, AND I want to find the most popular movies among everyone, not just linux geekdom.
All that should be one category or filter or personalized button away, because my preferences don't change that often, even though my search needs change every time. And I think that if everybody was using some personalization, it would become more difficult to build pages and link farms that would always rise to the top. Spamming would be harder because the optimizers would have to guess what personlization biases the vast majority of users would choose.
And even simple filtering and reranking should be easily possible in about a second. e.g. restrictions to selected categories in dmoz.com, or sites within a certain distance or weight of some human-edited directory, or that contain any synonym for some set of terms, and many others that are easy to think of. None of this stuff is rocket science or new or computationally expensive. But this sort of remembered customization would go a long way to saving me hours of search time every week.
I'm disappointed that it hasn't happened, even since altavista days.
My guess is that google will eventually have to provide logins and accounts to allow personalization, because cookies may not be enough. Cookies don't stay in sync with my laptop or when I go to the library or my friends computer. Yahoo, MSN, and AOL will easily be able to do this stuff. I'd like Google to remember me also. In other words, Google may have to start looking even more like a general purpose portal.
I don't think it was openness alone, but that they gave Microsoft exclusive rights to DOS for the clones. Then all the clones ran with the same MS-DOS, which was nearly identical to IBM's PC-DOS variant. So Microsoft got the software profits on the clones. It was a very, very smart move. I don't think that many big companies would have anticipated the consequences, given that Microsoft was a tiny company selling Basic.
I think the question is not of global warming, but global warming due to human actions. I think there is agreement that global warming has definately been occurring, but the disagreement is whether 20th century CO2 production is responsible for it.
I would be thrilled to not have to care about CO2. Then a great reason to avoid fossil fuels is because everyone involved in oil production is a bunch of fucking assholes, Exxon, OPEC, Texans, terrorists or their apologists, take your pick...
Heck, if we have fuel cell based cars, then you could also drink the water from the exhaust of your car. It would be more pure than Evian. Screw the French too.
I'd really like to see Linux on this sucker because I really want to hook up a digital camera to it. It's got USB, so it could suck in the pictures from a modern camera, but this requires a proper driver; e.g. developed on a Linux desktop and then ported to the handheld.
This way you could use the PSP to offload pictures and do more serious editing. Editing on the camera is a pain in the ass. What it needs is much, much larger storage. It sounds like it doesn't have a harddrive, since it's only got 1.8G storage. If it did, like the iPod variety, then you could listen to a
truckload of music and store hours of video and hundreds of thousands of shots from
a camera; all without lugging around a laptop.
Unfortunately I haven't seen any palmtop that could did this as cheaply, or laptop that would be as light. Anyone else heard of such a thing?
The ruling preserved the Constitutional requirements. Conservatives may be pro-state rights, but they, as any Justice, have an overriding duty to the Constituition, and this was the only outcome that would be consistent with the demands of the Constitution.
By definition, the SC is always supposed to clarify/defend/prescribe enforcement of Constitutional rights. I thought that was their only reason for existence, not primarily interpreting Federal or states laws, but only to determine whether either federal or states laws or their enforcement are in violation of the Constitution. And yet, they obviously differ in how the exact same 200+ year old amended text is interpreted and enforced. Their habitual differences is what labels them as generally leaning Conservative or Liberal.
IANAL, and I did not read the decision. But my statement about how the 5 conservative decisions were viewed is based solely on the commentary I've seen from legal analysts and academics on various outlets such as Lehrer Newshour, Charlie Rose, CNN, Fox, MSNBC/CNBC, CSpan, Wall Street Journal Editors and NYTimes editors. (I had a lot of spare time back then, so I watched and read a lot of news.) I know generally the bias of the commentator, so I always take that into account.
There was not unanimity among legal commentators and academics about the decision, but the general consensus I distinctly recall from the analysis and interviews of that time and subsequent books, was that the 5 justices acted in a manner more consistent with their political (i.e. assumed partisan) bias than with their legal bias. In particular, the majority decision was not widely regarded as a well-written one. I've also heard they some of the majority themselves regret how they wrote and defended their positions, whether or not they would have changed their vote. Some were privately concerned about how it was viewed.
If you wanted to label this as a troll, merely for disagreeing with me or my recollection, then Fox News would love to have you.
With Microsoft, Boies did win, but unfortunately the presiding judge needed a restraining order on his mouth. So maybe Boies did too good of a job.
As for Bush v Gore, it is generally acknowledged from every legal debate I've ever heard, that the 5 justices in favor of Bush were disresptected for it. In particular, because they were the conservatives on the court and a conservative interpretation of the Constitution generally sides with the states laws on matters. Hence, by voting for the Bush team, the conservatives were voting in direct opposition to their philosophy and general previous decisions. They cited "Irrepairable Harm", but by some accounts, even the majority justices did not like how they made or described the decision. In other words, Boies could not overcome the political bias of the judges, intentional bias or not.
Finally, as for Napster, they deserved to lose. It was a stupid case for a great toy but a lousy business. Nobody could have won it.
I must say my linear algebra experience was awesome, mostly because I never went to class. One day I finally made an appearance and saw my TA handing back graded tests; one that I never knew about. My TA politely informed me that I didn't take it, and asked when could I come in to make it up. Eventually, I took it after I had taken the final, and my TA was literally making up the questions on the spot, one by one, while I was working on the previous questions.
My numerical analysis class, on the other hand, was less fun than smoking saw dust.
I readily agree. as a software person, it boggles my mind that the hardware doesn't fail hourly. If Microsoft (or Oracle or any software company) were held to the same standards as Intel/AMD, etc, they wouldn't exist. Intel and CPU companies have been work against the immutable laws of physics, whereas software companies only have to manage their own incompetence and beat back their business departments, IMO.
Apparently those jets did not properly emit some signal that the Patriots use to identify friendly aircraft. They were flying low and may have switched the beacon off, or something like that. I think that was the initial assessment, but I don't know what they eventually concluded.
In his book, A Brief History of Time, I think he said something to the effect that he believes that we'll figure out most of the big questions about the nature of the universe within 10 years or so. That was about 15 years ago. Does anybody remember reading this?
When I saw that, I remember thinking that's naive and contrary to the entire history of scientific research.
Anyway, it reminds me that even some of the best minds say some of the stupidest things. Especially in physics.
I'm not a physicist but I'm pretty damn sure that Stephen Wolfram and Roger Penrose have had some pretty wacky theories when they venture away from straight physics, like into cellular biology, free will, philosophy,...
I didn't mean to imply that SIGGRAPH is for research on anything, but that it was the biggest annual research conference regardless of the field; e.g. medical, scientific, or technological research. When I first attended in 1994, with over 30,000 others, I heard that it was second in attendance only to some conference for Christian groups. Since then, CG has become even cheaper and more pervasive. I don't know about the Christians...
Sony and Nintendo are two major Japanese companies who have done more to spurn innovation in virtual reality and 3D audio/video technology than any other institution, including the military.
No, I think that there are plenty of universities, that you can read about from SIGGRAPH conferences, that solved a lot of the fast 3-D mathematics and algorithms for texturing, and other high-speed rendering techniques, LONG before Sony or Nintendo ever got involved. This research is not a "rehash". Those algorithms are constantly improved because they are never fast enough. I think SIGGRAPH is the very biggest research conference of anything. At least it was about 8 years ago.
Also, the fast 3-D hardware was made first by SGI and was pushed by Jim Clark, aka Netscape founder. VR would not be possible without fast hardware rendering. Then other companies, like ATI, Nvidia, etc made chip sets and graphics boards very cheap, for Wintel boxes.
Also, 3-D games are not the most demanding for VR. Much scientific visualization is FAR FAR more demanding, in all the important areas. It requires more polygons, higher frame rates, higher resolution, and texture memory. These game boxes used the technology many years later, once it was miniaturized and mass produced. N64, PS1 and PS2 all used technology that was already very well established in the research VR world.
Also, advanced dynamically computed sound algorithms are still too complex for game machines. The crap coming out of game machines is very primitive and sounds like simple modulations of samples and FM-synthesizer algorithms. But so far there isn't the same sort of hardware acceleration for these complex sound algorithms; at least not to the degree that OpenGL is implemented in hardware.
Slashdot Mirror
So I would guess that they already use more than 32 bits per item with everything in a single item ID space, or they use 32-bits plus some code indicating the ID-space, or more perhaps a variable length code depending on the item type, e.g. like UTF8. In any case, they should have exceeded 32-bits long ago.
An intelligent designer surely exists. The iPod proves it, not the universe.
More to the point, why are humans required at all in the manufacturing process. I would expect the entire manufacturing and testing process, from sand to plastic-encased chip, to be automated enough that people in bunny suits should not be needed. Maybe they are needed to replace the robots and fill up the supplies, but other than that, what do they do?
I like Google because it is fast, real fast and uncluttered, but the results are not better that Teoma or AllTheWeb. The link analysis that was unique to Google, 6 years ago, was the real quantum leap forward. But now everybody else has caught up. It appears to me that the differentiation is fast, bug-free quality of service and a clean UI.
Short of another breakthrough from Google, I think Microsoft could still clobber Google. Google has got no stickiness.
No, guns don't kill people. Bullets do.
The new desktops appear to be a slick interpretation of Microsoft's new licensing terms and a way to navigate customer demand for PCs without an OS installed. The Microsoft licensing terms, which were put in place on Aug. 1, specify that PC makers must ship PCs with an operating system. The new policy exists to prevent piracy and to better track OS shipments.
And of course, there is a free PGP timestamping service, but unfortunately, that does not have the backing of the USPS.
Anyone know of something similar that is cheap?
This is because humans can be "gamed" in the real world. That is, one can fabricate a "buzz" about things, not simply by overt measures like commercials, but plants in social situations. Sony or some other consumer electronics companies planted people in Times Square and other highly visible situations to pretend to use some cool new gadget. Then people see it and tell their friends and then eventually, they hope, there will be an underground sort of "buzz" generated about the cool product.
If people can't tell that they are being manipulated on the street, I see no hope for machines to be able to figure this out either. Furthermore, when applied to the web, it is not difficult to imagine writing programs that could post good or bad reviews of any product or company to blogs, newsgroups, or whatever.
The posted reviews would have to be varied and still capture the same meaning. That does not seem that hard. Even a simple randomized (CFG) grammar could provide enough variability to make that same statements but with different phrases. It doesn't have to fool a human, just a machine that is looking for grammatical sentences with key adjectives. Perhaps something like...
START := PRODUCT "is" COOL_PHRASE. "I installed one" TIME_PHRASE HOME_PLACE. := "wicked" | "phat" | "fly". := "during Christmas" | "after Kwansa" | "last weekend". := "at my crib" | "in my apartment" | "outside my mansion" | "in my old jetstream" | "in my mom's basement"
COOL_PHRASE
TIME_PHRASE
HOME_PLACE
Basically, link farms combined with natural language generation techniques could make this type of semantic inference about as hard as deciphering spam.
I imagine that a first pass of a WebFountain approach would try to cluster pages or comments based on the inferred sentiment. Then a human would have to read all the individual comments to filter out the semantic spam. Then the remainder would have to be reranked.
Well if you crumple and then scan them then they won't look right when printed back to paper.
Anyway, they could be guilted into doing some interesting science while engorging their egos.
I already use yahoo mail, and I love MyYahoo. I would gladly wait a whole second or two, or even wait for email, to get a useful page of results that is biased to my categories. I don't give a flying fuck about what is most popular to the general web audience. Fundamentally, there is no such thing as "most relevant" because that it usually (maybe exclusively) context-dependent and person-dependent. On a Monday, my searches could be linux or programming related, so I want a customizable filter or checkbox that biases everything for that. But on Friday, I may want to find localized pizza joints, AND I want to find the most popular movies among everyone, not just linux geekdom.
All that should be one category or filter or personalized button away, because my preferences don't change that often, even though my search needs change every time. And I think that if everybody was using some personalization, it would become more difficult to build pages and link farms that would always rise to the top. Spamming would be harder because the optimizers would have to guess what personlization biases the vast majority of users would choose.
And even simple filtering and reranking should be easily possible in about a second. e.g. restrictions to selected categories in dmoz.com, or sites within a certain distance or weight of some human-edited directory, or that contain any synonym for some set of terms, and many others that are easy to think of. None of this stuff is rocket science or new or computationally expensive. But this sort of remembered customization would go a long way to saving me hours of search time every week. I'm disappointed that it hasn't happened, even since altavista days.
My guess is that google will eventually have to provide logins and accounts to allow personalization, because cookies may not be enough. Cookies don't stay in sync with my laptop or when I go to the library or my friends computer. Yahoo, MSN, and AOL will easily be able to do this stuff. I'd like Google to remember me also. In other words, Google may have to start looking even more like a general purpose portal.
I don't think it was openness alone, but that they gave Microsoft exclusive rights to DOS for the clones. Then all the clones ran with the same MS-DOS, which was nearly identical to IBM's PC-DOS variant. So Microsoft got the software profits on the clones. It was a very, very smart move. I don't think that many big companies would have anticipated the consequences, given that Microsoft was a tiny company selling Basic.
I would be thrilled to not have to care about CO2. Then a great reason to avoid fossil fuels is because everyone involved in oil production is a bunch of fucking assholes, Exxon, OPEC, Texans, terrorists or their apologists, take your pick...
Heck, if we have fuel cell based cars, then you could also drink the water from the exhaust of your car. It would be more pure than Evian. Screw the French too.
Don't you remember? GPL is the cancer, and Windows is the virus that cures it.
This way you could use the PSP to offload pictures and do more serious editing. Editing on the camera is a pain in the ass. What it needs is much, much larger storage. It sounds like it doesn't have a harddrive, since it's only got 1.8G storage. If it did, like the iPod variety, then you could listen to a truckload of music and store hours of video and hundreds of thousands of shots from a camera; all without lugging around a laptop.
Unfortunately I haven't seen any palmtop that could did this as cheaply, or laptop that would be as light. Anyone else heard of such a thing?
By definition, the SC is always supposed to clarify/defend/prescribe enforcement of Constitutional rights. I thought that was their only reason for existence, not primarily interpreting Federal or states laws, but only to determine whether either federal or states laws or their enforcement are in violation of the Constitution. And yet, they obviously differ in how the exact same 200+ year old amended text is interpreted and enforced. Their habitual differences is what labels them as generally leaning Conservative or Liberal.
IANAL, and I did not read the decision. But my statement about how the 5 conservative decisions were viewed is based solely on the commentary I've seen from legal analysts and academics on various outlets such as Lehrer Newshour, Charlie Rose, CNN, Fox, MSNBC/CNBC, CSpan, Wall Street Journal Editors and NYTimes editors. (I had a lot of spare time back then, so I watched and read a lot of news.) I know generally the bias of the commentator, so I always take that into account.
There was not unanimity among legal commentators and academics about the decision, but the general consensus I distinctly recall from the analysis and interviews of that time and subsequent books, was that the 5 justices acted in a manner more consistent with their political (i.e. assumed partisan) bias than with their legal bias. In particular, the majority decision was not widely regarded as a well-written one. I've also heard they some of the majority themselves regret how they wrote and defended their positions, whether or not they would have changed their vote. Some were privately concerned about how it was viewed.
If you wanted to label this as a troll, merely for disagreeing with me or my recollection, then Fox News would love to have you.
All of these cases are herrings, I think.
With Microsoft, Boies did win, but unfortunately the presiding judge needed a restraining order on his mouth. So maybe Boies did too good of a job.
As for Bush v Gore, it is generally acknowledged from every legal debate I've ever heard, that the 5 justices in favor of Bush were disresptected for it. In particular, because they were the conservatives on the court and a conservative interpretation of the Constitution generally sides with the states laws on matters. Hence, by voting for the Bush team, the conservatives were voting in direct opposition to their philosophy and general previous decisions. They cited "Irrepairable Harm", but by some accounts, even the majority justices did not like how they made or described the decision. In other words, Boies could not overcome the political bias of the judges, intentional bias or not.
Finally, as for Napster, they deserved to lose. It was a stupid case for a great toy but a lousy business. Nobody could have won it.
Yeah, maybe Google or Yahoo or some internet company will become the next evil empire. Evil companies are always cute when they are still a young pup.
AOL already missed their shot as the big bad, since they are currently imploding.
I'm a vegetarian and a foulmouth. I'd prefer... Like turning shit backing into pizza.
My numerical analysis class, on the other hand, was less fun than smoking saw dust.
I readily agree. as a software person, it boggles my mind that the hardware doesn't fail hourly. If Microsoft (or Oracle or any software company) were held to the same standards as Intel/AMD, etc, they wouldn't exist. Intel and CPU companies have been work against the immutable laws of physics, whereas software companies only have to manage their own incompetence and beat back their business departments, IMO.
Apparently those jets did not properly emit some signal that the Patriots use to identify friendly aircraft. They were flying low and may have switched the beacon off, or something like that. I think that was the initial assessment, but I don't know what they eventually concluded.
When I saw that, I remember thinking that's naive and contrary to the entire history of scientific research. Anyway, it reminds me that even some of the best minds say some of the stupidest things. Especially in physics.
I'm not a physicist but I'm pretty damn sure that Stephen Wolfram and Roger Penrose have had some pretty wacky theories when they venture away from straight physics, like into cellular biology, free will, philosophy, ...
No, I think that there are plenty of universities, that you can read about from SIGGRAPH conferences, that solved a lot of the fast 3-D mathematics and algorithms for texturing, and other high-speed rendering techniques, LONG before Sony or Nintendo ever got involved. This research is not a "rehash". Those algorithms are constantly improved because they are never fast enough. I think SIGGRAPH is the very biggest research conference of anything. At least it was about 8 years ago.
Also, the fast 3-D hardware was made first by SGI and was pushed by Jim Clark, aka Netscape founder. VR would not be possible without fast hardware rendering. Then other companies, like ATI, Nvidia, etc made chip sets and graphics boards very cheap, for Wintel boxes.
Also, 3-D games are not the most demanding for VR. Much scientific visualization is FAR FAR more demanding, in all the important areas. It requires more polygons, higher frame rates, higher resolution, and texture memory. These game boxes used the technology many years later, once it was miniaturized and mass produced. N64, PS1 and PS2 all used technology that was already very well established in the research VR world.
Also, advanced dynamically computed sound algorithms are still too complex for game machines. The crap coming out of game machines is very primitive and sounds like simple modulations of samples and FM-synthesizer algorithms. But so far there isn't the same sort of hardware acceleration for these complex sound algorithms; at least not to the degree that OpenGL is implemented in hardware.