Win7 most definitely does some of the things mentioned in the article out of the box, such as loading resources from executables and producing thumbnails for images on USB drives.
Its likely that you can dig out of any modern OS sandbox (Linux or otherwise) when giving them malformed input.. look at how much effort Apple has put into protecting iOS, and contrast that with how many ways that its already been rooted... and thats a completely locked down example of failure. Now imagine how badly Windows, Linux, and mainstream BSD must be at sandboxing.
As Raymond Chen would say... this stuff is going on on the wrong side of the airtight hatchway.
What do you mean not knowing why? The reason was pretty clear. The user was looking for X, and then wen to Y. Thats the best reason possible for a search engine to correlate X with Y.
Every page has a relevance to every search term, in both google and bing.
Specifically there would be a function f(term, url) that gives a numerical weight of the association between the term and the url. In proper terms, this value would be a measure of the correlation between to two inputs to the function and would likely be constricted to values between [-1.0, +1.0,] fitting in precisely with the domain of mathematics dealing with correlation. Values below 0 are negative correlation, values above 0 are positive correlations.
Now, the engines will not be calling f() for every URL in its database when a user does a search, but will use various shortcuts to severely cull the set of url's it needs to generate a correlation value for. After this, they will then additionally cull the list using some cutoff value that is related to the highest value returned by all the calls to f() for a given term. If the highest value is 0.7 then perhaps all results that are not at least a 0.5 will be dropped from the list. Then the engine responds with the best 10 or so items and provides a link to get the next best 10, and so on. If the highest correlation value is below some cutoff (likely something well about 0) then the engine returns no results at all.
What you are arguing is that Bing did not have those URL's in its database at all, but that is simple not the case. Those URL's were already in Bings database. What Google did was manipulate some of the derived correlation values to push them above the cutoff, and additionally found that in many cases they could not even succeed at doing this (probably because the URL's never make it to f() for that search term, culled wholesale prior to that)
You're splitting hairs. The click-stream information is Google search results.
My site Y links to X due to my actions to specifically include that link on a page with content that I created, and indexers such as google takes note of this and modifies the rank of Y based on some metrics related to X.
And how did the "user" find out about the link? Ah, that's right, via Google!
I found out about slashdot via an IRC channel (long ago, before IRC was broken up)
Does that mean that when I associate ignorant posts on slashdot by iserlohn, who makes them in response to articles about microsoft, that I have stolen the association from #hack or #phreak?
So Microsoft is tracking users' searches for specific strings on Google, and giving links they commonly click higher rankings.
No, you still don't get it. You are ALMOST there tho.
Microsoft is tracking what users enter into search boxes, and then associates those search terms with the sites that they then visit.
This is, in fact, a component of how many search engines have worked for a very long time. From altavista to yahoo. The difference is only where the information is collected. It used to be that search engines would return search results that had specially crafted links that bounced the users later clicks through the search provider (via a dynamic redirect page. perhaps www.altavista.com/redirect?thesityoureallywanted.com/thepageyoureallywanted.html) These days search engines dont use redirects for this collection mechanism, but instead they all do it in other way.
No, I compared it with a i7-950, a Phenom II 1100T, and a i7-980X. Actually.. I only compared it with an i7-980X.. the rumors compared it to the others. Did you not read my post, or did you intentionally ignore what you read?
Well and good, but that is not the same thing as a 32nm Sandy Bridge chip [snip] a 2600 (roughly same price) is 3.4Ghz and has a higher turbo speed.
Nobody said it was, but I also mentioned the Sandy's. Did you not read my post, or did you intentionally ignore what you read?
The 2600 is a Sandy in the 9000+ PassMark category... oh.. exactly where the Bulldozer rumors put Bulldozer..Why would I talk about price on similarly performing parts in the same paragraph? hmmmmmm. I wonder... did you even read my post?
Second you have to consider that throwing more cores at a problem doesn't do a lot of good in the consumer arena.
Using your logic, a consumer is good to go with their dual core Pentium 4. email, facebook, and netflix... Right? Why are you talking about Sandy's again?
Given those things, the AMD chip doesn't sound all that impressive. If the 1.5x a 950 number is accurate it is probably not much faster, if at all, than a 2600. That will be made worse for consumers in that most apps will not be able to use a large part of the bulldozer's power and thus be even slower in practice.
Each pair of cores shares some resources, so if that 1.5x is for an 8-threaded task, it is even better than 1.5x per thread in 4-threaded scenarios. Nobody, however, knows if that 1.5x is exclusively multi threaded tasks.. you seem to want to argue as if you do know. Do you?
The 8 core bulldozer is rumored to benchmark 1.5x faster than an i7-950 (as well as 1.5x the similarly performing Phenom II x6 1100T).. this info was supposedly leaked a few days ago.
Could be a fake, but I wouldn't be surprised given the PassMark numbers for the i7-950 (6346 @ $286), the Phenom II x6 1100T (6174 @ $260), and the i7-980X (10472 @ $1000)
AMD is moving to the same process size that Intel has been enjoying, and a smaller process size usually means large performance increments (just as it did with Intel), AMD has been working on this one for a very long time and they havent redesigned their architecture since the athlon 64's, so I think there is a good chance that the rumors or more or less true. AMD will likely be offering a single chip that can attain 9000..10000 on PassMark, territory that Intel had an exclusive to on single-CPU systems this past year.
The most important thing, tho, is price. If AMD drops these off at $300 like I suspect then Intel will have to seriously rethink its pricing strategy at the high end, but if AMD can't push them out for under $500 then Intel probably has nothing to worry about with its Sandy's pushing 9000+ PassMark's for less.
What matters for Theora support is the Theora decoder, which has been finished for years. The Theora bitstream format has been frozen since 2004 and any decoder written since then can play Theora files.
The reference decoder implementation has been changed recently (15 months ago) and that includes bug fixes.
There is thus a set of input files that will not produce the same output, BY DEFINITION.
Well then it is good that this is not the case. That hardware is generic stuff, it could easily do WebM.
Yet nobody is doing it. There isnt a single device doing hardware accelerated WebM right now, but plenty of plans to add hardware acceleration to future devices. Nobody is adding hardware acceleration to existing devices, which tells me that in spite of your claim it "could easily" be done.. it is either not easy, or can't be done.
I like how you failed to properly and accurately quote the text:
Google hereby grants to you a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to...
"irrevocable.. 'cept for those cases where we reserve the right to revoke..."
es, it's not black and white, but it's also the case that Theora being free makes the lack of inclusion either a sign of a choice on their part or a belief that Theora is so underused that it ranks below a ton of old codec; that's a little hard to believe.
..and by "free" you mean that if Microsoft includes the theora codec with windows, then they also have to provide source code upon request... that whole GPL thing that Theora falls under...
..and lets suppose they included the theora codec in 2009.. well you know that the reference implementation has been changed multiple times since then, right? bug-fixes and all that? (how can anything in the reference be a bug? but hey... its not my project... I didnt call a beta a reference.. they did)
I remember the comparisons of Theora Vs "H.264" and the only one even remotely favorable for Theora was actually underhandedly using H.263 for most of the comparisons. If you need a citation beyond the one I just gave, notice how this page never mentions the fact that that is H.263, but instead propagandizes that it is H.264 in their comparison images (directly copied from the first citation, which does at least honestly note that it is H.263)
Sorry. H.264 is miles ahead of Theora, *especially* at low bit rates, as you can see here, take note of the 486Kbit still images halfway down. On the left is 1Mbit Theora, on the right is 486Kbit Theora, and in the middle are the best images... 486Kbit H.264
You can ask him if he read it or not, but I don't have to ask you. I know that if you did bother to read it, than you didn't comprehend it. Nowhere does the article state what you are claiming.. only the slashdot summary hints at what you are claiming.
That isnt a correct analogy. A correct analogy would be that you are looking to see what other people are buying and then produce your own top 100 list.
For example, one important parameter for rendering the Mandelbrot set is the iteration count after which you decide that you're not escaping. Doubling this count will more or less double the time it takes to render.
umm, no.. UNLESS the entire render is lake., AND there is no periodicity checking, AND... (I could go on for awhile)
I would have thought someone trying to come off as an expert would not have begun their "I'm smarter that you" post with something so obviously incorrect to anyone who has ever actually written a mandelbrot or julia renderer.
Increasing the iteration count does not change the escape time of any point outside the set, and it is the outside of the set that people look at and zoom in on. All the interesting areas has little no no lake at all in them.
Slashdot Mirror
Is that like a pinky swear?
Win7 most definitely does some of the things mentioned in the article out of the box, such as loading resources from executables and producing thumbnails for images on USB drives.
Its likely that you can dig out of any modern OS sandbox (Linux or otherwise) when giving them malformed input.. look at how much effort Apple has put into protecting iOS, and contrast that with how many ways that its already been rooted... and thats a completely locked down example of failure. Now imagine how badly Windows, Linux, and mainstream BSD must be at sandboxing.
As Raymond Chen would say... this stuff is going on on the wrong side of the airtight hatchway.
Bing added that association without knowing why.
What do you mean not knowing why? The reason was pretty clear. The user was looking for X, and then wen to Y. Thats the best reason possible for a search engine to correlate X with Y.
Google Engineers knowingly and willfully copied made up search results into Bing's database.
See, we can play this copy game if you want.
The point is that the users typed those search queries into Google
No they didn't. They types them into the bing toolbars searchbox, which then uses whatever search engine is configured.
Why are making arguments when your beginning premise is demonstrably wrong?
Every page has a relevance to every search term, in both google and bing.
Specifically there would be a function f(term, url) that gives a numerical weight of the association between the term and the url. In proper terms, this value would be a measure of the correlation between to two inputs to the function and would likely be constricted to values between [-1.0, +1.0,] fitting in precisely with the domain of mathematics dealing with correlation. Values below 0 are negative correlation, values above 0 are positive correlations.
Now, the engines will not be calling f() for every URL in its database when a user does a search, but will use various shortcuts to severely cull the set of url's it needs to generate a correlation value for. After this, they will then additionally cull the list using some cutoff value that is related to the highest value returned by all the calls to f() for a given term. If the highest value is 0.7 then perhaps all results that are not at least a 0.5 will be dropped from the list. Then the engine responds with the best 10 or so items and provides a link to get the next best 10, and so on. If the highest correlation value is below some cutoff (likely something well about 0) then the engine returns no results at all.
What you are arguing is that Bing did not have those URL's in its database at all, but that is simple not the case. Those URL's were already in Bings database. What Google did was manipulate some of the derived correlation values to push them above the cutoff, and additionally found that in many cases they could not even succeed at doing this (probably because the URL's never make it to f() for that search term, culled wholesale prior to that)
You're splitting hairs. The click-stream information is Google search results.
My site Y links to X due to my actions to specifically include that link on a page with content that I created, and indexers such as google takes note of this and modifies the rank of Y based on some metrics related to X.
By your logic, Google is stealing links from me.
And how did the "user" find out about the link? Ah, that's right, via Google!
I found out about slashdot via an IRC channel (long ago, before IRC was broken up)
Does that mean that when I associate ignorant posts on slashdot by iserlohn, who makes them in response to articles about microsoft, that I have stolen the association from #hack or #phreak?
So Microsoft is tracking users' searches for specific strings on Google, and giving links they commonly click higher rankings.
No, you still don't get it. You are ALMOST there tho.
Microsoft is tracking what users enter into search boxes, and then associates those search terms with the sites that they then visit.
This is, in fact, a component of how many search engines have worked for a very long time. From altavista to yahoo. The difference is only where the information is collected. It used to be that search engines would return search results that had specially crafted links that bounced the users later clicks through the search provider (via a dynamic redirect page. perhaps www.altavista.com/redirect?thesityoureallywanted.com/thepageyoureallywanted.html) These days search engines dont use redirects for this collection mechanism, but instead they all do it in other way.
First you are comparing the bulldozer to a 950.
No, I compared it with a i7-950, a Phenom II 1100T, and a i7-980X. Actually.. I only compared it with an i7-980X.. the rumors compared it to the others. Did you not read my post, or did you intentionally ignore what you read?
Well and good, but that is not the same thing as a 32nm Sandy Bridge chip [snip] a 2600 (roughly same price) is 3.4Ghz and has a higher turbo speed.
Nobody said it was, but I also mentioned the Sandy's. Did you not read my post, or did you intentionally ignore what you read?
The 2600 is a Sandy in the 9000+ PassMark category... oh.. exactly where the Bulldozer rumors put Bulldozer..Why would I talk about price on similarly performing parts in the same paragraph? hmmmmmm. I wonder... did you even read my post?
Second you have to consider that throwing more cores at a problem doesn't do a lot of good in the consumer arena.
Using your logic, a consumer is good to go with their dual core Pentium 4. email, facebook, and netflix... Right? Why are you talking about Sandy's again?
Given those things, the AMD chip doesn't sound all that impressive. If the 1.5x a 950 number is accurate it is probably not much faster, if at all, than a 2600. That will be made worse for consumers in that most apps will not be able to use a large part of the bulldozer's power and thus be even slower in practice.
Each pair of cores shares some resources, so if that 1.5x is for an 8-threaded task, it is even better than 1.5x per thread in 4-threaded scenarios. Nobody, however, knows if that 1.5x is exclusively multi threaded tasks.. you seem to want to argue as if you do know. Do you?
The 8 core bulldozer is rumored to benchmark 1.5x faster than an i7-950 (as well as 1.5x the similarly performing Phenom II x6 1100T) .. this info was supposedly leaked a few days ago.
Could be a fake, but I wouldn't be surprised given the PassMark numbers for the i7-950 (6346 @ $286), the Phenom II x6 1100T (6174 @ $260), and the i7-980X (10472 @ $1000)
AMD is moving to the same process size that Intel has been enjoying, and a smaller process size usually means large performance increments (just as it did with Intel), AMD has been working on this one for a very long time and they havent redesigned their architecture since the athlon 64's, so I think there is a good chance that the rumors or more or less true. AMD will likely be offering a single chip that can attain 9000..10000 on PassMark, territory that Intel had an exclusive to on single-CPU systems this past year.
The most important thing, tho, is price. If AMD drops these off at $300 like I suspect then Intel will have to seriously rethink its pricing strategy at the high end, but if AMD can't push them out for under $500 then Intel probably has nothing to worry about with its Sandy's pushing 9000+ PassMark's for less.
What matters for Theora support is the Theora decoder, which has been finished for years. The Theora bitstream format has been frozen since 2004 and any decoder written since then can play Theora files.
The reference decoder implementation has been changed recently (15 months ago) and that includes bug fixes.
There is thus a set of input files that will not produce the same output, BY DEFINITION.
Well then it is good that this is not the case. That hardware is generic stuff, it could easily do WebM.
Yet nobody is doing it. There isnt a single device doing hardware accelerated WebM right now, but plenty of plans to add hardware acceleration to future devices. Nobody is adding hardware acceleration to existing devices, which tells me that in spite of your claim it "could easily" be done.. it is either not easy, or can't be done.
Google hereby grants to you a perpetual, worldwide, non-exclusive, no-charge, royalty-free, irrevocable (except as stated in this section) patent license to...
"irrevocable.. 'cept for those cases where we reserve the right to revoke..."
es, it's not black and white, but it's also the case that Theora being free makes the lack of inclusion either a sign of a choice on their part or a belief that Theora is so underused that it ranks below a ton of old codec; that's a little hard to believe.
I remember the comparisons of Theora Vs "H.264" and the only one even remotely favorable for Theora was actually underhandedly using H.263 for most of the comparisons. If you need a citation beyond the one I just gave, notice how this page never mentions the fact that that is H.263, but instead propagandizes that it is H.264 in their comparison images (directly copied from the first citation, which does at least honestly note that it is H.263)
Sorry. H.264 is miles ahead of Theora, *especially* at low bit rates, as you can see here, take note of the 486Kbit still images halfway down. On the left is 1Mbit Theora, on the right is 486Kbit Theora, and in the middle are the best images... 486Kbit H.264
I think he was addressing it to the person that can't figure out to whom it is addressed.
Isnt that what Google does with analytics, only not nearly as bad a privacy violation (since Bing Toolbar is opt-in)?
He read the article, you didn't.
You can ask him if he read it or not, but I don't have to ask you. I know that if you did bother to read it, than you didn't comprehend it. Nowhere does the article state what you are claiming.. only the slashdot summary hints at what you are claiming.
That isnt a correct analogy. A correct analogy would be that you are looking to see what other people are buying and then produce your own top 100 list.
For example, one important parameter for rendering the Mandelbrot set is the iteration count after which you decide that you're not escaping. Doubling this count will more or less double the time it takes to render.
umm, no.. UNLESS the entire render is lake., AND there is no periodicity checking, AND ... (I could go on for awhile)
I would have thought someone trying to come off as an expert would not have begun their "I'm smarter that you" post with something so obviously incorrect to anyone who has ever actually written a mandelbrot or julia renderer.
Increasing the iteration count does not change the escape time of any point outside the set, and it is the outside of the set that people look at and zoom in on. All the interesting areas has little no no lake at all in them.
woooooooooooosh!!
Who decided that search = process, and who gave them that authority?
A search engine would index web pages and shit. Alpha doesnt do that.
Alpha only indexes tabulated data and provides tools to process those tables.
Alpha is not a search engine.