Thanks for pointing that out. It caused me to think more deeply about the matter and learn more about drives and video signals than I knew when I made. if performance is not a problem, I have to wonder why bother with RAID0 at all then? As I understand it, RAID0's only real benefit is better performance.
I still think they used RAID0 here for performance. My own tuner card can crank out video streams at up to 12mbit/sec, which, if we plug 12 into your calculation instead of 4, would swamp the drive, or come really close, right? HDTV quality, as I understand it, is 15mbit/second... so a 12mbit maximum would still entail a loss of quality over a direct signal to screen feed. I'd have to assume in this case that the owner of the PVR would not want to degrade the signal any more than necessary (since the summary mentions they have HDTV capability) and that they really are maxing out their tuner cards whenever possible.
That's what I was going to say, eleven tuners and only 1TB of HD? And the HD is only RAID0 to boot (and the primary reason for RAID0 here is performance because you are going to need fast disk I/O if you get 11 tuners all trying to work at the same time). In this situation if one of the four drives goes down, you lose a lot of data. I'd rather see multiple PVR backends managing the tuners, each tuner with its own dedicated drive. Double the number of drives and go to RAID1 and then you don't have to worry about a drive failure much either.
I mean, can you imagine a beowulf cluster of PVRs?;)
I not only KNEW ABOUT no-cache, but have EVEN USED.htaccess to PREVENT certain types of DEEP LINKING in the past (although only for images because I HATE WHEN SOMEONE LINKS TO MY IMAGES BUT MAKES IT LOOK LIKE IT'S THEIR IMAGE--but these days I realized that they were offering me something special: the opportunity to put whatever image I wanted on their website).
Anyway, got any guesses what happens to your traffic if Google can't cache your site? The user who gets used to using the color-coded cache to find stuff is likely to skim right past results listings that don't have caches at all. As for deep linking, I think you're missing the point. Which isn't that web site owners want to piss off users or make life difficult by adding hurdles, but that the search engines are theoretically making it less likely that users will stay on the site past finding the page they wanted. Again, if the users ever actually get to the site.
As to whether this is really a problem is hard to say, and I don't think I was implying that I have any strong feelings about it one way or another right now. Generally the only people who care about web site "stickiness" are people trying to sell advertising space. I'm not one of those people. And I'm not sure I believe that using ad revenues to pay for "free" content is a guarantee that the content will be worth having. Look at television, the only thing that TV has more of than ads is crappy content. So will it be a loss if some web sites find their ads revenues dry up? Maybe not.
And I don't think that others, like the widget sellers that Nielson refers to, are going to have as much trouble with this as he makes it sound like. Those web sites are not the value those businesses are adding to the world. And I don't think there is going to be any "race to the top" with respect to bidding up ad impressions like he mentions. Because the scenario he paints is one with an infinite supply of customers, not one where there is constant competition for a limited pool of customers. Ultimately factors like providing quality products, good customer service experiences, and decent prices will do more to drive customers back to those web sites than ad price wars.
There is some question whether a search engine will actually generate traffic. Through the miracle of caching the user may never really leave the search engine to view a web site's content. Another problem is that if the user is looking for one specific thing, no more, no less, then they go directly to that page on your site, get what they want, and leave just as quickly. TFA says this is great for users, but for someone operating a site, it may just be that better searches are lowering their traffic since the user doesn't really even need to interact with your site to get something from it.
As long as the browser has the ability to respond to all pings or respond to some pings or respond to no pings, depending on a user pref, I think the default should be to respond to all pings. Just like when I load slashdot.org they link in Javascript scripts from TWO different 3rd parties (Google Analytics and something else). These pings don't do anything different than URLs like http://www.example.com/redirect.cgi?http://www.foo bar.org do. In fact, the pings discourage the use of lame URLs like http://www.example.com/redirect.cgi?s0m3_w3bs1t3 where you have no idea where you're headed until you click the link, so in that sense they'd be a marked improvement.
That is true, and with Cygwin you can work on Windows almost like it was a more serious operating system, but it's always a little bit more difficult. Where the F/OSS stuff "just works" on Linux often requires a bit of handholding on Windows.
The guy I admire is this "Eric Reynolds", who according to TFA wrote some open source manifesto called "The Cathedral and the Bazaar". He sounds like a real hacker's hacker... no mere geek, y'know.
Or maybe I just admire ZDNet's ability to publish pure tripe and get it to the front page of Slashdot.
If you think a tinfoil-hatter is going to refuse all cookies, but still allow this sort of Javascript web bug I think you've vastly underestimated the power of paranoia. Personally, after reading this very article, I decided to install the NoScript extension to Firefox, which makes Google Analytics completely useless for tracking me, since no Javascript is ever run without my explicit permission. And like for Slashdot's web bugs, I can block the JS coming from google.com, but not JS coming from slashdot.org. So if the site uses Google Analytics JS code I can always keep that off, even while allowing site-based code for things like menus.
I routinely accept site-originated cookies, because these are critical to managing per site user prefs. Additionally, cookies don't have any power that I can't manage myself (so there's much less to be paranoid about in the first place). I can see the full text of the cookie, and the cookie mechanism can only send back the cookie as I received it. This Javascript bug tells Google whatever Javascript can learn about my system. It is significantly more invasive than cookies.
And you can't build this sort of page view history on a per user basis yourself using a cookie and a tracking database? This isn't exactly rocket science here. It's just another chapter in Google's "pretty soon we'll have one of everything" approach to doing business on the web. And it will all be tied into making you either want to sign up to put ads on your site, or getting you to view pages that have ads on them.
He may have been thinking of "ghost in the machine" (and even then wouldn't really be using the phrase correctly)... but still, can you believe an editor let this go? This isn't just from their online version, it says the article is from their print edition. If it were just the online version I'd be inclined to forgive, but it seems like the quality of print media has been nosediving the last few years. I see egregious typos all the time. Really. Every hardcover book I've read recently has had some embarrassingly bad typos. Which is nothing compared to a completely inept use of a fancy sounding phrase in a puff piece like this. Get me out of this hand-basket!
Spoken like someone who never experienced the joy of trying to find a shop to make prints off your old 110 or 126 film.
Prints are not a function of their original capture method. You can get prints of digital images that are at least as archival in quality as an photomat 35mm 4x6 glossy.
As an added bonus, making a print of a digital image does not expose your film media to any danger. Even cartridge formats like 110 and APS carry some risk to the negative in terms of dust, scratching, etc, when making a print. Not to mention that while film has improved markedly over the years, it is still a highly unstable base on which to store an image.
All file transfers must be initiated by user action.
This seems overly broad. How do you automate internal file transfers with a policy like this? Do you have no operational systems that need to provide data extracts to analysis systems or the like? Or do you allow automated transfer in documented and approved situations?
Re:but wait did the MS apologist not say
on
Two New WMF Bugs Found
·
· Score: 2, Informative
Yes becuase breaking hundreds of people off their regular duties
Actually, given MS' scope and resources I fully expect them to have a staff whose regular duties consist solely of fixing these types of problems.
Personally I'm at least as concerned about the ability to sell used media. Right now when I buy a book, CD, or DVD I can watch it as many times as I want and sell it when I'm done/bored/broke. But how does this work with digital media, especially if it's got any sort of DRM on it that is linked to a specific user?
For example: I have a significant chunk of the Cannibal Corpse discography in MP3 that I "bought" from emusic.com... can I sell it? And how much worse does this question get if the file is somehow DRMed? How would I sell it if the DRM won't let anyone else play it? And if I can't sell it, can my daughter inherit it tax free, rather than having to pay taxes on something that she probably can't sell either? Will she even be able to inherit it?
The fact that I can even ask these questions indicates to me that the whole model of "intellectual property" is fundamentally broken and bound to cause more problems than it solves at this point. And if Google is going to attempt to further the cause of "intellectual property" by selling DRMed products, then I think this goes against their whole "don't be evil" motto.
Then one might respectfully suggest that you go and learn something about system administration.
One might. Except that what we're discussing here is not administering large, multi-user servers, rather desktops. And I have plenty of experience following the recommendations to use multiple partitions on desktops, and all of it is negative. In eveery case where I've done that I've ended up regretting it because one partition or another was full long before the others, requiring a nasty mess of symlinks to keep the system functional without wiping the drive entirely. These recommendations to use multiple partitions are just what I said: cargo cult systems administration. Desktop Linux users looking at how admins set up servers and just doing the same, without really thinking about why or how.
Yes, there are times when using partitions as a poor-man's quota system (like your web server example) makes some sense. It's a good point. But again, my main objection to multiple partitioning isn't that it shouldn't be done, but rather that it's recommended all the time by people who haven't really thought about the cases where it actually makes sense.
How many of you own Apple notebooks? How many have blown away OS X to put a PPC linux distro on there?
Well, I might be interested in buying an Apple notebook, except that I can guarantee that I'd blow away OS X to install Linux. Given that I have no intention of ever using OS X why would I want to buy a machine which is premium priced precisely because it includes OS X?
I think I'll stick to buying OS-free used laptops (usually corporate castoffs) and installing an operating system I know I'll like on there.
Nonsense. The idea of having *any* separate partitions on a single disk machine is just cargo cult systems administration. On a multi-disk system it's obviously necessary, but even there the strategy of partitioning by system or data is only really useful if one drive is fairly small and the other drive is quite large. if you have two 120gb drives in the same system you are going to be wasting gigs and gigs of drive space on the drive devoted only to systems stuff.
I don't even buy the idea of having swap space in its own partition. What if you need more swap? Then you have to create a swap file on an existing partition and manage two separate swap files in the future. Ugh.
As for reinstalling, generally the only time I'm reinstalling the whole OS is when I am switching distros or when I feel like starting fresh on new, larger HD. Which I do this a lot less frequently than having to create a bunch of annoying symlinks across partitions to help alleviate problems due to running out of space in some partitions. Even if I'm just swapping a drive and not trying to reinstall, having data in a separate partition doesn't make the drive swap any easier, you still have to copy the whole thing.
My opening sentence was too vague--maybe even a full blown misstatement. I should have said "Public domain works can be used to create copyrighted works" and then focused on how this is not true for public property. My point wasn't that the public domain work is no longer in the public domain, but that the originaly analogy (that public domain is like public property) is poor because privatizing public property does mean the property is no longer public.
Slashdot Mirror
Thanks for pointing that out. It caused me to think more deeply about the matter and learn more about drives and video signals than I knew when I made. if performance is not a problem, I have to wonder why bother with RAID0 at all then? As I understand it, RAID0's only real benefit is better performance.
I still think they used RAID0 here for performance. My own tuner card can crank out video streams at up to 12mbit/sec, which, if we plug 12 into your calculation instead of 4, would swamp the drive, or come really close, right? HDTV quality, as I understand it, is 15mbit/second... so a 12mbit maximum would still entail a loss of quality over a direct signal to screen feed. I'd have to assume in this case that the owner of the PVR would not want to degrade the signal any more than necessary (since the summary mentions they have HDTV capability) and that they really are maxing out their tuner cards whenever possible.
That's what I was going to say, eleven tuners and only 1TB of HD? And the HD is only RAID0 to boot (and the primary reason for RAID0 here is performance because you are going to need fast disk I/O if you get 11 tuners all trying to work at the same time). In this situation if one of the four drives goes down, you lose a lot of data. I'd rather see multiple PVR backends managing the tuners, each tuner with its own dedicated drive. Double the number of drives and go to RAID1 and then you don't have to worry about a drive failure much either.
I mean, can you imagine a beowulf cluster of PVRs? ;)
So do you recommend for Aunt Tillie instead of Debian?
I not only KNEW ABOUT no-cache, but have EVEN USED .htaccess to PREVENT certain types of DEEP LINKING in the past (although only for images because I HATE WHEN SOMEONE LINKS TO MY IMAGES BUT MAKES IT LOOK LIKE IT'S THEIR IMAGE--but these days I realized that they were offering me something special: the opportunity to put whatever image I wanted on their website).
Anyway, got any guesses what happens to your traffic if Google can't cache your site? The user who gets used to using the color-coded cache to find stuff is likely to skim right past results listings that don't have caches at all. As for deep linking, I think you're missing the point. Which isn't that web site owners want to piss off users or make life difficult by adding hurdles, but that the search engines are theoretically making it less likely that users will stay on the site past finding the page they wanted. Again, if the users ever actually get to the site.
As to whether this is really a problem is hard to say, and I don't think I was implying that I have any strong feelings about it one way or another right now. Generally the only people who care about web site "stickiness" are people trying to sell advertising space. I'm not one of those people. And I'm not sure I believe that using ad revenues to pay for "free" content is a guarantee that the content will be worth having. Look at television, the only thing that TV has more of than ads is crappy content. So will it be a loss if some web sites find their ads revenues dry up? Maybe not.
And I don't think that others, like the widget sellers that Nielson refers to, are going to have as much trouble with this as he makes it sound like. Those web sites are not the value those businesses are adding to the world. And I don't think there is going to be any "race to the top" with respect to bidding up ad impressions like he mentions. Because the scenario he paints is one with an infinite supply of customers, not one where there is constant competition for a limited pool of customers. Ultimately factors like providing quality products, good customer service experiences, and decent prices will do more to drive customers back to those web sites than ad price wars.
There is some question whether a search engine will actually generate traffic. Through the miracle of caching the user may never really leave the search engine to view a web site's content. Another problem is that if the user is looking for one specific thing, no more, no less, then they go directly to that page on your site, get what they want, and leave just as quickly. TFA says this is great for users, but for someone operating a site, it may just be that better searches are lowering their traffic since the user doesn't really even need to interact with your site to get something from it.
As long as the browser has the ability to respond to all pings or respond to some pings or respond to no pings, depending on a user pref, I think the default should be to respond to all pings. Just like when I load slashdot.org they link in Javascript scripts from TWO different 3rd parties (Google Analytics and something else). These pings don't do anything different than URLs like http://www.example.com/redirect.cgi?http://www.foo bar.org do. In fact, the pings discourage the use of lame URLs like http://www.example.com/redirect.cgi?s0m3_w3bs1t3 where you have no idea where you're headed until you click the link, so in that sense they'd be a marked improvement.
That is true, and with Cygwin you can work on Windows almost like it was a more serious operating system, but it's always a little bit more difficult. Where the F/OSS stuff "just works" on Linux often requires a bit of handholding on Windows.
Or maybe I just admire ZDNet's ability to publish pure tripe and get it to the front page of Slashdot.
Or maybe not.
The only part of that statement that's true is "Windows... has a large selection of ... proprietary software available for it."
I routinely accept site-originated cookies, because these are critical to managing per site user prefs. Additionally, cookies don't have any power that I can't manage myself (so there's much less to be paranoid about in the first place). I can see the full text of the cookie, and the cookie mechanism can only send back the cookie as I received it. This Javascript bug tells Google whatever Javascript can learn about my system. It is significantly more invasive than cookies.
And you can't build this sort of page view history on a per user basis yourself using a cookie and a tracking database? This isn't exactly rocket science here. It's just another chapter in Google's "pretty soon we'll have one of everything" approach to doing business on the web. And it will all be tied into making you either want to sign up to put ads on your site, or getting you to view pages that have ads on them.
No kidding. Let me know when they go VHS. I'm not interested in buying any videos in Beta.
[/curmudgeon]
Definitely agree. I guess I missed the part where that piece of the policy was only for IM.
Prints are not a function of their original capture method. You can get prints of digital images that are at least as archival in quality as an photomat 35mm 4x6 glossy.
As an added bonus, making a print of a digital image does not expose your film media to any danger. Even cartridge formats like 110 and APS carry some risk to the negative in terms of dust, scratching, etc, when making a print. Not to mention that while film has improved markedly over the years, it is still a highly unstable base on which to store an image.
How about if motorists stop killing 40k+ people in the USA every year "accidentally"?
This seems overly broad. How do you automate internal file transfers with a policy like this? Do you have no operational systems that need to provide data extracts to analysis systems or the like? Or do you allow automated transfer in documented and approved situations?
Actually, given MS' scope and resources I fully expect them to have a staff whose regular duties consist solely of fixing these types of problems.
For example: I have a significant chunk of the Cannibal Corpse discography in MP3 that I "bought" from emusic.com... can I sell it? And how much worse does this question get if the file is somehow DRMed? How would I sell it if the DRM won't let anyone else play it? And if I can't sell it, can my daughter inherit it tax free, rather than having to pay taxes on something that she probably can't sell either? Will she even be able to inherit it?
The fact that I can even ask these questions indicates to me that the whole model of "intellectual property" is fundamentally broken and bound to cause more problems than it solves at this point. And if Google is going to attempt to further the cause of "intellectual property" by selling DRMed products, then I think this goes against their whole "don't be evil" motto.
Yes, I can usually see into the future and anticipate each and every potential need. Cripes. If that's the level this discussion is at, then I'm done.
One might. Except that what we're discussing here is not administering large, multi-user servers, rather desktops. And I have plenty of experience following the recommendations to use multiple partitions on desktops, and all of it is negative. In eveery case where I've done that I've ended up regretting it because one partition or another was full long before the others, requiring a nasty mess of symlinks to keep the system functional without wiping the drive entirely. These recommendations to use multiple partitions are just what I said: cargo cult systems administration. Desktop Linux users looking at how admins set up servers and just doing the same, without really thinking about why or how.
Yes, there are times when using partitions as a poor-man's quota system (like your web server example) makes some sense. It's a good point. But again, my main objection to multiple partitioning isn't that it shouldn't be done, but rather that it's recommended all the time by people who haven't really thought about the cases where it actually makes sense.
Well, I might be interested in buying an Apple notebook, except that I can guarantee that I'd blow away OS X to install Linux. Given that I have no intention of ever using OS X why would I want to buy a machine which is premium priced precisely because it includes OS X?
I think I'll stick to buying OS-free used laptops (usually corporate castoffs) and installing an operating system I know I'll like on there.
I don't even buy the idea of having swap space in its own partition. What if you need more swap? Then you have to create a swap file on an existing partition and manage two separate swap files in the future. Ugh.
As for reinstalling, generally the only time I'm reinstalling the whole OS is when I am switching distros or when I feel like starting fresh on new, larger HD. Which I do this a lot less frequently than having to create a bunch of annoying symlinks across partitions to help alleviate problems due to running out of space in some partitions. Even if I'm just swapping a drive and not trying to reinstall, having data in a separate partition doesn't make the drive swap any easier, you still have to copy the whole thing.
My opening sentence was too vague--maybe even a full blown misstatement. I should have said "Public domain works can be used to create copyrighted works" and then focused on how this is not true for public property. My point wasn't that the public domain work is no longer in the public domain, but that the originaly analogy (that public domain is like public property) is poor because privatizing public property does mean the property is no longer public.