Oh come on, that's baloney. The rules for using the apostrophe are quite logical.
"it's" is a contraction of "it is". The ' means "something was taken out and the phrase was shortened". If you can logically replace the word in question with "it is" then you want the apostrophe. Same with "you're" versus "your". If you can use "you are" and the sentence still makes sense then you want "you're", otherwise "your".
"Customer's" is the posessive of one customer. "Customers'" is the posessive of many customers. If you're talking about one customer then you want the first, if you mean two or more customers you want the latter. You can think of everything to the left of the apostrophe as being the original word, and everything to the right being something added to show that it's a possessive.
How is any of that not logical? And yet the article submitter apparently did not have even the slightest notion of considering the proper usage. If someone posted code on the front page that didn't compile because it was grammatically invalid, you'd damn well hear about it from angry readers.
This file is 12.0 MB and contains only 1:10 (70 seconds) of footage. The video stream is 512x384 DivX 5 (DX50 fourcc) at 25 FPS, with an average bitrate of 738kbit/s. That's pretty respectable. However the audio track is UNCOMPRESSED PCM at 22kHz stereo, which comes works out to a bitrate of 706kbit/s. That is just insane - approximately half of this file is UNCOMPRESSED AUDIO.
Had the person used a decent codec, such as VBR mp3 at 128kbit/s average rate, the file size would have shrunk to about 7.3 MB. Or in other words this file is approximately 70% larger than it should be. I don't know who's running their web site but I'm sure they'd appreciate the bandwidth savings if their trailer wasn't made by someone lacking clue.
Just make him save EVERYTHING under "My Documents" or somewhere under "Documents and Settings". Everything. Tell him his computer will break if he doesn't store data files there. Most programs these days default to that directory for the "Save As" and "Open" dialogs, so that's a plus. Then just back that up in whole regularly (make sure to purge the "Temporary files" folder and browser cache first, or exclude them, as you don't want to bother backing that up as well.)
It's not perfect. Most program settings are either stored in a config file in "Program Files" or in teh registry. But it's really hard to back up program settings reliably and you will probably end up reinstalling anyway. So just focus on the data files, i.e. the things that you actually load and save inside the program, not its internal settings and configuration options.
So one tard somewhere can't uninstall a SP. People find all sorts of ways to mess up Windows, that doesn't mean that you can't uninstall a SP. I've successfully uninstalled service packs myself many times.
Did you even read the link I posted? Each of the "issues fixed in SP2" has a link to a KB article with a corresponding hotfix that you can download. All of the security issues (MSxx-yyy) have an individual download that you can install, without installing all of XP SP2. Some of them are marked as "cumulative update for Internet Explorer" but that's not the same as installing a service pack. It just means you're getting all the IE fixes up to that point. Go read the link and search in the page for "internet explorer" and you'll see that you can patch ala carte.
What I find most impressive...... is how well your site's holding up under the slashdotting!
That's probably because it's very simple static HTML and static images, with a moderate number of images per page. What people seem to forget is that regular Apache combined with the provisions in http for caching of static content make for extraordinarily robust servers. When you get into trouble is when you leave that behind and start to make pages dynamically, with database backends and scripting languages like PHP, Perl, ASP, etc. Most sites that die a hideous death from a slashdotting are usually one of the following:
- some idiot posted 100 full size images to a single page - some idiot posted unscaled images from his digital camera (i.e. hundreds of kb or even several megabytes each) - any site that has dynamic content (user accounts, logins, page customizations, content generated from database backend, ad rotations, etc) - some page on a freebie hosting service not intended for real hosting
Static HTML and images, with a sane page design (not "let's put it all on one page") scales very nicely and will almost always do well to a flash-crowd, providing it's on a real server and the metered bandwidth does not run out.
Why is this still rated highly? It's completely wrong. Mach is a dimensionless ratio of current speed to LOCAL speed of sound. It's not an absolute speed.
That's a non sequitor. If you're going to use commercial software you have to be prepared to pay for upgrades every now and then. If that bothers you then use free software. You can't have your cake and eat it too. It's the user's fault if they won't or can't run the latest version. And besides, XP SP2 is a free upgrade if you have any version of XP currently.
By that line of reasoning, using some antiquated version of windows 98 with truckloads of bugs and security holes is somehow microsoft's fault, even though they fixed most of them years and years ago. Sorry, that's just specious reasoning.
Just playing devil's advocate here, but if there was a security vulnerabilty in an open-source project which affected older versions of the software -- but not the current released/stable version -- then this would be a non-story. "Foo v1.25 has a vulnerability? Well it's the user's fault for not running v1.30 which fixed that bug." But it's Microsoft, so somehow all the laws of software are different....
I always feel like the record companies fooled us when CD's came out. Sure, they sounded great, but when CD's were new, vinyl LP's were now $7, and the CD was $14-16, with the excuse that "we are capacity constrained, when we get more capacity, prices will be much cheaper, because these things are cheap to make".
US$7 in 1983 (the year the compact disc was introduced) is equivalent to $12.73 in 2003 after accounting for inflation. The average retail CD price in the first quarter of 2004 was $13.29. Seems like CDs these days are selling for about the same as your vinyl LPs back in the day, so that line of reasoning really doesn't go very far. You can't compare monetary amounts spanning two decades without accounting for inflation.
When you see a regular expression like that it's a good indicator that the person that wrote it wasn't very familiar with how to write good REs. The above suffers from "leaning toothpick syndrome." If you are trying to match the '/' character, then don't use it as the delimiter of the RE. For example, compare the following REs, which are equivalent:
m/\/\/(.*)\/\//i
and
m,//(.*)//,i
Using ',' for the delimiter of the RE means you don't have to backslash-quote the forward slash to use it in a match.
This is just basic RE stuff and has nothing to do with perl. In fact, perl gives you many tools such as the "/x" modifier that allow you to vastly clarify the meaning of long and complicated REs, to the point of having indentation, extra whitespace, and even embedded comments. It's not perl's fault that many people only have a cursory knowledge of REs, and so they tend to write terrible REs.
Wrong. There is no such thing as a negotiation port in BitTorrent. The clients connect directly to each other without any prior intervention.
The way it works is that the client contacts the tracker, and tells it what port it's listenting on. That information (along with the client's IP address) is added to the tracker's peer list. When reporting to the tracker the client gets part or all of this peer list, resulting in a list of IP addresses and ports that it will attempt to contact.
All of this happens on arbitrary ports. The tracker can be on any port (based on the announce URL in the.torrent file) and peers can listen on any port or ports. There is no "negotiation port" like with FTP involved anywhere.
The only accurate way to classify BT traffic is packet inspection at layer 4 or higher, to look for the client handshake.
I read an article once by someone familiar with security/fraud divisions of large banks and the like. (Sorry, I can't seem to find a link or reference at the moment.)
But essentially the gist was they need the fraud to occur in order to do anything about it. Their entire machinery for dealing with fraud requires that it has already happened, i.e. the money has changed hands. They start to act after the fraudulent transactions have been processed. Apparently they need that amount of concrete evidence in order to investigate further and work with law enforcement. So, when you report "hey, this website is trying to phish your customers" you are naturally going to get the silent treatment, a virtual shrug, because they don't give a crap until the fraud has taken place. Basically they have decided to be entirely retroactive and fight the fraud from the perspective of "mopping up after it's happened" and are entirely uninterested in trying to investigate web sites, domain name registrations, open proxies, etc. If you report a phishing scam, they are going to need records of the bad person logging in to your account you and removing your money -- they just aren't interested in looking into it until that point.
If your data is based on port number then it's wildly underestimating BitTorrent and likely useless for anything meaningful.
Most clients these days let you choose an arbitrary or random port. Some trackers require you to change your port to a nonstandard one. Open up your details page on your client[1] some time (or netstat, or TCPView) and look at the remote ports for the peer connections. You'll see they're all over the place, because each individual peer gets to decide which port its client is going to listen on. The standard allows for any port to be used.
Thus to accurately assess the BitTorrent traffic you have to look at the content of the packets, not the port number. There are devices to do this, but it takes significantly more CPU/memory resources compared to just classifying traffic by port.
[1] The crappy mainstream client aside -- but nobody uses it as it has next to zero features compared to the rest like Azureus, ABC, or BitTornado.
No, there are no IP addresses in the torrent. THe client contacts the tracker and gets a list of peers from there. The only thing that's in the.torrent that contains anything resembling an address is the announce url, which is how the client contacts the tracker. You could make a case for the fact that the announce URL could contain raw IP addresses but it usually contains a domain name. In any case, it's not the IP address where the client downloads any of the actual file data, which is what the parent post was implying. It would be next to useless to store peer lists in the.torrent file because the swarm changes often.
Note: you might look at.torrent files that have been molested by Azureus and say that there are indeed peer lists in there. That is Azureus' fast-resume data, so that it has a peer list ready when you resume instead of having to first contact the tracker. That is a local modification of the.torrent file done by Azureus after you started the transfer. It is not distributed with the.torrent file that you download. Azureus is just using the file to store some state, it's an undocumented nonstandard thing that this particular client does. It doesn't change the fact that peer addresses are not distributed in the.torrent.
Think about a particularly well known deathmatch level from your favorite FPS shooter game -- if the place were real, could you find your way around it quickly for the "first" time? Would you know where to hide? Be able to pick out good sniper holes?
Oh great, now you've got me thinking about mp_dawnville and wondering what it'd be like in real life.:-P
And yes, I think I would know of all the good spots to hang out with a kar98k.
It's only dependency hell if you try to install RPMs by hand. Use yum and all you have to do is type "yum install whatever" and you get all the dependencies for 'whatever' automatically downloaded and installed in the proper order. You can even have yum upgrade across major versions for you, e.g. 9 to 10 and so on.
If you find yourself installing from source just to get headers then you're not using your package management system correctly. You were probably just missing the right -devel RPM as the other poster mentioned.
"RPM dependency hell" is nothing more that an unfortunate but preventable outcome of people not knowing what they're doing and installing RPMs by hand.
Configure yum.conf with your closest redhat/fedora/mandrake mirrors, and then try "yum install foobar" for the exact same effect. This is hardly something unique to Debian.
Signed, A Debian user who isn't afraid to admit that automatic dependency resolution should be taken for granted in all the major distros.
You're making a specious argument though. Your point assumes that the number of people using Firefox in no way affects the amount of funding that they get.
If you spend $1000 to convince 1000 people to use Firefox, then isn't it possible that some percentage of those thankful people might contribute via paypal to the Mozilla foundation? Wouldn't some of them perhaps do bug triage or support in forums and free up time for developers to do more important things? What if seeing the increased user-base convinces a company that firefox is ready for prime time and they decide to alter their webpages to work with standards? Or they start supporting FF internally... or they donate some of their programmers' time to the project? Or they donate.
The point is that giving money directly to the developers is nice, but so is increasing the size of the community, which can result in similar side effects: more money (donations) or time (volunteers) for the developers. It's not like you throw away dollars by spending them on advertising.
Put differently, just because Mozilla isn't about making a profit doesn't mean they won't benefit monetarily from a larger user-base or market share.
That's exactly why if you leave your upload rate set to unlimited or uncapped (and you have an asymmetric connection, as most of us do) you will kill your download speed, as the ACKs start to drop. Duh. Problem solved, it's far from the end of the world. Just cap your upload to about 80% of your max. Use netlimter if you want, but I think it's a very poor quality program. Most p2p apps let you do this limiting in their settings which is better.
Slashdot Mirror
Oh come on, that's baloney. The rules for using the apostrophe are quite logical.
"it's" is a contraction of "it is". The ' means "something was taken out and the phrase was shortened". If you can logically replace the word in question with "it is" then you want the apostrophe. Same with "you're" versus "your". If you can use "you are" and the sentence still makes sense then you want "you're", otherwise "your".
"Customer's" is the posessive of one customer. "Customers'" is the posessive of many customers. If you're talking about one customer then you want the first, if you mean two or more customers you want the latter. You can think of everything to the left of the apostrophe as being the original word, and everything to the right being something added to show that it's a possessive.
How is any of that not logical? And yet the article submitter apparently did not have even the slightest notion of considering the proper usage. If someone posted code on the front page that didn't compile because it was grammatically invalid, you'd damn well hear about it from angry readers.
Here is a jpeg version (718x1151) of the ad, converted from the above PDF in the parent's post.
On their site is a link to a trailer clip of the game/mod.
This file is 12.0 MB and contains only 1:10 (70 seconds) of footage. The video stream is 512x384 DivX 5 (DX50 fourcc) at 25 FPS, with an average bitrate of 738kbit/s. That's pretty respectable. However the audio track is UNCOMPRESSED PCM at 22kHz stereo, which comes works out to a bitrate of 706kbit/s. That is just insane - approximately half of this file is UNCOMPRESSED AUDIO.
Had the person used a decent codec, such as VBR mp3 at 128kbit/s average rate, the file size would have shrunk to about 7.3 MB. Or in other words this file is approximately 70% larger than it should be. I don't know who's running their web site but I'm sure they'd appreciate the bandwidth savings if their trailer wasn't made by someone lacking clue.
(assuming you're referring to Windows)
Just make him save EVERYTHING under "My Documents" or somewhere under "Documents and Settings". Everything. Tell him his computer will break if he doesn't store data files there. Most programs these days default to that directory for the "Save As" and "Open" dialogs, so that's a plus. Then just back that up in whole regularly (make sure to purge the "Temporary files" folder and browser cache first, or exclude them, as you don't want to bother backing that up as well.)
It's not perfect. Most program settings are either stored in a config file in "Program Files" or in teh registry. But it's really hard to back up program settings reliably and you will probably end up reinstalling anyway. So just focus on the data files, i.e. the things that you actually load and save inside the program, not its internal settings and configuration options.
So one tard somewhere can't uninstall a SP. People find all sorts of ways to mess up Windows, that doesn't mean that you can't uninstall a SP. I've successfully uninstalled service packs myself many times.
Did you even read the link I posted? Each of the "issues fixed in SP2" has a link to a KB article with a corresponding hotfix that you can download. All of the security issues (MSxx-yyy) have an individual download that you can install, without installing all of XP SP2. Some of them are marked as "cumulative update for Internet Explorer" but that's not the same as installing a service pack. It just means you're getting all the IE fixes up to that point. Go read the link and search in the page for "internet explorer" and you'll see that you can patch ala carte.
You can uninstall a service pack. You can install the individual fixes included in the SP. Please get a clue.
Way to defend yourself as AC, dickcheese.
That's probably because it's very simple static HTML and static images, with a moderate number of images per page. What people seem to forget is that regular Apache combined with the provisions in http for caching of static content make for extraordinarily robust servers. When you get into trouble is when you leave that behind and start to make pages dynamically, with database backends and scripting languages like PHP, Perl, ASP, etc. Most sites that die a hideous death from a slashdotting are usually one of the following:
- some idiot posted 100 full size images to a single page
- some idiot posted unscaled images from his digital camera (i.e. hundreds of kb or even several megabytes each)
- any site that has dynamic content (user accounts, logins, page customizations, content generated from database backend, ad rotations, etc)
- some page on a freebie hosting service not intended for real hosting
Static HTML and images, with a sane page design (not "let's put it all on one page") scales very nicely and will almost always do well to a flash-crowd, providing it's on a real server and the metered bandwidth does not run out.
Why is this still rated highly? It's completely wrong. Mach is a dimensionless ratio of current speed to LOCAL speed of sound. It's not an absolute speed.
Come on moderators.
That's a non sequitor. If you're going to use commercial software you have to be prepared to pay for upgrades every now and then. If that bothers you then use free software. You can't have your cake and eat it too. It's the user's fault if they won't or can't run the latest version. And besides, XP SP2 is a free upgrade if you have any version of XP currently.
By that line of reasoning, using some antiquated version of windows 98 with truckloads of bugs and security holes is somehow microsoft's fault, even though they fixed most of them years and years ago. Sorry, that's just specious reasoning.
Just playing devil's advocate here, but if there was a security vulnerabilty in an open-source project which affected older versions of the software -- but not the current released/stable version -- then this would be a non-story. "Foo v1.25 has a vulnerability? Well it's the user's fault for not running v1.30 which fixed that bug." But it's Microsoft, so somehow all the laws of software are different....
US$7 in 1983 (the year the compact disc was introduced) is equivalent to $12.73 in 2003 after accounting for inflation. The average retail CD price in the first quarter of 2004 was $13.29. Seems like CDs these days are selling for about the same as your vinyl LPs back in the day, so that line of reasoning really doesn't go very far. You can't compare monetary amounts spanning two decades without accounting for inflation.
This is just basic RE stuff and has nothing to do with perl. In fact, perl gives you many tools such as the "/x" modifier that allow you to vastly clarify the meaning of long and complicated REs, to the point of having indentation, extra whitespace, and even embedded comments. It's not perl's fault that many people only have a cursory knowledge of REs, and so they tend to write terrible REs.
Wrong. There is no such thing as a negotiation port in BitTorrent. The clients connect directly to each other without any prior intervention.
.torrent file) and peers can listen on any port or ports. There is no "negotiation port" like with FTP involved anywhere.
The way it works is that the client contacts the tracker, and tells it what port it's listenting on. That information (along with the client's IP address) is added to the tracker's peer list. When reporting to the tracker the client gets part or all of this peer list, resulting in a list of IP addresses and ports that it will attempt to contact.
All of this happens on arbitrary ports. The tracker can be on any port (based on the announce URL in the
The only accurate way to classify BT traffic is packet inspection at layer 4 or higher, to look for the client handshake.
I read an article once by someone familiar with security/fraud divisions of large banks and the like. (Sorry, I can't seem to find a link or reference at the moment.)
But essentially the gist was they need the fraud to occur in order to do anything about it. Their entire machinery for dealing with fraud requires that it has already happened, i.e. the money has changed hands. They start to act after the fraudulent transactions have been processed. Apparently they need that amount of concrete evidence in order to investigate further and work with law enforcement. So, when you report "hey, this website is trying to phish your customers" you are naturally going to get the silent treatment, a virtual shrug, because they don't give a crap until the fraud has taken place. Basically they have decided to be entirely retroactive and fight the fraud from the perspective of "mopping up after it's happened" and are entirely uninterested in trying to investigate web sites, domain name registrations, open proxies, etc. If you report a phishing scam, they are going to need records of the bad person logging in to your account you and removing your money -- they just aren't interested in looking into it until that point.
If your data is based on port number then it's wildly underestimating BitTorrent and likely useless for anything meaningful.
Most clients these days let you choose an arbitrary or random port. Some trackers require you to change your port to a nonstandard one. Open up your details page on your client[1] some time (or netstat, or TCPView) and look at the remote ports for the peer connections. You'll see they're all over the place, because each individual peer gets to decide which port its client is going to listen on. The standard allows for any port to be used.
Thus to accurately assess the BitTorrent traffic you have to look at the content of the packets, not the port number. There are devices to do this, but it takes significantly more CPU/memory resources compared to just classifying traffic by port.
[1] The crappy mainstream client aside -- but nobody uses it as it has next to zero features compared to the rest like Azureus, ABC, or BitTornado.
No, there are no IP addresses in the torrent. THe client contacts the tracker and gets a list of peers from there. The only thing that's in the .torrent that contains anything resembling an address is the announce url, which is how the client contacts the tracker. You could make a case for the fact that the announce URL could contain raw IP addresses but it usually contains a domain name. In any case, it's not the IP address where the client downloads any of the actual file data, which is what the parent post was implying. It would be next to useless to store peer lists in the .torrent file because the swarm changes often.
.torrent files that have been molested by Azureus and say that there are indeed peer lists in there. That is Azureus' fast-resume data, so that it has a peer list ready when you resume instead of having to first contact the tracker. That is a local modification of the .torrent file done by Azureus after you started the transfer. It is not distributed with the .torrent file that you download. Azureus is just using the file to store some state, it's an undocumented nonstandard thing that this particular client does. It doesn't change the fact that peer addresses are not distributed in the .torrent.
Note: you might look at
Humm.. interesting.. now I'm wondering what media player it was.. BSPlayer? (just a guess)
Good for you for not letting their BS get past you.
And yes, I think I would know of all the good spots to hang out with a kar98k.
It's only dependency hell if you try to install RPMs by hand. Use yum and all you have to do is type "yum install whatever" and you get all the dependencies for 'whatever' automatically downloaded and installed in the proper order. You can even have yum upgrade across major versions for you, e.g. 9 to 10 and so on.
If you find yourself installing from source just to get headers then you're not using your package management system correctly. You were probably just missing the right -devel RPM as the other poster mentioned.
"RPM dependency hell" is nothing more that an unfortunate but preventable outcome of people not knowing what they're doing and installing RPMs by hand.
for those too retarded to search for themselves here are some:
nforce: site feed selection page
#bt on EFNet: site feed
#tvtorrents on EFNet: site feed
youceff: site feed selection page
torrentbits: site feed
torrentmind: site feed (there are other feeds for individual categories)
FUD indeed.
Configure yum.conf with your closest redhat/fedora/mandrake mirrors, and then try "yum install foobar" for the exact same effect. This is hardly something unique to Debian.
Signed,
A Debian user who isn't afraid to admit that automatic dependency resolution should be taken for granted in all the major distros.
You're making a specious argument though. Your point assumes that the number of people using Firefox in no way affects the amount of funding that they get.
If you spend $1000 to convince 1000 people to use Firefox, then isn't it possible that some percentage of those thankful people might contribute via paypal to the Mozilla foundation? Wouldn't some of them perhaps do bug triage or support in forums and free up time for developers to do more important things? What if seeing the increased user-base convinces a company that firefox is ready for prime time and they decide to alter their webpages to work with standards? Or they start supporting FF internally... or they donate some of their programmers' time to the project? Or they donate.
The point is that giving money directly to the developers is nice, but so is increasing the size of the community, which can result in similar side effects: more money (donations) or time (volunteers) for the developers. It's not like you throw away dollars by spending them on advertising.
Put differently, just because Mozilla isn't about making a profit doesn't mean they won't benefit monetarily from a larger user-base or market share.
Not so much a transcript, but a translation of the entire tape: http://www.cnn.com/2004/WORLD/meast/10/29/bin.lade n.transcript/
That's exactly why if you leave your upload rate set to unlimited or uncapped (and you have an asymmetric connection, as most of us do) you will kill your download speed, as the ACKs start to drop. Duh. Problem solved, it's far from the end of the world. Just cap your upload to about 80% of your max. Use netlimter if you want, but I think it's a very poor quality program. Most p2p apps let you do this limiting in their settings which is better.