It doesn't help to deny changing the status bar text. The way google manages this is by rewriting the link on a mousedown event. So, it starts out going to the proper place, but when you click or right-click it is re-written to go to the redirect link. Ad links are a bit different in that the container of the ad prevents the status bar from changing by overwriting the normal mouseover event.
Check out any search link on Google. Mouse over. See the text? Now right click on the link. See the new redirection status text (in firefox only, IE will still show the normal link)? This can be done with any link using the proper javascript.
It is actually quite clever scripting. One advantage is that without javascript you still get the proper search results.
You can't spin-down any hard disks that is mounted. File systems are too advanced these days. Whether journaling or not, they all write fs metadata every few seconds.
I fixed this problem by using special options when mounting (noatime) and reconfiguring several system utilities (cron, syslogd). The filesystem doesn't just magically need to write/read data every few seconds (there may be a few that mark checkpoints regardless of writes but I've never seen one in action). In linux you can also use laptop_mode to optimize fs writes for battery life at the expense of a potential to loose some maximum amount of time worth of data (usually 5min or so).
Another interesting option is to use solid-state storage for some disk locations (/var/log,/tmp). You may still have to tweak some running programs not to query the disk so often. Many laptops these days have built in memory card readers and a 2GiB card is less than $50.
I do agree though that most of the power is not going to be saved by spinning down the disk. Laptop drives only use a few watts of power compared to the cpu and display.
The cable company in my area (Charter) sells fiber links for a reasonable price. Then AFAIK they also don't bother big uploaders in the residential side... yet.
1. All other commercially available operating systems use flat files to store configuration information. And almost every other operating system out there works better than Windows in a variety of ways, not least of which being performance.
Hmm... OSX would be one to disagree with you there (netinfo is not in flat files and neither is LDAP). OSX does have flat files on the BSD end but they aren't the ones you configure with the gui. The os that used netinfo before OSX was NeXTSTEP and everyone thought it was a bad idea then too. There is also the gconf database that Gnome uses, while similar to the registry in function it uses the file system directories for hierarchy and xml files for storage so it is easy to edit by hand if you need to.
3. People who run operating systems that use flat files tend to READ those flat files. The registry, on the other hand, is so huge and byzantine (again, WHY???) that finding entries in it is like going on a fishing expedition. Nobody really knows what's in their registry. I believe this is by design, not by accident.
Not to knock flat files (I prefer them myself) but one could say the same about finding them. The registry was designed to be a structured database of configuration data, Microsoft just screwed things up badly with it.
4. The registry is IN FACT used to make piracy difficult. Virtually every piece of commercial Windows software stores registration information in the registry, usually in literally dozens of different locations so that to clear out a botched install you have to use a search tool and guess at all the possible names the company may have used for its keys. First, do you think Microsoft isn't doing the same thing??? Second, do you think this isn't by design???
Umm... no. If you look at Microsoft's reference documentation on how companies should and should not use the windows registry I think you will note that the locations where things *should* be stored are quite well defined. Microsoft themselves may have issues with using there own standards but I have never found anti-piracy to be one of their reasons for being incompetent.
5. When a hacker creates a Word Macro Virus and the cops catch him like, a week later, how do you think that happens? Word, installed, puts serial number information in the registry and later, into documents. Again, by design.
Now you are starting to sound paranoid. Can you point out the place in a word document where a serial number is kept? Give me a link or I call BS. I know it is stored in the registry (Where else do you think it would be?) but so are several other bits of license related data. HINT: it is not called an install code, it's called a license key. If you removed the key from the registry it stops working and asks for a key when you start it up.
6. When they spent millions of dollars building Windows 95 and created long filename support, do you think it was by mistake that they just happened to leave long filename support out of their new version of DOS? Or that you couldn't boot to a command prompt that had long filename support?...
IIRC this feature was included in the version of DOS that shipped with Windows 98 so that must not have been their reasoning otherwise they would have left it out.
IF YOU ARE CORRECT, you must have a reasonable justification for the use of the registry that is credibly better than using a flat-file approach. I bet you don't have one.:)
Hmm... I can think of 2 design decisions why a databased approach to configuration has advantages.
Structured data types in the configuration. (Strings are strings, numbers are numbers, and hey, you can store binary data as well.)
The only part of the kernel that needs assembly is the parts that interact directly with hardware (small parts of driver code). Everything else can (and is) written in higher level languages.
And that was argued in a copyright case. The claim was that copying a program to the memory on your computer could count as copyright violation. The courts ruled that copying a program to memory or to a hard disk was not copyright infringement when such use was necessary to the operation of the program.
You are over complicating the opt-out procedure. A usual opt-out request is done once and stored or attached to the data. You don't have to tell everyone who searches the web that you opt-out, you don't even have to tell every search engine, just opt-out in a robots.txt file or in meta headers.
Actually it has more to do with DNS, the lists based on static/dynamic have several methods of checking on addresses before they are listed as dynamic most of them are based on DNS lookups. If your address has no RDNS it is likely going to be listed. If your address has an RDNS of dhcp-ip-ip-ip-ip.yourisp or ip-ip-ip-ip.yourisp it will be listed. An RDNS of static-ip-ip-ip-ip.yourisp will not be listed, usually anyway. Other non-address based RDNS also usually lands you in the clear. A list can't check to see if you are on DHCP or not unless your ISP publishes that kind of information but the RDNS name gives most people a clue about what kind of IP they are dealing with.
NiMH batteries AFAIK are trickle charged when full. This means that power is ALWAYS flowing into the battery but at such a low rate that 100% of it is wasted to the internal resistance of a fully charged battery. It's how a charging circuit knows when it battery is charged. Also note that NiMH batteries loose charge even when not under a load at a slow rate (30%/month). If you take the battery out when you don't need it and you don't need it for a month at a time then it will be (mostly) discharged when you do need it.
LI batteries have no natural discharge rate so they are not charged up to full unless their charge drops below a threshold. The loss rate of the charging/monitoring circuit is only 5%/month. A good reason to remove LI batteries when not in use is lifespan rather than safety. LI batteries last longer at lower temperatures and at partial charges. Recommended storage is at 40% charge and 0C (32F).
We had a lab full of PCs that would lockup if jolted the wrong way while upright. To get them to unlock they would have to be powered down and dropped on their side from about 8 inches up. It turned out that the motherboards had bad memory slots and the memory contacts would loose connection. I eventually tried several techniques to keep the memory in place (solder, silicone, hot glue). One of the best methods was to use solder to tin the contacts on one side of the memory, took me about an hour per DIMM to perform the procedure so I never got around to more than just the worst computers in the bunch.
Ah, but they could be advertising with no delivery. An analogy to this would be putting an add in the newspaper that read, "Contact me and we will swap photocopies of our books." and suing everyone who contacted you wanting to trade you a copy of your book.
Putting your book in the public domain? No. Entrapment? Yes.
Some learning uses are simply by-products of normal use and can be treated as normal use; other techniques such as reverse engineering are generally outlawed one way or another.
Well, I think you are confusing law with contract (EULA). Reverse engineering is perfectly acceptable unless you have agreed to a contract that forbids it. This is why many software titles in recent publication requires you to agree to a EULA before you USE the software (and when you install it).
All that aside, your argument about worth and cost of physical items is flawed. Some books cost much more than they are worth as an object. I've had books that cost $300 for school. You do not need an EULA to protect your software, you only need it to limit liability and/or limit your customers use of said software. As a contract they are judged on a case by case basis as to enforceability and if the user agreed to them.
Mine is a set of 5x250gig 7200 rpm SATA drives. I've rebuilt the raid three times now, twice from within the OS (linux) and once from the RAID configuration BIOS. The BIOS configuration seems to take forever to rebuild (more than 24 hours) but from within the OS it only takes about 2 hours. All on a P-III 600.
I have another hardware RAID 5 that is SCSI (6 x 72 gig @ 10k rpm), it rebuilds in about 45 min.
The only software RAID that I have done was using much smaller drives (6 gig) and never needed to be rebuilt.
Indicate that rebuild times in software are not bad on a 1ghz system. We can extrapolate out that a rebuild of 4x250 gig drives should take less than 2 hours (calculating parody on 3 rather than 2 drives for 2x as much data). Unless you are using the drives or have problems with a controller, then you might have more problems.
I have a 5 disk array using a highpoint raid controller, it rebuilds in about 2 hours when the computer is booted. It takes a full day to rebuild from the BIOS however.
Just so you know there is a quick fix. Replace ever instance of <embed src="stuff"> with <script src="embedStuff.js"></script> and have embedStuff.js be document.write('<embed src="stuff">');
Simpe to do with a python/perl/shell script...:)
Hmm... I guess it depends on what you do. From experience I would guess that IT where you work is implementing policy directed to them by management and implementing infrastructure to make their life easier. I've had management direct me to implement some seriously silly policies, usually I have a discussion with my manager about them and we work out a compromise, sometimes I'll implement them and direct comments about them directed to the management that required them.
It's been my experience that a small % of the user base can be trusted to know what all they need, how it works and how to maintain it. Many of those who THINK they can do it on their own end up screwing up more than just their own systems. Most of what we have to deal with in this is software related.
I like the vast majority of my users, I consider myself lucky in this. The few users that I have that cause problems (from my standpoint) are either because: They know nothing about technology and have no desire to learn (I've had users refuse training because they "didn't have the time" and ask me to drive over to show them how to operate basic software) or they think they know how things should be done and have budget authority for their department (think users buying their own wireless access points).
If your input is disregarded without a second thought or not even sought after, I feel your pain, it happens to me all the time. I make it a point to get my user's opinions on technical matters that they are knowledgeable about, but sometimes there are other needs that outweigh their advice. Part of my job is to research the BEST solution to problems as a whole.
A man-in-the-middle attack can still be performed in this approach. Z has broken the wires and is now acting as an intermediary between signals. Z monitors both wires for disconnects and then disconnects the other end when that happens. Z then knows the state of both ends at any given time so Z knows which sequence each has chosen.
At least Blizzard doesn't charge you for accessing customer support where I think it is EXTREMELY SHOCKING that in North America, phone companies charge a monhtly fee for access to 911, whatever the fee is.
I think you are confused. The 911 charge is not to access 911 services it is to recover some of the cost of supporting 911 services on all phones. The phone company still pays for cell phones with discontinued service and pay phones access to 911.
Wouldn't it be odd if I went to a pay phone and dialed 911 only to hear "please insert 75 cents to continue."
AOL's big mistake was resting on their laurels. They were once pioneers of the internet but somewhere along the line they came up with some terrible problems that they continued to ignore. Now they are playing catch-up with the other big players on the net and doing a terrible job of it. They want to be like everyotherjoe on the net and have doomed themselves to a loosing battle.
The problem with this is that most corperations earn next to nothing. They pay no taxes because they have ~$0 net income. The money goes back to investors, pays employees, funds research, etc.
The main problem is defining what is earnings. If you say grose income is earnings then those in low-margin/high-volume markets can earn money and go out of business at the same time.
If you say net income is earnings then most corperations make nothing.
Google is considered an ISP for the sake of the services they offer. That said here is a quote from 17 U.S.C. 512(d)
"by reason of the provider referring or linking users to an online location containing infringing material or infringing activity, by using information location tools, including a directory, index, reference, pointer, or hypertext link, if the service provider . ..
(3) upon notification of claimed infringement . . . responds expeditiously to remove, or disable access to, the material that is claimed to be infringing or to be the subject of infringing activity, except for purposes of this paragraph . . . the information [for notification] shall be identification of the reference or link, to material or activity claimed to be infringing, that is to be removed or access to which is to be disabled, and information reasonably sufficient to permit the service provider to locate that reference or link."
So google is not liable for the service provided as long as they respond to a request to remove content.
IIRC the case you are referring to was a case of trademark infringement not of copyright. In that case it was ruled that it WAS infringement of Nintendo's trademark for the game to run the initialization sequence on the NES that caused it to display the Nintendo logo.
I had an aunt with the same problem. Only the first book was put on the publishers back shelf (no marketing.) Now she has two more in the series but she can't get them published because the first didn't sell enough copies.
Just thought I would point out that it does not accept regex. It accpets wildcards which are quite different and limited compared to regular expressions.
Slashdot Mirror
It doesn't help to deny changing the status bar text. The way google manages this is by rewriting the link on a mousedown event. So, it starts out going to the proper place, but when you click or right-click it is re-written to go to the redirect link. Ad links are a bit different in that the container of the ad prevents the status bar from changing by overwriting the normal mouseover event.
Check out any search link on Google. Mouse over. See the text? Now right click on the link. See the new redirection status text (in firefox only, IE will still show the normal link)? This can be done with any link using the proper javascript.
It is actually quite clever scripting. One advantage is that without javascript you still get the proper search results.
I fixed this problem by using special options when mounting (noatime) and reconfiguring several system utilities (cron, syslogd). The filesystem doesn't just magically need to write/read data every few seconds (there may be a few that mark checkpoints regardless of writes but I've never seen one in action). In linux you can also use laptop_mode to optimize fs writes for battery life at the expense of a potential to loose some maximum amount of time worth of data (usually 5min or so).
Another interesting option is to use solid-state storage for some disk locations (/var/log, /tmp). You may still have to tweak some running programs not to query the disk so often. Many laptops these days have built in memory card readers and a 2GiB card is less than $50.
I do agree though that most of the power is not going to be saved by spinning down the disk. Laptop drives only use a few watts of power compared to the cpu and display.
The cable company in my area (Charter) sells fiber links for a reasonable price. Then AFAIK they also don't bother big uploaders in the residential side... yet.
Hmm... OSX would be one to disagree with you there (netinfo is not in flat files and neither is LDAP). OSX does have flat files on the BSD end but they aren't the ones you configure with the gui. The os that used netinfo before OSX was NeXTSTEP and everyone thought it was a bad idea then too. There is also the gconf database that Gnome uses, while similar to the registry in function it uses the file system directories for hierarchy and xml files for storage so it is easy to edit by hand if you need to.
Not to knock flat files (I prefer them myself) but one could say the same about finding them. The registry was designed to be a structured database of configuration data, Microsoft just screwed things up badly with it.
Umm... no. If you look at Microsoft's reference documentation on how companies should and should not use the windows registry I think you will note that the locations where things *should* be stored are quite well defined. Microsoft themselves may have issues with using there own standards but I have never found anti-piracy to be one of their reasons for being incompetent.
Now you are starting to sound paranoid. Can you point out the place in a word document where a serial number is kept? Give me a link or I call BS. I know it is stored in the registry (Where else do you think it would be?) but so are several other bits of license related data. HINT: it is not called an install code, it's called a license key. If you removed the key from the registry it stops working and asks for a key when you start it up.
IIRC this feature was included in the version of DOS that shipped with Windows 98 so that must not have been their reasoning otherwise they would have left it out.
Hmm... I can think of 2 design decisions why a databased approach to configuration has advantages.
The only part of the kernel that needs assembly is the parts that interact directly with hardware (small parts of driver code). Everything else can (and is) written in higher level languages.
And that was argued in a copyright case. The claim was that copying a program to the memory on your computer could count as copyright violation. The courts ruled that copying a program to memory or to a hard disk was not copyright infringement when such use was necessary to the operation of the program.
You are over complicating the opt-out procedure. A usual opt-out request is done once and stored or attached to the data. You don't have to tell everyone who searches the web that you opt-out, you don't even have to tell every search engine, just opt-out in a robots.txt file or in meta headers.
Actually it has more to do with DNS, the lists based on static/dynamic have several methods of checking on addresses before they are listed as dynamic most of them are based on DNS lookups. If your address has no RDNS it is likely going to be listed. If your address has an RDNS of dhcp-ip-ip-ip-ip.yourisp or ip-ip-ip-ip.yourisp it will be listed. An RDNS of static-ip-ip-ip-ip.yourisp will not be listed, usually anyway. Other non-address based RDNS also usually lands you in the clear. A list can't check to see if you are on DHCP or not unless your ISP publishes that kind of information but the RDNS name gives most people a clue about what kind of IP they are dealing with.
NiMH batteries AFAIK are trickle charged when full. This means that power is ALWAYS flowing into the battery but at such a low rate that 100% of it is wasted to the internal resistance of a fully charged battery. It's how a charging circuit knows when it battery is charged. Also note that NiMH batteries loose charge even when not under a load at a slow rate (30%/month). If you take the battery out when you don't need it and you don't need it for a month at a time then it will be (mostly) discharged when you do need it.
LI batteries have no natural discharge rate so they are not charged up to full unless their charge drops below a threshold. The loss rate of the charging/monitoring circuit is only 5%/month. A good reason to remove LI batteries when not in use is lifespan rather than safety. LI batteries last longer at lower temperatures and at partial charges. Recommended storage is at 40% charge and 0C (32F).
Speaking of smacking the computer...
We had a lab full of PCs that would lockup if jolted the wrong way while upright. To get them to unlock they would have to be powered down and dropped on their side from about 8 inches up. It turned out that the motherboards had bad memory slots and the memory contacts would loose connection. I eventually tried several techniques to keep the memory in place (solder, silicone, hot glue). One of the best methods was to use solder to tin the contacts on one side of the memory, took me about an hour per DIMM to perform the procedure so I never got around to more than just the worst computers in the bunch.
Ah, but they could be advertising with no delivery. An analogy to this would be putting an add in the newspaper that read, "Contact me and we will swap photocopies of our books." and suing everyone who contacted you wanting to trade you a copy of your book.
Putting your book in the public domain? No. Entrapment? Yes.
Some learning uses are simply by-products of normal use and can be treated as normal use; other techniques such as reverse engineering are generally outlawed one way or another.
Well, I think you are confusing law with contract (EULA). Reverse engineering is perfectly acceptable unless you have agreed to a contract that forbids it. This is why many software titles in recent publication requires you to agree to a EULA before you USE the software (and when you install it).
All that aside, your argument about worth and cost of physical items is flawed. Some books cost much more than they are worth as an object. I've had books that cost $300 for school. You do not need an EULA to protect your software, you only need it to limit liability and/or limit your customers use of said software. As a contract they are judged on a case by case basis as to enforceability and if the user agreed to them.
With a bulk license you agree to the terms before they sell you the software. One of the terms is no-resale.
Mine is a set of 5x250gig 7200 rpm SATA drives. I've rebuilt the raid three times now, twice from within the OS (linux) and once from the RAID configuration BIOS. The BIOS configuration seems to take forever to rebuild (more than 24 hours) but from within the OS it only takes about 2 hours. All on a P-III 600.
8 4256
d =10677428
I have another hardware RAID 5 that is SCSI (6 x 72 gig @ 10k rpm), it rebuilds in about 45 min.
The only software RAID that I have done was using much smaller drives (6 gig) and never needed to be rebuilt.
Comments from this thread:
http://ask.slashdot.org/article.pl?sid=04/10/30/1
and specifically this comment:
http://ask.slashdot.org/comments.pl?sid=127776&ci
Indicate that rebuild times in software are not bad on a 1ghz system. We can extrapolate out that a rebuild of 4x250 gig drives should take less than 2 hours (calculating parody on 3 rather than 2 drives for 2x as much data). Unless you are using the drives or have problems with a controller, then you might have more problems.
I have a 5 disk array using a highpoint raid controller, it rebuilds in about 2 hours when the computer is booted. It takes a full day to rebuild from the BIOS however.
Just so you know there is a quick fix. Replace ever instance of <embed src="stuff"> with <script src="embedStuff.js"></script> and have embedStuff.js be document.write('<embed src="stuff">'); Simpe to do with a python/perl/shell script... :)
Hmm... I guess it depends on what you do. From experience I would guess that IT where you work is implementing policy directed to them by management and implementing infrastructure to make their life easier. I've had management direct me to implement some seriously silly policies, usually I have a discussion with my manager about them and we work out a compromise, sometimes I'll implement them and direct comments about them directed to the management that required them.
It's been my experience that a small % of the user base can be trusted to know what all they need, how it works and how to maintain it. Many of those who THINK they can do it on their own end up screwing up more than just their own systems. Most of what we have to deal with in this is software related.
I like the vast majority of my users, I consider myself lucky in this. The few users that I have that cause problems (from my standpoint) are either because: They know nothing about technology and have no desire to learn (I've had users refuse training because they "didn't have the time" and ask me to drive over to show them how to operate basic software) or they think they know how things should be done and have budget authority for their department (think users buying their own wireless access points).
If your input is disregarded without a second thought or not even sought after, I feel your pain, it happens to me all the time. I make it a point to get my user's opinions on technical matters that they are knowledgeable about, but sometimes there are other needs that outweigh their advice. Part of my job is to research the BEST solution to problems as a whole.
A man-in-the-middle attack can still be performed in this approach. Z has broken the wires and is now acting as an intermediary between signals. Z monitors both wires for disconnects and then disconnects the other end when that happens. Z then knows the state of both ends at any given time so Z knows which sequence each has chosen.
Wouldn't it be odd if I went to a pay phone and dialed 911 only to hear "please insert 75 cents to continue."
AOL's big mistake was resting on their laurels. They were once pioneers of the internet but somewhere along the line they came up with some terrible problems that they continued to ignore. Now they are playing catch-up with the other big players on the net and doing a terrible job of it. They want to be like every other joe on the net and have doomed themselves to a loosing battle.
The problem with this is that most corperations earn next to nothing. They pay no taxes because they have ~$0 net income. The money goes back to investors, pays employees, funds research, etc.
The main problem is defining what is earnings. If you say grose income is earnings then those in low-margin/high-volume markets can earn money and go out of business at the same time.
If you say net income is earnings then most corperations make nothing.
IIRC the case you are referring to was a case of trademark infringement not of copyright. In that case it was ruled that it WAS infringement of Nintendo's trademark for the game to run the initialization sequence on the NES that caused it to display the Nintendo logo.
I had an aunt with the same problem. Only the first book was put on the publishers back shelf (no marketing.) Now she has two more in the series but she can't get them published because the first didn't sell enough copies.