The only reason why Oracle exists is because their database is far and away the best database server on the planet. Informix? Ingres? Sybase? Progress? Rdb? Anyone hear of them anymore? There is SQL server if you are a Microsoft shop, and DB2 if you are an IBM shop, occasionally PostgreSQL and MySQL. All those other databases are dead or dying because they are vastly inferior to Oracle for general purpose use.
I suspect Oracle itself will choose ZFS. The only question is whether they want ZFS support on Linux or not. If yes, btrfs will probably go the way of all the earth. If no, btrfs will probably survive in other hands.
I seriously doubt that Oracle is going to rename MySQL to anything other than "Oracle MySQL" because it would lead to massive confusion, and throw away a perfectly good brand name and risk an enormous amount of mind share / marketing awareness.
The number one reason why it can't possibly be called "MyOracle" is that MySQL and Oracle server are not remotely compatible. The number two reason is that it would seriously dilute the Oracle brand to have people confusing the Oracle database server with Oracle MySQL server. Oracle did this before. They purchased Rdb from Compaq/Digital. They called the result "Oracle Rdb".
Re:This doesn't sound like a good move.
on
Oracle Buys Sun
·
· Score: 1
Steer them to what? IBM has DB2, but that preference is nothing new. Anybody inclined to use SQL Server is probably a Microsoft shop already. Oracle practically owns the Unix mid range database server market. PostgreSQL is a decent option, but one that probably won't make HP / Dell any money, unless they get involved in it in a big way.
Oracle didn't get to be where it is today with a sub par database server. Oracle is far and away the best database server on the planet, bar none. No one is likely come close for at least a decade. Not Microsoft, not IBM, not PostgreSQL, not anybody. If anything the gap between Oracle and its database competitors is getting wider.
There is little reason why Oracle will not gradually phase the MySQL code base out of existence. There is a narrow class of applications that MySQL can run more effectively, but that is more a like a division between "read only" and "transactional" database, not "division" and "enterprise". MySQL isn't a good fit even for most "departmental" applications when compared to Oracle. About the only things it has going for it in an Oracle world are mind share, compatibility with existing applications, and installed base.
The idea of Oracle encouraging any existing Oracle database server customer to replace what they have with something derived from the MySQL code base is borderline ludicrous. Oracle was more advanced than where MySQL is now twenty years ago. MySQL is extremely primitive by comparison in almost any way you care to name.
If an employer told me we were going to switch to MySQL from Oracle for our database of record I would start looking for a new job tommorrow. If I were given a job maintaining the MySQL code base, my first question would be "when do we rewrite this from scratch"?
The worst part of MySQL is backward compatibility with dozens of design decisions that were made in a complete vacuum, that make MySQL about the least SQL compliant database on the planet. It is a decent simple minded database for a narrow class of one off applications, but if I were Oracle I would phase out all development of "enterprise" features for MySQL immediately. If they want to keep it around they are going to have to rewrite most of it from scratch to make it respectable anyway.
Oracle using ZFS transactional features
on
Oracle Buys Sun
·
· Score: 1
That is so impractical it is not even funny. It would amount to rewriting ~30-40% of the internal core of the Oracle database, redoing more than twenty years of work, and would make it so that Oracle would no longer be portable to any other filesystem.
Not only that, ZFS is designed for a completely different workload than Oracle anyway. Oracle is designed to run on the equivalent of raw disk devices. If you snapshot an active Oracle datafile, a copy on write filesystem like ZFS will start relocating newly written blocks, which will gradually defeat all of Oracle's internal I/O optimization logic until the datafiles are defragmented back into linear order.
The abbreviation is the summary is wrong. The correct unit is "Newton seconds" or "Ns", not "ns". Newton-seconds are a measure of impulse, i.e. the average thrust generated times the duration of the motor burn.
Of course the peak thrust matters too, if it is not greater than the weight of the rocket, the rocket won't lift off of the ground. Any net upward acceleration will only occur during the period the rocket is pointed up and the motor thrust is greater than the weight. Of course the rocket will keep moving upward for a considerable period after the net upward acceleration has stopped. A typical projectile (i.e. a non rocket) experiences net downward acceleration (due to gravity) during its entire flight.
I agree that all metro level cable and telecom distribution networks should be regulated monopolies, with open access required at reasonable rates where feasible. Competition in many areas is nonexistent (where service is available at all) due to cherry picking, often down to the level of individual cul-de-sacs.
A reasonable rate for a service that requires a network buildout, of course, includes sufficient price to cover the risk and cost to the distribution network provider of building out the network in the first place.
However, even then, the economics will not support everyone watching all their television on their own schedule over the internet for sometime to come. Hybrid multicast and DVR architectures are the only practical way to deliver IPTV to a large customer base.
The practical service limits are not numbers you can just pull out of thin air. It depends on the economics and technical capacity of the distribution network, how much it costs to expand that capacity, how long before the equipment is obsolete, and so on.
More to the point, the current cable television distribution network is hundreds if not thousands of times more efficient than ordinary internet television. The reason is because the cable television network delivers a fixed number of channels on a fixed time schedule to all customers. It doesn't matter how many people are watching the cost to deliver a channel remains the same.
The Internet is not like that because Internet multicast is non-existent and everyone wants to watch on their own schedule anyways. So if 10,000 people are watching LOST online, ABC generally has to deliver 10000 separate HD video feeds, which consumes a non-trivial amount of bandwidth. It is going to be a decade or more before everyone can economically watch television over their primary Internet connection. Current IPTV services generally do it over a local multicast network, which is hundreds of times more efficient, provided there is a fixed programming schedule and lots of people watching the same thing at the same time, etc.
It costs a lot more to deliver bandwidth to homes than to a data center. The extra cost is due to the cost of constructing and maintaining the distribution network. The per connection cost needs to be excluded to make a serious rate comparison. Electrical wholesale rates are about 1/3 of retail rates, and the cost of the distribution network is the major part of the difference. Of course if the electric company was allowed to use a rational pricing policy and charge customers about thirty dollars a month whether they used any electricity or not, the differential (i.e. the difference in the marginal rate for each additional kilowatt hour) would go down.
What annoys me is the silliness of articles like the recent one at arsTechnica that assume that the cost of a connection with zero data transfer is actually zero. It is not. For a cable company it is probably twenty to thirty dollars a month, depending on the amortization schedule. All actual services are on top of that. So dividing the base charge by how many GB you get at that level gives you a number that has nothing to do with how much it costs the cable company to purchase that bandwidth wholesale. The cost for low transfer users is overwhelmingly the cost of the physical connection itself.
I would guess that when this settles down, the cost is likely to be equivalent to $30 a month plus $0.20/GB. The $30/month may be lower of course if you have regular cable television service in addition to cable internet service. If anyone gets a plan that costs less than $30/mo without being required to purchase cable television service in addition to that, they are probably being cross subsidized by other users.
According to the article, the team is based at the Space and Naval Warfare Systems Command (SPAWAR) in San Diego, California. The announcement was made at a conference in Utah.
It doesn't matter if the file was uploaded in Italy. Unless there is some proof of a criminal conspiracy between Google and the uploading individual, it is the latter that has presumably committed a crime within Italian jurisdiction.
First of all, this sort of prosecution is likely to be immensely counterproductive. What kind of businessperson would want to travel to Italy when they prosecute individuals for supervisory responsibility of departments that have made a diligent, good faith effort to comply with the local laws from 10,000 miles away?
In addition, the modern conception of legal jurisdiction is all screwed up. Traditionally, jurisdiction comes with the territory. Physical presence in the jurisdiction is required when an essential part of the crime is committed. That is why, for example, states cannot force companies who do not have a physical presence in their state to collect sales tax on online purchases for them. The idea that you can prosecute somebody for an ordinary crime when all the relevant actions occurred outside your physical jurisdiction is a very bad precedent.
So rather than arresting visiting Google executives, if Italy feels so strongly about this, why not just shut down Google's local operations (if any), or create a national firewall and filter Google at the border? Or require ISPs to filter their entries from local DNS servers? Or threaten to do so unless Google pays some civil fine?
Copyright does not protect portions of derived works that lack originality. That principle is the basis of the recent court decision Bridgeman Art Library vs. Corel Corp.. Unless the publisher has made substantial changes to Kipling's work, I dare say we are not dirty rotten intellectual property stealing criminals after all.
Stroustrup is wrong (or perhaps over-enthusiastic about teaching C++ to beginners). If it were any other language, sure, but using C++ for anything but trivial applications without a knowledge of C is an exercise in futility.
For example, you can't use "new" for anything (other than wasting memory) without using pointers in C++, so one better have some semblance of an idea about how pointers work before trying to write anything that doesn't look like a FORTRAN program.
The problem is that (for all its benefits - and there are many), the design of C++ is forever compromised by the requirement to be backwards compatible with C. String literals aren't strings, they are *pointers* to a series of signed eight bit integers. When new objects are allocated, by default they contain random garbage data. No array bounds checking anywhere. Does a language really need both references *and* pointers? Other than backward compatibility?
I wouldn't dream of teaching C++ to anyone who didn't have a working knowledge of C, because using C++ to do anything worth the extra complexity of the language requires it - and a lot more.
The "feature" of MSSQL that allows compound statements separated by semicolons, without any sort of begin / end block, makes it particularly vulnerable to this type of SQL injection attack.
This type of attack requires more than just modifying a where clause or changing a value - it requires injecting one or more complete statements. MSSQL allows you to begin a new statement in the middle of any unchecked parameter. Most databases don't. That is why MSSQL is unusually vulnerable to this sort of attack.
Oracle applications with unchecked parameters (for example) are difficult to exploit this way because compound statements must occur in a PL/SQL begin end block and it is usually impossible to inject text at both the beginning and end of a statement.
If you do not hold a copyright in the material being distributed, you lack legal standing to enforce the license. That may be the reason why they are ignoring you. You need to contact someone who is a pertinent copyright holder and who is interested in enforcing the license to his or her work.
I think this decision could have gone the other way. In the court's opinion what the modelers did was not even minimally creative. Another court might disagree.
Aside: The contract was not at issue here. However, if Toyota paid the full cost for someone to create a wireframe model of one of their cars and did not require the copyright to be assigned back to them, someone didn't do their job very well.
The summary is wrong. The court did not determine that digital models are not subject to copyright. They merely decided that these particular wireframe models (of Toyota vehicles) were not - in and of themselves - original works of authorship in which new copyright privileges rest with the modelers (MeshWerks).
The vehicle designer (Toyota) retains its design patents on the vehicles and the presumptive copyright on any creative expression reflected in the design of their vehicles. The models here are clearly derivative works. The court ruled nothing substantially new was added to grant new rights to the modelers. However, Toyota designed the vehicles, that design is reflected in immaculate detail in the models, and as such the models presumably may not be copied without Toyota's permission, barring some sort of fair use exception.
This decision rests on a landmark Supreme Court precedent called Feist Publications v. Rural Telephone Service (1992), in which the Court held that the lists of names, addresses, and phone numbers in telephone directories were compilations of facts not creative works of authorship protectable by copyright.
This decision opens new ground, however, suggesting that much of the contents of any comprehensive digital model of the real world (digital maps come to mind) may not be independently protectable by copyright to the degree that those contents are intended to accurately reflect pre-existing reality rather than the creative selection or arrangement of the creator.
Please explain how you would exploit the following purposely vulnerable example in Oracle, assuming you had arbitrary control over the employee_nbr variable:
string employee_nbr;
sql = "SELECT * FROM EMPLOYEE WHERE EMPLOYEE_ID = " + employee_nbr; Execute(sql);
In MS SQL you could submit "1; DELETE FROM EMPLOYEE" and it would be game over. The same thing doesn't work in Oracle. The best you might do is submit "1 OR 1 = 1" and get information you weren't supposed to have.
Slashdot Mirror
The only reason why Oracle exists is because their database is far and away the best database server on the planet. Informix? Ingres? Sybase? Progress? Rdb? Anyone hear of them anymore? There is SQL server if you are a Microsoft shop, and DB2 if you are an IBM shop, occasionally PostgreSQL and MySQL. All those other databases are dead or dying because they are vastly inferior to Oracle for general purpose use.
I suspect Oracle itself will choose ZFS. The only question is whether they want ZFS support on Linux or not. If yes, btrfs will probably go the way of all the earth. If no, btrfs will probably survive in other hands.
I seriously doubt that Oracle is going to rename MySQL to anything other than "Oracle MySQL" because it would lead to massive confusion, and throw away a perfectly good brand name and risk an enormous amount of mind share / marketing awareness.
The number one reason why it can't possibly be called "MyOracle" is that MySQL and Oracle server are not remotely compatible. The number two reason is that it would seriously dilute the Oracle brand to have people confusing the Oracle database server with Oracle MySQL server. Oracle did this before. They purchased Rdb from Compaq/Digital. They called the result "Oracle Rdb".
Steer them to what? IBM has DB2, but that preference is nothing new. Anybody inclined to use SQL Server is probably a Microsoft shop already. Oracle practically owns the Unix mid range database server market. PostgreSQL is a decent option, but one that probably won't make HP / Dell any money, unless they get involved in it in a big way.
Oracle didn't get to be where it is today with a sub par database server. Oracle is far and away the best database server on the planet, bar none. No one is likely come close for at least a decade. Not Microsoft, not IBM, not PostgreSQL, not anybody. If anything the gap between Oracle and its database competitors is getting wider.
There is little reason why Oracle will not gradually phase the MySQL code base out of existence. There is a narrow class of applications that MySQL can run more effectively, but that is more a like a division between "read only" and "transactional" database, not "division" and "enterprise". MySQL isn't a good fit even for most "departmental" applications when compared to Oracle. About the only things it has going for it in an Oracle world are mind share, compatibility with existing applications, and installed base.
The idea of Oracle encouraging any existing Oracle database server customer to replace what they have with something derived from the MySQL code base is borderline ludicrous. Oracle was more advanced than where MySQL is now twenty years ago. MySQL is extremely primitive by comparison in almost any way you care to name.
If an employer told me we were going to switch to MySQL from Oracle for our database of record I would start looking for a new job tommorrow. If I were given a job maintaining the MySQL code base, my first question would be "when do we rewrite this from scratch"?
The worst part of MySQL is backward compatibility with dozens of design decisions that were made in a complete vacuum, that make MySQL about the least SQL compliant database on the planet. It is a decent simple minded database for a narrow class of one off applications, but if I were Oracle I would phase out all development of "enterprise" features for MySQL immediately. If they want to keep it around they are going to have to rewrite most of it from scratch to make it respectable anyway.
That is so impractical it is not even funny. It would amount to rewriting ~30-40% of the internal core of the Oracle database, redoing more than twenty years of work, and would make it so that Oracle would no longer be portable to any other filesystem.
Not only that, ZFS is designed for a completely different workload than Oracle anyway. Oracle is designed to run on the equivalent of raw disk devices. If you snapshot an active Oracle datafile, a copy on write filesystem like ZFS will start relocating newly written blocks, which will gradually defeat all of Oracle's internal I/O optimization logic until the datafiles are defragmented back into linear order.
Re-read the sentence. The key word is "net".
The abbreviation is the summary is wrong. The correct unit is "Newton seconds" or "Ns", not "ns". Newton-seconds are a measure of impulse, i.e. the average thrust generated times the duration of the motor burn.
Of course the peak thrust matters too, if it is not greater than the weight of the rocket, the rocket won't lift off of the ground. Any net upward acceleration will only occur during the period the rocket is pointed up and the motor thrust is greater than the weight. Of course the rocket will keep moving upward for a considerable period after the net upward acceleration has stopped. A typical projectile (i.e. a non rocket) experiences net downward acceleration (due to gravity) during its entire flight.
I agree that all metro level cable and telecom distribution networks should be regulated monopolies, with open access required at reasonable rates where feasible. Competition in many areas is nonexistent (where service is available at all) due to cherry picking, often down to the level of individual cul-de-sacs.
A reasonable rate for a service that requires a network buildout, of course, includes sufficient price to cover the risk and cost to the distribution network provider of building out the network in the first place.
However, even then, the economics will not support everyone watching all their television on their own schedule over the internet for sometime to come. Hybrid multicast and DVR architectures are the only practical way to deliver IPTV to a large customer base.
The practical service limits are not numbers you can just pull out of thin air. It depends on the economics and technical capacity of the distribution network, how much it costs to expand that capacity, how long before the equipment is obsolete, and so on.
More to the point, the current cable television distribution network is hundreds if not thousands of times more efficient than ordinary internet television. The reason is because the cable television network delivers a fixed number of channels on a fixed time schedule to all customers. It doesn't matter how many people are watching the cost to deliver a channel remains the same.
The Internet is not like that because Internet multicast is non-existent and everyone wants to watch on their own schedule anyways. So if 10,000 people are watching LOST online, ABC generally has to deliver 10000 separate HD video feeds, which consumes a non-trivial amount of bandwidth. It is going to be a decade or more before everyone can economically watch television over their primary Internet connection. Current IPTV services generally do it over a local multicast network, which is hundreds of times more efficient, provided there is a fixed programming schedule and lots of people watching the same thing at the same time, etc.
It costs a lot more to deliver bandwidth to homes than to a data center. The extra cost is due to the cost of constructing and maintaining the distribution network. The per connection cost needs to be excluded to make a serious rate comparison. Electrical wholesale rates are about 1/3 of retail rates, and the cost of the distribution network is the major part of the difference. Of course if the electric company was allowed to use a rational pricing policy and charge customers about thirty dollars a month whether they used any electricity or not, the differential (i.e. the difference in the marginal rate for each additional kilowatt hour) would go down.
What annoys me is the silliness of articles like the recent one at arsTechnica that assume that the cost of a connection with zero data transfer is actually zero. It is not. For a cable company it is probably twenty to thirty dollars a month, depending on the amortization schedule. All actual services are on top of that. So dividing the base charge by how many GB you get at that level gives you a number that has nothing to do with how much it costs the cable company to purchase that bandwidth wholesale. The cost for low transfer users is overwhelmingly the cost of the physical connection itself.
I would guess that when this settles down, the cost is likely to be equivalent to $30 a month plus $0.20/GB. The $30/month may be lower of course if you have regular cable television service in addition to cable internet service. If anyone gets a plan that costs less than $30/mo without being required to purchase cable television service in addition to that, they are probably being cross subsidized by other users.
According to the article, the team is based at the Space and Naval Warfare Systems Command (SPAWAR) in San Diego, California. The announcement was made at a conference in Utah.
He is not talking about "macro substitution", he is talking about "command substitution". Big difference.
It doesn't matter if the file was uploaded in Italy. Unless there is some proof of a criminal conspiracy between Google and the uploading individual, it is the latter that has presumably committed a crime within Italian jurisdiction.
First of all, this sort of prosecution is likely to be immensely counterproductive. What kind of businessperson would want to travel to Italy when they prosecute individuals for supervisory responsibility of departments that have made a diligent, good faith effort to comply with the local laws from 10,000 miles away?
In addition, the modern conception of legal jurisdiction is all screwed up. Traditionally, jurisdiction comes with the territory. Physical presence in the jurisdiction is required when an essential part of the crime is committed. That is why, for example, states cannot force companies who do not have a physical presence in their state to collect sales tax on online purchases for them. The idea that you can prosecute somebody for an ordinary crime when all the relevant actions occurred outside your physical jurisdiction is a very bad precedent.
So rather than arresting visiting Google executives, if Italy feels so strongly about this, why not just shut down Google's local operations (if any), or create a national firewall and filter Google at the border? Or require ISPs to filter their entries from local DNS servers? Or threaten to do so unless Google pays some civil fine?
Solar energy in what unit of time? A day? A second?
Copyright does not protect portions of derived works that lack originality. That principle is the basis of the recent court decision Bridgeman Art Library vs. Corel Corp.. Unless the publisher has made substantial changes to Kipling's work, I dare say we are not dirty rotten intellectual property stealing criminals after all.
Stroustrup is wrong (or perhaps over-enthusiastic about teaching C++ to beginners). If it were any other language, sure, but using C++ for anything but trivial applications without a knowledge of C is an exercise in futility.
For example, you can't use "new" for anything (other than wasting memory) without using pointers in C++, so one better have some semblance of an idea about how pointers work before trying to write anything that doesn't look like a FORTRAN program.
The problem is that (for all its benefits - and there are many), the design of C++ is forever compromised by the requirement to be backwards compatible with C. String literals aren't strings, they are *pointers* to a series of signed eight bit integers. When new objects are allocated, by default they contain random garbage data. No array bounds checking anywhere. Does a language really need both references *and* pointers? Other than backward compatibility?
I wouldn't dream of teaching C++ to anyone who didn't have a working knowledge of C, because using C++ to do anything worth the extra complexity of the language requires it - and a lot more.
The "feature" of MSSQL that allows compound statements separated by semicolons, without any sort of begin / end block, makes it particularly vulnerable to this type of SQL injection attack.
This type of attack requires more than just modifying a where clause or changing a value - it requires injecting one or more complete statements. MSSQL allows you to begin a new statement in the middle of any unchecked parameter. Most databases don't. That is why MSSQL is unusually vulnerable to this sort of attack.
Oracle applications with unchecked parameters (for example) are difficult to exploit this way because compound statements must occur in a PL/SQL begin end block and it is usually impossible to inject text at both the beginning and end of a statement.
If you do not hold a copyright in the material being distributed, you lack legal standing to enforce the license. That may be the reason why they are ignoring you. You need to contact someone who is a pertinent copyright holder and who is interested in enforcing the license to his or her work.
I think this decision could have gone the other way. In the court's opinion what the modelers did was not even minimally creative. Another court might disagree.
Aside: The contract was not at issue here. However, if Toyota paid the full cost for someone to create a wireframe model of one of their cars and did not require the copyright to be assigned back to them, someone didn't do their job very well.
The summary is wrong. The court did not determine that digital models are not subject to copyright. They merely decided that these particular wireframe models (of Toyota vehicles) were not - in and of themselves - original works of authorship in which new copyright privileges rest with the modelers (MeshWerks).
The vehicle designer (Toyota) retains its design patents on the vehicles and the presumptive copyright on any creative expression reflected in the design of their vehicles. The models here are clearly derivative works. The court ruled nothing substantially new was added to grant new rights to the modelers. However, Toyota designed the vehicles, that design is reflected in immaculate detail in the models, and as such the models presumably may not be copied without Toyota's permission, barring some sort of fair use exception.
This decision rests on a landmark Supreme Court precedent called Feist Publications v. Rural Telephone Service (1992), in which the Court held that the lists of names, addresses, and phone numbers in telephone directories were compilations of facts not creative works of authorship protectable by copyright.
This decision opens new ground, however, suggesting that much of the contents of any comprehensive digital model of the real world (digital maps come to mind) may not be independently protectable by copyright to the degree that those contents are intended to accurately reflect pre-existing reality rather than the creative selection or arrangement of the creator.
Please explain how you would exploit the following purposely vulnerable example in Oracle, assuming you had arbitrary control over the employee_nbr variable:
string employee_nbr;
sql = "SELECT * FROM EMPLOYEE WHERE EMPLOYEE_ID = " + employee_nbr;
Execute(sql);
In MS SQL you could submit "1; DELETE FROM EMPLOYEE" and it would be game over. The same thing doesn't work in Oracle. The best you might do is submit "1 OR 1 = 1" and get information you weren't supposed to have.