Since I am qualified by my experience at the CBOE diagnosing problems sending and (mostly) receiving market data from SIAC, I can say that they mostly provide the service of a stock price reflector. They have (previously: redundant T1 lines) WAN links going from SIAC in New York to all registered stock exchanges trading NYSE listed stocks. Prices come in asynchronously, and they are broadcast (via IGMP) to stock exchanges and subscribers (like Reuters and ADP). The rest of us get our stock quotes from SIAC exchanges and subscribers.
Currently, you have to write your own interface to parse the SIAC quote streams. It is supposed to be a lossless protocol with sequence numbers that almost ran out during the market frenzy last year. Proprietary protocols are scary when they don't scale... The stock market, on the busiest day in history, might have had to be closed early for lack of quote message ids.
Linux is much more modern than AIX in security and modularity, and has that low development cost appeal. Thus IBM can upgrade SIAC, replacing old scary crufty protocols with up-to-date standards based protocols, and IBM doesn't have to staff up so much to support it (though they can still charge the same).
Oh, and IBM is selling Linux on 390 mainframes... a "Lin-Frame" machine... You still think you have to run MVS/OS390 or TPF on a mainframe? It's a Linux box with N processors and N terabytes of "DASD".
A filesystem is a service. It (should) provide storage and retrieval of any string of bits that will fit. The only essential (key) metadata is the filename/location (address) information. Dates and permissions/ownership are metadata that the filesystem provides for itself, and related file-management software.
File type metadata is redundant. For proof, I offer the example of steganography software. The essential part of steganography is the ability to look at data which appears to be unusable and yet somehow extract data which conforms to our expectation. My point is that file type metadata can be computed from the data itself, and therefore any stored designation is arbitrary and unneccesary.
You might think "why not talk about magic numbers?" Let's do: steganography can be computationally intensive. If you don't need to obscure the data's purpose, why not embed a label in the data itself? Oh! We do! Most (well designed) file format specifications include such a designation in the first few bytes of their internal structure.
Theoretically, a filesystem can do its job without knowing anything about the file contents. Oh! Some filesystems do! Utilities which manipulate the data served up by the filesystem cannot read and write files without prior knowledge of the file's contents. Which (application or filesystem) should be authoritative on file types? That's right: applications!
What happens when you ask the filesystem to choose which application to handle any given file? Then the filesystem needs prior knowledge of *all* file types for every file. It needs some kind of file type metadata on every file type. Windows handles this with the registry each time an application is installed: it tells the filesystem what kind of files it uses and how to recognise them (by yucky filename extensions). But, wait! That's Windows Explorer I'm talking about: a utility/application NOT the FAT32 or NTFS filesystem. Unix has a similar function called "magic". It is also an application/utility and not part of the filesystem. Even NeXT bundles (meta-files?) in MacOS X (and GnuStep) place the onus for file handling on the applications.
The file browser is an application, and it has needs which are a subset of filesystem needs, a subset of each application, and a set of its own peculiar features' dependancies. The Mac Finder, Windows Explorer, and every X-Windows filesystem browser (browser for short) has strategies to deal with each of these sets of needs. Let the browser handle the file typing! What is the problem?
Oh, that "creator" metadata that assumes you have several applications that handle a given file type, but you prefer to open a file with the application that created it, creates a knowledge gap between the application and the browser. Doesn't that break the assumption that a file type is a standard and that software should not assume one application is better than another?
Oh, I forgot: Mac users have been promised they won't have to make any unnecessary choices (like which JPEG application to use for each file). This is an optional feature of a "Mac flavored" filesystem browser and should be implemented within the browser itself. How should (say in MacOS X) the filesystem browser get the creator metadata from the application? How about using RCS style metadata? Isn't creator information really historical data? Who should handle logging a file's history? Traditionally, that would be the application. Maybe the API should provide transparent historical metadata maintainance and hooks for the filesystem browser to access this history?
First I have a system, and it works, but leaves me with too many choices. Next I want the system to have intuition so I can avoid making some choices. Now my once clear system is as complex and muddy as my own thinking as I digress and eventually undermine my working system's design.
You see, eventually the historical metadata can become a neural network and anticipate data handling in more sophisticated ways without even the programmer having to make decisions about data handling...
The Sherman Antitrust Act forbids using dominance in one market (PC Processor) to leverage control over another market (PC Chipsets). If you feel Intel is harming your consumer choice, file a complaint with the US FTC online, but please RTFM first.
If you read carefully, the article warns "they [prima donna programmers] will hurt you" meaning that the middle/low manager has lost control of what they can be held accountable for." This is a classic power struggle, and the particular examples of prima donna behavior in the article are irrelevant. If a person [programmer/admin] has such a relationship with upper management, the hierarchy is upset and the middle/low manager will have the motivation to label said person as prima donna to compete for upper management's trust.
I don't know why people think corporations are the answer to everything. They provide advantages for sure, like legal indemnity for the people running them, but really what is that good for in the world of free software?
I think the world may be better served by something that doesn't have such a slant towards ripping people (investors, employees, customers) off. The next time someone says (insert company name here) is doing (insert project here) ask yourself if it really means anything and what and why.
If the software solves a real problem for the developer, his work is already worth something to himself. Why does he need a corporation to pay him for that work? Greed? Stupidity? Maybe he doesn't need to be concerned about that problem he was trying to solve for himself (and others) because some corporation is paying him to do something else which needs no other justification than a paycheck.
Think about why people write free software. People, I mean, and corporations I do not mean. A guy writes a program and posts it up on the web with GPL/BSD license. How has he been rewarded for the actual toil he invested in this? No paycheck is necessary. Remember: when you have a job the job also has you.
Please read Cornell Law School's Antitrust Primer. It will explain that it is illegal to use dominance in one market, like the PC OS dominance of Microsoft, inorder to influence another market, like the recording and distribution of music. Start posting the *DAMAGES* to consumer choice so we can talk about the monetary value of what they are taking away. Then we can sue them.
Using dominance in the OS market to push products in another market (ie web browsers or multimedia devices/formats/applications) is a violation of the Sherman Antitrust Act.
The music record industry is not a business of producing music.
The record industry is based on the value of the service they *do* provide.
The record industry *distributes* music, and the business is tied up in lucrative *distribution* agreements.
It used to be difficult to distribute music because of the scarcity of high-power radio transmitters and the bulkiness of the recorded music media.
The Internet has made music distribution cheap and easy.
The value of the services provided by the record industry is now diminished.
Producers and artists (some anyway..) will (achingly) slowly abandon the fantasy of getting obscenely rich overnight through big recording/distribution deals. (Ian MacKaye)
Musicians and producers will learn to make money promoting their fanclubs and networking with their fans who will buy product because they would be despondent if the band stopped making songs.
Music like N'Sync will not sell because their fans will not buy when they can download for free.
One-hit-wonders will make less money for big record companies
Big record companies will have to give better deals to artistic productions because their risk is lower after having established a fan base.
Good musicians will make more, but the most popular will prolly make less.
Bad musicians will have to keep their day jobs for sure.
Record company executives will have to take their golden parachutes in droves.
CC's come with insurance that has a $50 deductible. Coverage is null if you let someone else use your card or fail to notify the issuer's company in reasonable time that the card has been abused.
Paypal is just a CC vendor. They are like a card holder, but instead of card number, they get a vendor ID, and authority to put debits and credits on cardholders' accounts. They have credit reports on file like cardholders. If they break vendor rules, they get penalized. Maintaining the secrecy of your card number is part of their vendor contract. If their CC# storage is compromised, they get penalized by the CC company.
You will probably only have to pay $50 if your card number is stolen, but they may try to get you to pay more. Get a lawyer (one of those fix-your-credit guys) if there's a lot of money involved.
With that context, and to answer your question, it is possible there is a flaw in PayPal's software. However, considering the potential liability of that flaw, you should assume "someone" is getting paid to evaluate the system. It is doubtful there is any flaw as easy to exploit as a luser. This all depends on the "security consultant" that signed off on this thing.
I don't understand what the big deal is.
I'm a Unix Admin, and I can "become" a user any day of the week, and have had that capability for the ENTIRE ten years I've had root on ANY box.
First the person calls me. I answer some questions, trying to avoid having to type my own solution. If the user can't do it, they the user asks me to do "superuser" things (for example, "become" them and look at their files). This is OK because there is established accountability for things that happen here. Not so with an ASP. This is why Unix vendors have a hard time selling remote administration services to their clients. It is only done in a very restricted way, usually with (somewhat qualified) company personell supervising.
And you know what else them ignern't whaaht trash'll bleev? Why, lissen tah thee'is: Jesus wuz whaaht. He wern't no sand-nigger lahk them folks on teevee is always a sayin'. Lookit them pitchers of Jesus on yer granny's wall. That wern't no midl-eesturn sand-nigger. Why, ma pappy used ta say thet whaaht folk used ta live in them parts, but they got forced tah move up north on acount uh them sand-niggers.
Ah don't have issues now do ah?
OK, ok, ok.. I apologise for replying to this stupid rant, but I'm weak! He did the setup! I'm just the punch line! A bad joke is better than a good rant regardless..
The article showed slightly better performance
with MySQL and sendmail/fetchmail on linux because
of an intentional tradeoff of speed vs. reliability that each system makes. The #1 bottleneck in any system is mass storage. Tuning this part will net you the largest gains for your effort in any loaded system. Power them off --unplug them-- in the middle of an extra trial and compare what you have when the system boots. Weigh that against your disk performance.
FreeBSD is configured to mount its filesystems *synchronously* by default. This is slightly slower, but you are *MUCH* less likely to lose a filesystem in a hard crash. Linux mounts its ext2fs filesystems *ASYNCronously*. Rerun the tests with both mounting synchronously, and then run the Linux async vs. the FreeBSD server running with softupdates to see the relative effects of the disk tradeoff.
You might also try sticking a couple of extra gig of RAM in the box and mounting/var on a filesystem in memory. That would (mostly) take away the factor of relative DMA/PIO hardware I/O performance in the disk drivers.
The article takes a legal perspective: the compatibility monopolies section fails to acknowlege the connector conspiracy principle that we all know from the famed JARGON file. It completely misses vital necessary points that are common knowledge among the technical elite, and in the bias the conclusion falls on the wrong side.
I am acutely aware of the writers' ignorance on the subject. The problem with this fantastic assumption is that Microsoft products are *not* the best, even on price/performance ratio. The rest of the article is suspended over the vaccum of what *wanted* to be a supporting pillar of truth.
What I am afraid this article may suggest (to the wrong people) is that the Microsofts and IBMs of the world deserve the right to "innovate" new interfaces, but they must be mediated by a central authority. I fear this authority who will guarantee nothing about the under the table licensing schemes and pre-release engineering samples, but *will* guarantee that software patents worm their way in like a liver fluke as a safeguard for the incentive MS and IBM had to create all that superior technology.
Then again, this is preaching to the choir of "Standards" people. Just remember: standards are created by consortiums of big corporations when and if their legal departments see fit. Also remember to use free software: it rules (and The Law just drools).
Demand for bandwidth has peaked. How much crap
do believe is out there actually *worth* saturating your connection/senses with? You're not going to pay more for it. That much makes sense.
Demand for the Web has peaked. Anyone who was going to post a web page/file/whatever-content to an Internet server probably already has, or can without too much difficulty. (The Third-World respectfully excluded). Wether you find what you thought you wanted or not, there is so much available to you that the bottleneck is in your frontal lobes (your ability to understand/comprehend) and not in the server and not in the connections.
The Internet has NOT peaked.
Connectivity of the 300bps internet-phone variety will be extended to billions of new low-power nodes as flash ROM and CPU resources get cheap enough to make *everything* smart and connected. The Third-World *will* eventually reach similar per-capita node densities. The Internet will continue to grow, and it will fruit with practical examples of the theoretical capabilities (vaporware) people were foaming at the mouth about for the last two years.
I don't think anyone disagrees that tracking people by MAC address gives the supply side an edge in their marketing powers. What I think people are concerned about is the worst-case scenario of a more solid,
focused, revenue model.
Adam Smith capitalism is supposed to be consumer driven, but lately we've been seeing power shift to the producers. That gets dangerously close to fascism when you see how government involvement plays in.
I'LL TELL YOU Re:Why Even Bother?
on
GPG vs. PGP?
·
· Score: 1
Law Enforcement "profiling"
Law Enforcement's legal power of "Discretion"
Predator
strong crypto isn't just for privacy, its for authenticity and plausible deniability.
your military firearm won't stop anyone from emptying your bank accounts
This article is all snake oil. I'm disappointed that I read the whole thing and didn't learn anything about comparative DBMS architecture and implementations. The simply *is* no *how or why* to the results. What a useless piece of sales drivel! I feel no better equipped to decide on a favorite database backend.
The one who has an ear had better hear what the Spirit says to the churches. To
the one who conquers, I will give him some of the hidden manna, and I will give him a white stone, and on that stone will be
written a new name that no one can understand except the one who receives it.'
The verse is one of the worst understood, and the footnotes suggest that the greek word translated as "white" here could mean "bright" as in glowing. Imagine a backlit PDA with your PK keyrings on it that you can use to recieve, sign, encrypt, and send irrefutable tamper-proof messages. Why, you could use that in lieu of the mark-of-the-beast at least within the network/trust-ring of the holy apocolyptic black-market (white market?).
Status Quo: It is possible for a person to redeem themselves for past crimes. A necessary condition of this is that the person must be able to dissociate themselves from the link to the past crime. Past crimes are recorded with a reference to a criminal's identity. If it is possible to live without revealing that information, then and only then is it in a person's interest to pay a debt they owe to society.
World of implanted ID chips: Nobody can ever redeem themselves, because *somebody* will always be around to steal official records before they are erased, store them, and resell the data on the black market. Your "credit report" will contain inane crap like your school kindergarten disciplinary record and how much beer you bought in college. You will not be able to control this because the black market will shadow any official records. Anyone you deal with will use this data to discriminate against you. Since civil law has the "preponderance of evidence" burden of proof, you will be at a legal disadvantage if you do not use others' data as leverage against them. If you ever make a mistake, no matter how small, you can look forward to paying for it repeatedly the rest of your life. The concept of "benefit of the doubt" will no longer exist, and all trust decisions will effectively be arbitrated by the authorities who sell the data.
Your comment seems to unfairly suggest a few things:
A class in economics automatically disproves some necessary condition of RMS's position.
The models presented in such a class apply to RMS's position.
The lack of coherence between your pedantic economic model and RMS's position somehow elimanates the possibility of an unknown, sound economic model that *does* gel with RMS.
Your readers will automatically agree with you on the previous assumptions if they are aware of viable alternatives.
I hereby declare youre "broke and somewhat off your rocker" comment to be genuine FUD. Go FUD yourself.
I read a good book that helped me in the Unix learning curve. Unix is a system that adheres to principles, and is descended from the Original AT&T project.
BOYCOTT AMAZON
The UNIX Philosophy
by Mike Gancarz ISBN: 1555581234 Butterworth-Heinemann 1994
What makes you so sure they actually buy elections without the Internet? Maybe they coerce the vote-counters! Bullets are cheaper. Oh... You mean the *AMERICAN* presidential elections..
Maybe these early Internet voting schemes are straw-men to keep the old system (very opaque to the voters) in place? Maybe you're astroturfing for the Dept. of State!
Anyone who has discovered public-key crypto knows that Internet voting is both workable, desirable, and is *FAR* cheaper than punch-ballots if done *CORRECTLY*...
No name; no one to blame. This is how the Bill of Rights intended.
There will be no mandatory Internet user tracking. The reason is each thing you do with an Internet client is a REQUEST for information. Working this proposal would require practical violation of the First Amendment on two grounds, and possibly the Third Amendment on a third ground.
<rant>Bill Clinton should be ashamed that his middle name is Jefferson, and he will be recorded in history of attempting to sell out the rights American Citizen for short term political gain.</rant>
Requiring people to use a specific protocol which identifies them, abridges their right to make requests for information with impunity. It would also abridge (have a chilling effect) on an individual's likelihood of using such a request to establish an unpopular idea on a public bulletin board. It would also require agents of those services to enforce the protocol on the individual users, which constitutes abridging the freedom of the press.
IMnsHO, while the world considers the term cyber-warfare we should be prepared to establish the constitutional interperetation of quartering of cyber-soldiers in our homes. President Clinton's statement is preparatory for yet another key-escrow proposal. We need to get the idea out that escrowed keys are agents of the government, and are essentially cyber-soldiers which the President, and (thus far) potential Presidential Candidates Gore and McCain would like to install in our cyber-homes.
The Internet is a collection of computers and users who volunteer adherence to IP.
The new Corporate Internet is migrating away from IP. As a result, it will not be peer-to-peer, and it will not be open, and it will not be (is not) reliable.
Corporate networks appear to be DAMAGED (Dain Bramaged?) to traditional (good) IP hosts.
IP routes around damage in the network. Check out the (RBL) evolution of the Internet's Killer App: email. This is a strong and specific example of the old-school Internet segmenting the new-school Pseudo-internet. The new school sues, and the RBL lives! Paul Vixie is free to write software and distribute it, and we are all (somewhat) free to run sendmail, preserving the usefulness of our email system.
Like email, the rest of the real internet, loyal to our proven principles of good hosting, will simply fork off and let the Corpses rot in isolation.
Slashdot Mirror
Since I am qualified by my experience at the CBOE diagnosing problems sending and (mostly) receiving market data from SIAC, I can say that they mostly provide the service of a stock price reflector. They have (previously: redundant T1 lines) WAN links going from SIAC in New York to all registered stock exchanges trading NYSE listed stocks. Prices come in asynchronously, and they are broadcast (via IGMP) to stock exchanges and subscribers (like Reuters and ADP). The rest of us get our stock quotes from SIAC exchanges and subscribers.
Currently, you have to write your own interface to parse the SIAC quote streams. It is supposed to be a lossless protocol with sequence numbers that almost ran out during the market frenzy last year. Proprietary protocols are scary when they don't scale... The stock market, on the busiest day in history, might have had to be closed early for lack of quote message ids.
Linux is much more modern than AIX in security and modularity, and has that low development cost appeal. Thus IBM can upgrade SIAC, replacing old scary crufty protocols with up-to-date standards based protocols, and IBM doesn't have to staff up so much to support it (though they can still charge the same).
Oh, and IBM is selling Linux on 390 mainframes... a "Lin-Frame" machine... You still think you have to run MVS/OS390 or TPF on a mainframe? It's a Linux box with N processors and N terabytes of "DASD".
A filesystem is a service. It (should) provide storage and retrieval of any string of bits that will fit. The only essential (key) metadata is the filename/location (address) information. Dates and permissions/ownership are metadata that the filesystem provides for itself, and related file-management software.
File type metadata is redundant. For proof, I offer the example of steganography software. The essential part of steganography is the ability to look at data which appears to be unusable and yet somehow extract data which conforms to our expectation. My point is that file type metadata can be computed from the data itself, and therefore any stored designation is arbitrary and unneccesary.
You might think "why not talk about magic numbers?" Let's do: steganography can be computationally intensive. If you don't need to obscure the data's purpose, why not embed a label in the data itself? Oh! We do! Most (well designed) file format specifications include such a designation in the first few bytes of their internal structure.
Theoretically, a filesystem can do its job without knowing anything about the file contents. Oh! Some filesystems do! Utilities which manipulate the data served up by the filesystem cannot read and write files without prior knowledge of the file's contents. Which (application or filesystem) should be authoritative on file types? That's right: applications!
What happens when you ask the filesystem to choose which application to handle any given file? Then the filesystem needs prior knowledge of *all* file types for every file. It needs some kind of file type metadata on every file type. Windows handles this with the registry each time an application is installed: it tells the filesystem what kind of files it uses and how to recognise them (by yucky filename extensions). But, wait! That's Windows Explorer I'm talking about: a utility/application NOT the FAT32 or NTFS filesystem. Unix has a similar function called "magic". It is also an application/utility and not part of the filesystem. Even NeXT bundles (meta-files?) in MacOS X (and GnuStep) place the onus for file handling on the applications.
The file browser is an application, and it has needs which are a subset of filesystem needs, a subset of each application, and a set of its own peculiar features' dependancies. The Mac Finder, Windows Explorer, and every X-Windows filesystem browser (browser for short) has strategies to deal with each of these sets of needs. Let the browser handle the file typing! What is the problem?
Oh, that "creator" metadata that assumes you have several applications that handle a given file type, but you prefer to open a file with the application that created it, creates a knowledge gap between the application and the browser. Doesn't that break the assumption that a file type is a standard and that software should not assume one application is better than another?
Oh, I forgot: Mac users have been promised they won't have to make any unnecessary choices (like which JPEG application to use for each file). This is an optional feature of a "Mac flavored" filesystem browser and should be implemented within the browser itself. How should (say in MacOS X) the filesystem browser get the creator metadata from the application? How about using RCS style metadata? Isn't creator information really historical data? Who should handle logging a file's history? Traditionally, that would be the application. Maybe the API should provide transparent historical metadata maintainance and hooks for the filesystem browser to access this history?
First I have a system, and it works, but leaves me with too many choices. Next I want the system to have intuition so I can avoid making some choices. Now my once clear system is as complex and muddy as my own thinking as I digress and eventually undermine my working system's design.
You see, eventually the historical metadata can become a neural network and anticipate data handling in more sophisticated ways without even the programmer having to make decisions about data handling...
The Sherman Antitrust Act forbids using dominance in one market (PC Processor) to leverage control over another market (PC Chipsets). If you feel Intel is harming your consumer choice, file a complaint with the US FTC online, but please RTFM first.
If you read carefully, the article warns "they [prima donna programmers] will hurt you" meaning that the middle/low manager has lost control of what they can be held accountable for." This is a classic power struggle, and the particular examples of prima donna behavior in the article are irrelevant. If a person [programmer/admin] has such a relationship with upper management, the hierarchy is upset and the middle/low manager will have the motivation to label said person as prima donna to compete for upper management's trust.
I don't know why people think corporations are the answer to everything. They provide advantages for sure, like legal indemnity for the people running them, but really what is that good for in the world of free software?
I think the world may be better served by something that doesn't have such a slant towards ripping people (investors, employees, customers) off. The next time someone says (insert company name here) is doing (insert project here) ask yourself if it really means anything and what and why.
If the software solves a real problem for the developer, his work is already worth something to himself. Why does he need a corporation to pay him for that work? Greed? Stupidity? Maybe he doesn't need to be concerned about that problem he was trying to solve for himself (and others) because some corporation is paying him to do something else which needs no other justification than a paycheck.
Think about why people write free software. People, I mean, and corporations I do not mean. A guy writes a program and posts it up on the web with GPL/BSD license. How has he been rewarded for the actual toil he invested in this? No paycheck is necessary. Remember: when you have a job the job also has you.
Please read Cornell Law School's Antitrust Primer. It will explain that it is illegal to use dominance in one market, like the PC OS dominance of Microsoft, inorder to influence another market, like the recording and distribution of music. Start posting the *DAMAGES* to consumer choice so we can talk about the monetary value of what they are taking away. Then we can sue them.
Using dominance in the OS market to push products in another market (ie web browsers or multimedia devices/formats/applications) is a violation of the Sherman Antitrust Act.
CC's come with insurance that has a $50 deductible. Coverage is null if you let someone else use your card or fail to notify the issuer's company in reasonable time that the card has been abused.
Paypal is just a CC vendor. They are like a card holder, but instead of card number, they get a vendor ID, and authority to put debits and credits on cardholders' accounts. They have credit reports on file like cardholders. If they break vendor rules, they get penalized. Maintaining the secrecy of your card number is part of their vendor contract. If their CC# storage is compromised, they get penalized by the CC company.
You will probably only have to pay $50 if your card number is stolen, but they may try to get you to pay more. Get a lawyer (one of those fix-your-credit guys) if there's a lot of money involved.
With that context, and to answer your question, it is possible there is a flaw in PayPal's software. However, considering the potential liability of that flaw, you should assume "someone" is getting paid to evaluate the system. It is doubtful there is any flaw as easy to exploit as a luser. This all depends on the "security consultant" that signed off on this thing.
I don't understand what the big deal is. I'm a Unix Admin, and I can "become" a user any day of the week, and have had that capability for the ENTIRE ten years I've had root on ANY box.
First the person calls me. I answer some questions, trying to avoid having to type my own solution. If the user can't do it, they the user asks me to do "superuser" things (for example, "become" them and look at their files). This is OK because there is established accountability for things that happen here. Not so with an ASP. This is why Unix vendors have a hard time selling remote administration services to their clients. It is only done in a very restricted way, usually with (somewhat qualified) company personell supervising.
And you know what else them ignern't whaaht trash'll bleev? Why, lissen tah thee'is: Jesus wuz whaaht. He wern't no sand-nigger lahk them folks on teevee is always a sayin'. Lookit them pitchers of Jesus on yer granny's wall. That wern't no midl-eesturn sand-nigger. Why, ma pappy used ta say thet whaaht folk used ta live in them parts, but they got forced tah move up north on acount uh them sand-niggers.
Ah don't have issues now do ah?
The article showed slightly better performance with MySQL and sendmail/fetchmail on linux because of an intentional tradeoff of speed vs. reliability that each system makes. The #1 bottleneck in any system is mass storage. Tuning this part will net you the largest gains for your effort in any loaded system. Power them off --unplug them-- in the middle of an extra trial and compare what you have when the system boots. Weigh that against your disk performance.
FreeBSD is configured to mount its filesystems *synchronously* by default. This is slightly slower, but you are *MUCH* less likely to lose a filesystem in a hard crash. Linux mounts its ext2fs filesystems *ASYNCronously*. Rerun the tests with both mounting synchronously, and then run the Linux async vs. the FreeBSD server running with softupdates to see the relative effects of the disk tradeoff.
You might also try sticking a couple of extra gig of RAM in the box and mounting /var on a filesystem in memory. That would (mostly) take away the factor of relative DMA/PIO hardware I/O performance in the disk drivers.
The article takes a legal perspective: the compatibility monopolies section fails to acknowlege the connector conspiracy principle that we all know from the famed JARGON file. It completely misses vital necessary points that are common knowledge among the technical elite, and in the bias the conclusion falls on the wrong side.
When I read things like
I am acutely aware of the writers' ignorance on the subject. The problem with this fantastic assumption is that Microsoft products are *not* the best, even on price/performance ratio. The rest of the article is suspended over the vaccum of what *wanted* to be a supporting pillar of truth.What I am afraid this article may suggest (to the wrong people) is that the Microsofts and IBMs of the world deserve the right to "innovate" new interfaces, but they must be mediated by a central authority. I fear this authority who will guarantee nothing about the under the table licensing schemes and pre-release engineering samples, but *will* guarantee that software patents worm their way in like a liver fluke as a safeguard for the incentive MS and IBM had to create all that superior technology.
Then again, this is preaching to the choir of "Standards" people. Just remember: standards are created by consortiums of big corporations when and if their legal departments see fit. Also remember to use free software: it rules (and The Law just drools).
On Has The Internet Peaked? :
How much crap do believe is out there actually *worth* saturating your connection/senses with? You're not going to pay more for it. That much makes sense.
Anyone who was going to post a web page/file/whatever-content to an Internet server probably already has, or can without too much difficulty. (The Third-World respectfully excluded). Wether you find what you thought you wanted or not, there is so much available to you that the bottleneck is in your frontal lobes (your ability to understand/comprehend) and not in the server and not in the connections.
Connectivity of the 300bps internet-phone variety will be extended to billions of new low-power nodes as flash ROM and CPU resources get cheap enough to make *everything* smart and connected. The Third-World *will* eventually reach similar per-capita node densities. The Internet will continue to grow, and it will fruit with practical examples of the theoretical capabilities (vaporware) people were foaming at the mouth about for the last two years.
I don't think anyone disagrees that tracking people by MAC address gives the supply side an edge in their marketing powers. What I think people are concerned about is the worst-case scenario of a more solid, focused, revenue model.
Adam Smith capitalism is supposed to be consumer driven, but lately we've been seeing power shift to the producers. That gets dangerously close to fascism when you see how government involvement plays in.
Flamebait!
This article is all snake oil. I'm disappointed that I read the whole thing and didn't learn anything about comparative DBMS architecture and implementations. The simply *is* no *how or why* to the results. What a useless piece of sales drivel! I feel no better equipped to decide on a favorite database backend.
Guess what?
Anonymity is impunity.The white stones beat the mark of the beast anyways.
It's not as bad as it could be because Rev 2:17 could be interpereted to mean peer-certified public-key cryptographic ID:
New English Translation Revelations 2:17:
The verse is one of the worst understood, and the footnotes suggest that the greek word translated as "white" here could mean "bright" as in glowing. Imagine a backlit PDA with your PK keyrings on it that you can use to recieve, sign, encrypt, and send irrefutable tamper-proof messages. Why, you could use that in lieu of the mark-of-the-beast at least within the network/trust-ring of the holy apocolyptic black-market (white market?).
Status Quo: It is possible for a person to redeem themselves for past crimes. A necessary condition of this is that the person must be able to dissociate themselves from the link to the past crime. Past crimes are recorded with a reference to a criminal's identity. If it is possible to live without revealing that information, then and only then is it in a person's interest to pay a debt they owe to society.
World of implanted ID chips: Nobody can ever redeem themselves, because *somebody* will always be around to steal official records before they are erased, store them, and resell the data on the black market. Your "credit report" will contain inane crap like your school kindergarten disciplinary record and how much beer you bought in college. You will not be able to control this because the black market will shadow any official records. Anyone you deal with will use this data to discriminate against you. Since civil law has the "preponderance of evidence" burden of proof, you will be at a legal disadvantage if you do not use others' data as leverage against them. If you ever make a mistake, no matter how small, you can look forward to paying for it repeatedly the rest of your life. The concept of "benefit of the doubt" will no longer exist, and all trust decisions will effectively be arbitrated by the authorities who sell the data.
Your comment seems to unfairly suggest a few things:
I hereby declare youre "broke and somewhat off your rocker" comment to be genuine FUD. Go FUD yourself.
I read a good book that helped me in the Unix learning curve. Unix is a system that adheres to principles, and is descended from the Original AT&T project.
BOYCOTT AMAZON
What makes you so sure they actually buy elections without the Internet? Maybe they coerce the vote-counters! Bullets are cheaper. Oh... You mean the *AMERICAN* presidential elections..
Maybe these early Internet voting schemes are straw-men to keep the old system (very opaque to the voters) in place? Maybe you're astroturfing for the Dept. of State!
Anyone who has discovered public-key crypto knows that Internet voting is both workable, desirable, and is *FAR* cheaper than punch-ballots if done *CORRECTLY*...
No name; no one to blame. This is how the Bill of Rights intended.
There will be no mandatory Internet user tracking. The reason is each thing you do with an Internet client is a REQUEST for information. Working this proposal would require practical violation of the First Amendment on two grounds, and possibly the Third Amendment on a third ground.
<rant>Bill Clinton should be ashamed that his middle name is Jefferson, and he will be recorded in history of attempting to sell out the rights American Citizen for short term political gain.</rant>
Requiring people to use a specific protocol which identifies them, abridges their right to make requests for information with impunity. It would also abridge (have a chilling effect) on an individual's likelihood of using such a request to establish an unpopular idea on a public bulletin board. It would also require agents of those services to enforce the protocol on the individual users, which constitutes abridging the freedom of the press.
IMnsHO, while the world considers the term cyber-warfare we should be prepared to establish the constitutional interperetation of quartering of cyber-soldiers in our homes. President Clinton's statement is preparatory for yet another key-escrow proposal. We need to get the idea out that escrowed keys are agents of the government, and are essentially cyber-soldiers which the President, and (thus far) potential Presidential Candidates Gore and McCain would like to install in our cyber-homes.
I agree with this guy.