Actually the technology works pretty well if properly applied! Like in all AI systems the ruling principal is "Junk in Junk out".
I am not sure what the quality factors considered by the often ill-informed journalists were and I am not sure what the pictures uploaded by the/.ers look like but I would guess there is a difference in orientation, lighting, as well as size.
The technology originated in the computer vision labs at MIT and if the orientation and lighting as well as size of the images are normalized they have achieved verification rates in excess of 98% on subject sets exceeding 3000 people.
Unfortunately it seems that in this case the journalists fed un-normalised faces into the system which means that the eigenvectors (or eigenfaces) are classified according to the size and orientation of the images rather than facial features. The results you saw are the telltale signs of this!
You find the terrorist, hunt them down, and get them - you don't randomly start throwing missiles at other countries.
Gee! That should have been rated FUNNY and not INFORMATIVE!
Like you mean the US has not been randomly throwing missiles at Afghanistan, Iraq and who was next again? I think it was North Korea, Siria and then Pakistan? Didn't get the whole list though and I might have messed up the order a little, sorry.
And oh yes, they did of course get both Bin Laden and Sadam as I recall, no wait something is wrong here...
I always believed that the MHz war would stop short of 2.4GHz as that is the absorbtion band that Microwave ovens operate in.
The latest 2.4GHz machines I have a couple of questions. 1) Does it emit 2.4GHz Microwaves and how many? 2) Do these fry your brains while sitting close to the machine? (Safety tests?) 3) Will it reboot if I superheat my cup of coffee in the microwave across the room?
Hmmmmm.... That is definately worth an experiment!
These guys seem to claim the noise is good for you! I quote the bit of interest for the guys to busy to read the whole thing.
A recent advancement in technology that is becoming widely used in psychological counseling and health service settings is the random noise generator. These devices, similar to the size and configuration of a smoke detector and/or an air purification device, emit a wide frequency band described as "white" or "pink noise." Adjusted to a relatively low level, these can be effectively utilized in the spaces in which the client interaction occurs to mask undesirable environmental sounds without negatively impacting the client interactions by being intrusive in nature.
AFAIK the noise from a fan is pretty close to white.
Seems that in medicine though there are always conflicting studies so I expect to be presented with the opposite result in a case study.
I must agree with your comment. This was however before the Win2K project according to what I have seen.
It is widely accepted that Win2000 is much more stable and controlled than any other version of windows has been.
There are a lot of assumptions in the CMM system. One of these is that you actually follow things you have documented.
What people seem to miss is the fact that CMM is about the capability to perform reliably and repeatably as an organisation. I have stopped counting the number of places I worked where there is no formal way of doing things. Everybody follows his own head. This is what the CMM is addressing. It attempts to put structure and process in place an that Microsoft definately have done.
If you are up for a challenge I would suggest you take the CMM documentation and compare that to the MSF process used by Microsoft.
Have a look at their processes for fixing problems, making releases and dealing with customers.
This is what the CMM is about, the process as a whole and not writing a 10-line perl regexp and getting it to work...
I know of very little cases where a bug, fixed in some version of Windows e.g. re-surfaced at a later stage. On the smaller CMM level 1 projects I have worked on it happens more often than not!
I will be prepared to put money that Microsoft rates at least a 3 and I would rate them a 4 for the Windows 2000 and XP projects.
These projects were the biggest software projects ever undertaken and both a success. Acording to the big consulting firms at least 80% of projects fail and thus Microsoft are not doing too badly I would say...
To answer your questions: 1) Yes and 2) Sorry but I have to site Microsoft Windows and point out that Microsoft does definately (by the CMM definition) score at least 3.
The value of Software Engineering as opposed to programming is only perceivable if you have done the prior and experienced the latter. I have found that all projects with budgets over $5 Million with more than 100 developers can only succeed if proper engineering practices are followed.
I inevitably ask people who oppose the model two questions. 1) Have you ever worked on a project of such scale (one with a budget, not open-ended-open-source)? 2) Have you read the book (Mythical Man Month)?
If they answer no to one I answer "Interesting!" and no to two I answer "Go read it and lets then continue this discussion."
I would like to know the answer of your CMU buddies and yourself to these (out of curiosity of course)
According to the Capability Maturity model the only way a company on level 1 - Initial (also called chaos) can be a success is if there are so-called super-programmers saving but most of the time.
In my experience of working on open-source most projects are governed by chaos and judging them according to the CMM there cannot be a single open source project on anything above level 1.
This leaves one to, by logical deduction, assume that open source projects can only be a success if there are a couple of super-human programmers involved.
I think most of you guys will agree that this has been proven in practice over and over again.
Unfortunately due to the eccentric nature of these extraordinary programmers it is very seldom that more than 2-3 of them can agree and co-exist without trying to prove themselves superior.
If it was possible to move a opensource project to level 3 or 4 and get a team of say 100 programmers working on it we could tackele a large project in a small timeframe and thus speed up progress by a couple of orders in magnitude.
According to F.P. Brooks in The Mythical Man Month this has been one of the major limitations in getting big projects done - they need big teams to be complete before they are obsolte - and this is the area where opensource needs some work.
For GPS satellites, GR predicts that the atomic clocks at GPS orbital altitudes will tick faster by about 45,900 ns/day because they are in a weaker gravitational field than atomic clocks on Earth's surface. Special Relativity (SR) predicts that atomic clocks moving at GPS orbital speeds will tick slower by about 7,200 ns/day than stationary ground clocks. Rather than have clocks with such large rate differences, the satellite clocks are reset in rate before launch to compensate for these predicted effects. In practice, simply changing the international definition of the number of atomic transitions that constitute a one-second interval accomplishes this goal. Therefore, we observe the clocks running at their offset rates before launch. Then we observe the clocks running after launch and compare their rates with the predictions of relativity, both GR and SR combined. If the predictions are right, we should see the clocks run again at nearly the same rates as ground clocks, despite using an offset definition for the length of one second.
We note that this post-launch rate comparison is independent of frame or observer considerations. Since the ground tracks repeat day after day, the distance from satellite to ground remains essentially unchanged. Yet, any rate difference between satellite and ground clocks continues to build a larger and larger time reading difference as the days go by. Therefore, no confusion can arise due to the satellite clock being located some distance away from the ground clock when we compare their time readings. One only needs to wait long enough and the time difference due to a rate discrepancy will eventually exceed any imaginable error source or ambiguity in such comparisons.
This in the other hand sound pretty scientific and conclusive to me...
What I fail to understand is why the database was password protected in the first place!
If it is in a museum then everybody had to have free access to the information. If the database is purely password protected for writing (as in read-only) I would understand it as well!
Maybe it is one of those Microsoft Word pop-ups where it asks you a password if you want to open the file for writing. I have stopped counting the number of people mailing me for the password for documents 'cause they cannot read it, turning out the password was not required at all!
BTW DBase is so old and the algorithms used for protection are all well-known. They could not be using anything better than DES and the average 2.4GHz machine can crack that in a couple of weeks nowadays so I fail to grasp the crisis.
AFAIK the current GPS satelite system makek adjustments for relativity in the signals it is sending around and they have been adjusting for this for years.
See the articles at Metaresearch and
lsu.edu for more info.
And that this self same well-known company happens to be in posession of the private keys of a large number of commercial servers.
If they climbed into bed with the FED's, as it seems they already have, not even encrypted IP conversations will be safe.
The way the world is changeing the next step will be ordering of wire taps on internet connections, even SSL ones, and this the government will only be able to do in conjunction with the only bunch with the key to unlock the conversations.
It is pretty easy to tap into a SSL or IPSEC session if you have the private keys of both the individuals!
I cannot help but wonder how usefull the efforts put into this service is going to be over the next 3-5 years during the take-over of the world by broadband and IP telephony. With the dissapointing earnings produced by all the major telco's they are all putting a lot of effort into getting IP telephony going in order to boost sales.
I must say this whole thing is going to let me think twice about that Verisign Certificate I bought which only I have the private key for...
I guess the moment we have our SSL encrypted, fully fledged PKI infrastructure based IP telephony system up and running Verisign will be selling our Private keys to the highest bidder!
Now if you take that into account this is not all that far off the Business Model that Verisign has been following...
Maybe they are just one step ahead of the rest of the pack!
Generally people are quick to indemnify themselves from everything that may ever happen but usually these statements are not worth the paper they are printed on in a court of law.
Now I am no lawyer but would Microsoft for example be indemnified from the antitrust allegations if they put a clause to that effect in their licence agreement?
More often than not copyright notices and licence agreements are there purely for FUD purposes. I have always seriously doubted the legal grounds a company has to stand on if they claim things in a license agreement which nobody really reads, seeks legal council on or sign.
I would go out and buy one and claim that I never received the licence agreement! Would they then have to prove that I received, read and agreed to it before they can take further steps?
Come-on you Law-infested-geeks out there! What is the answer?
I was surprised to see the sizes noted by people as large. It seems that the Linux (or at least the/.) community are not really big on RDBMS systems!
If they were they would know about the TPC benchmarks which frequently refer to terabyte sized databases.
Generally speaking Linux systems work well, usually better than comercial systems, if the databases are very small. This again generally due to the fact that they were optimized to work on small configurations, which is common but not the norm in enterprize.
We use Sybase on a RS6000 (IBM S85) with a EMC^2 storage array delivering aroung 80GB. Our system is small compared to others I have been exposed to. As for 10 000 articles, we have over 2000 tables and many of them have more than 20Million records.
We had a project to test feasibility of Sybase on Linux for our system (although it was 3 years ago), which failed miserably and today we still use AIX, IBM RS6000's and Sybase...
I think Google is probably the only enterprise size RDBMS on Linux in existence and it is a success because it consists of 8000 or so small systems in the Gigabyte range clevely made to work together. It is also all proprietary, as well as a sponsored research project.
I was surprised to see the sizes noted by people as large. It seems that the Linux (or at least the/.) community are not really big on RDBMS systems!
If they were they would know about the TPC benchmarks which frequently refer to terabyte sized databases.
Generally speaking Linux systems work well, usually better than comercial systems, if the databases are very small. This again generally due to the fact that they were optimized to work on small configurations, which is common but not the norm in enterprize.
We use Sybase on a RS6000 (IBM S85) with a EMC^2 storage array delivering aroung 80GB. Our system is small compared to others I have been exposed to. As for 10 000 articles, we have over 2000 tables and many of them have more than 20Million records.
We had a project to test feasibility of Sybase on Linux for our system (although it was 3 years ago), which failed miserably and today we still use AIX, IBM RS6000's and Sybase...
According to me the challenge was aimed at proving that it was impossible to build a perfect decompressor and that all compressors make use of redundency in data. This is a dangerous assumption to make for this challenge ! If the challenge required 2 files to be compressed it would have been much closer to being safe.
Consider this:
0x23 ^ 0x34 = 0x77B352C55E214991
Thus we could "compress" the 64bit binary data to 24bit binary data by storing only 0x239434. This indicates 2x8 bit numbers separated with a mathematical symbol. This leaves 56 bits for code, etc/64 bits of data -> 87.5% of data size for code in the decompressor.
The point being that if the decompressor need only work on a single dataset there are usually cunning methods of representing the data. Such a decompressor will not be usable to compress day-to-day data (as is well known from compression theory) but the number of possibilities to find a pattern that is known to exist in a file increases the probablity of meeting such a challenge greatly.
Slashdot Mirror
Actually the technology works pretty well if properly applied! Like in all AI systems the ruling principal is "Junk in Junk out".
/.ers look like but I would guess there is a difference in orientation, lighting, as well as size.
I am not sure what the quality factors considered by the often ill-informed journalists were and I am not sure what the pictures uploaded by the
The technology originated in the computer vision labs at MIT and if the orientation and lighting as well as size of the images are normalized they have achieved verification rates in excess of 98% on subject sets exceeding 3000 people.
Unfortunately it seems that in this case the journalists fed un-normalised faces into the system which means that the eigenvectors (or eigenfaces) are classified according to the size and orientation of the images rather than facial features. The results you saw are the telltale signs of this!
You find the terrorist, hunt them down, and get them - you don't randomly start throwing missiles at other countries.
...
Gee! That should have been rated FUNNY and not INFORMATIVE!
Like you mean the US has not been randomly throwing missiles at Afghanistan, Iraq and who was next again? I think it was North Korea, Siria and then Pakistan? Didn't get the whole list though and I might have messed up the order a little, sorry.
And oh yes, they did of course get both Bin Laden and Sadam as I recall, no wait something is wrong here
I always believed that the MHz war would stop short of 2.4GHz as that is the absorbtion band that Microwave ovens operate in.
.... That is definately worth an experiment!
The latest 2.4GHz machines I have a couple of questions.
1) Does it emit 2.4GHz Microwaves and how many?
2) Do these fry your brains while sitting close to the machine? (Safety tests?)
3) Will it reboot if I superheat my cup of coffee in the microwave across the room?
Hmmmmm
These guys seem to claim the noise is good for you! I quote the bit of interest for the guys to busy to read the whole thing.
A recent advancement in technology that is becoming widely used in psychological counseling and health service settings is the random noise generator. These devices, similar to the size and configuration of a smoke detector and/or an air purification device, emit a wide frequency band described as "white" or "pink noise." Adjusted to a relatively low level, these can be effectively utilized in the spaces in which the client interaction occurs to mask undesirable environmental sounds without negatively impacting the client interactions by being intrusive in nature.
AFAIK the noise from a fan is pretty close to white.
Seems that in medicine though there are always conflicting studies so I expect to be presented with the opposite result in a case study.
Have you never tried strapping that to your back sitting on a scateboard?
I am really dissapointed in you.
Yes and if speed is not of the essence you can just switch it off and relax in the peace and quiet that follows !
That is sure one way of having your CGI proxi /.ted. I bet that went straight into a GOOGLE of bookmark lists!
I must agree with your comment. This was however before the Win2K project according to what I have seen.
It is widely accepted that Win2000 is much more stable and controlled than any other version of windows has been.
There are a lot of assumptions in the CMM system. One of these is that you actually follow things you have documented.
What people seem to miss is the fact that CMM is about the capability to perform reliably and repeatably as an organisation. I have stopped counting the number of places I worked where there is no formal way of doing things. Everybody follows his own head. This is what the CMM is addressing. It attempts to put structure and process in place an that Microsoft definately have done.
If you are up for a challenge I would suggest you take the CMM documentation and compare that to the MSF process used by Microsoft.
...
...
Have a look at their processes for fixing problems, making releases and dealing with customers.
This is what the CMM is about, the process as a whole and not writing a 10-line perl regexp and getting it to work
I know of very little cases where a bug, fixed in some version of Windows e.g. re-surfaced at a later stage. On the smaller CMM level 1 projects I have worked on it happens more often than not!
I will be prepared to put money that Microsoft rates at least a 3 and I would rate them a 4 for the Windows 2000 and XP projects.
These projects were the biggest software projects ever undertaken and both a success. Acording to the big consulting firms at least 80% of projects fail and thus Microsoft are not doing too badly I would say
To answer your questions:
1) Yes
and
2) Sorry but I have to site Microsoft Windows and point out that Microsoft does definately (by the CMM definition) score at least 3.
The value of Software Engineering as opposed to programming is only perceivable if you have done the prior and experienced the latter. I have found that all projects with budgets over $5 Million with more than 100 developers can only succeed if proper engineering practices are followed.
I inevitably ask people who oppose the model two questions.
1) Have you ever worked on a project of such scale (one with a budget, not open-ended-open-source)?
2) Have you read the book (Mythical Man Month)?
If they answer no to one I answer "Interesting!" and no to two I answer "Go read it and lets then continue this discussion."
I would like to know the answer of your CMU buddies and yourself to these (out of curiosity of course)
According to the Capability Maturity model the only way a company on level 1 - Initial (also called chaos) can be a success is if there are so-called super-programmers saving but most of the time.
In my experience of working on open-source most projects are governed by chaos and judging them according to the CMM there cannot be a single open source project on anything above level 1.
This leaves one to, by logical deduction, assume that open source projects can only be a success if there are a couple of super-human programmers involved.
I think most of you guys will agree that this has been proven in practice over and over again.
Unfortunately due to the eccentric nature of these extraordinary programmers it is very seldom that more than 2-3 of them can agree and co-exist without trying to prove themselves superior.
If it was possible to move a opensource project to level 3 or 4 and get a team of say 100 programmers working on it we could tackele a large project in a small timeframe and thus speed up progress by a couple of orders in magnitude.
According to F.P. Brooks in The Mythical Man Month this has been one of the major limitations in getting big projects done - they need big teams to be complete before they are obsolte - and this is the area where opensource needs some work.
According to me that is - of course
And also from one of the links:
For GPS satellites, GR predicts that the atomic clocks at GPS orbital altitudes will tick faster by about 45,900 ns/day because they are in a weaker gravitational field than atomic clocks on Earth's surface. Special Relativity (SR) predicts that atomic clocks moving at GPS orbital speeds will tick slower by about 7,200 ns/day than stationary ground clocks. Rather than have clocks with such large rate differences, the satellite clocks are reset in rate before launch to compensate for these predicted effects. In practice, simply changing the international definition of the number of atomic transitions that constitute a one-second interval accomplishes this goal. Therefore, we observe the clocks running at their offset rates before launch. Then we observe the clocks running after launch and compare their rates with the predictions of relativity, both GR and SR combined. If the predictions are right, we should see the clocks run again at nearly the same rates as ground clocks, despite using an offset definition for the length of one second.
We note that this post-launch rate comparison is independent of frame or observer considerations. Since the ground tracks repeat day after day, the distance from satellite to ground remains essentially unchanged. Yet, any rate difference between satellite and ground clocks continues to build a larger and larger time reading difference as the days go by. Therefore, no confusion can arise due to the satellite clock being located some distance away from the ground clock when we compare their time readings. One only needs to wait long enough and the time difference due to a rate discrepancy will eventually exceed any imaginable error source or ambiguity in such comparisons.
This in the other hand sound pretty scientific and conclusive to me
What I fail to understand is why the database was password protected in the first place!
If it is in a museum then everybody had to have free access to the information. If the database is purely password protected for writing (as in read-only) I would understand it as well!
Maybe it is one of those Microsoft Word pop-ups where it asks you a password if you want to open the file for writing. I have stopped counting the number of people mailing me for the password for documents 'cause they cannot read it, turning out the password was not required at all!
BTW DBase is so old and the algorithms used for protection are all well-known. They could not be using anything better than DES and the average 2.4GHz machine can crack that in a couple of weeks nowadays so I fail to grasp the crisis.
And for the slow ones out there, how does a satelite (such as the ISS) stay in orbit?
Might it just have something to do with gravity?
AFAIK the current GPS satelite system makek adjustments for relativity in the signals it is sending around and they have been adjusting for this for years. See the articles at Metaresearch and lsu.edu for more info.
Bet they will use GPG with a Verisign provided key!
Bet VERISIGN will not be able to decrypt that !
And that this self same well-known company happens to be in posession of the private keys of a large number of commercial servers.
If they climbed into bed with the FED's, as it seems they already have, not even encrypted IP conversations will be safe.
The way the world is changeing the next step will be ordering of wire taps on internet connections, even SSL ones, and this the government will only be able to do in conjunction with the only bunch with the key to unlock the conversations.
It is pretty easy to tap into a SSL or IPSEC session if you have the private keys of both the individuals!
I cannot help but wonder how usefull the efforts put into this service is going to be over the next 3-5 years during the take-over of the world by broadband and IP telephony. With the dissapointing earnings produced by all the major telco's they are all putting a lot of effort into getting IP telephony going in order to boost sales.
...
...
I must say this whole thing is going to let me think twice about that Verisign Certificate I bought which only I have the private key for
I guess the moment we have our SSL encrypted, fully fledged PKI infrastructure based IP telephony system up and running Verisign will be selling our Private keys to the highest bidder!
Now if you take that into account this is not all that far off the Business Model that Verisign has been following
Maybe they are just one step ahead of the rest of the pack!
Generally people are quick to indemnify themselves from everything that may ever happen but usually these statements are not worth the paper they are printed on in a court of law.
Now I am no lawyer but would Microsoft for example be indemnified from the antitrust allegations if they put a clause to that effect in their licence agreement?
More often than not copyright notices and licence agreements are there purely for FUD purposes. I have always seriously doubted the legal grounds a company has to stand on if they claim things in a license agreement which nobody really reads, seeks legal council on or sign.
I would go out and buy one and claim that I never received the licence agreement! Would they then have to prove that I received, read and agreed to it before they can take further steps?
Come-on you Law-infested-geeks out there! What is the answer?
Congratulations! You just managed to hack the server in question it seems :0)
...
Now is it not interesting that this got posted right past our demigod moderators.
Guess slashdot CAN be hacked afterall
I was surprised to see the sizes noted by people as large. It seems that the Linux (or at least the /.) community are not really big on RDBMS systems!
...
If they were they would know about the TPC benchmarks which frequently refer to terabyte sized databases.
Generally speaking Linux systems work well, usually better than comercial systems, if the databases are very small. This again generally due to the fact that they were optimized to work on small configurations, which is common but not the norm in enterprize.
We use Sybase on a RS6000 (IBM S85) with a EMC^2 storage array delivering aroung 80GB. Our system is small compared to others I have been exposed to. As for 10 000 articles, we have over 2000 tables and many of them have more than 20Million records.
We had a project to test feasibility of Sybase on Linux for our system (although it was 3 years ago), which failed miserably and today we still use AIX, IBM RS6000's and Sybase
I think Google is probably the only enterprise size RDBMS on Linux in existence and it is a success because it consists of 8000 or so small systems in the Gigabyte range clevely made to work together. It is also all proprietary, as well as a sponsored research project.
I was surprised to see the sizes noted by people as large. It seems that the Linux (or at least the /.) community are not really big on RDBMS systems!
...
If they were they would know about the TPC benchmarks which frequently refer to terabyte sized databases.
Generally speaking Linux systems work well, usually better than comercial systems, if the databases are very small. This again generally due to the fact that they were optimized to work on small configurations, which is common but not the norm in enterprize.
We use Sybase on a RS6000 (IBM S85) with a EMC^2 storage array delivering aroung 80GB. Our system is small compared to others I have been exposed to. As for 10 000 articles, we have over 2000 tables and many of them have more than 20Million records.
We had a project to test feasibility of Sybase on Linux for our system (although it was 3 years ago), which failed miserably and today we still use AIX, IBM RS6000's and Sybase
According to me the challenge was aimed at proving that it was impossible to build a perfect decompressor and that all compressors make use of redundency in data. This is a dangerous assumption to make for this challenge ! If the challenge required 2 files to be compressed it would have been much closer to being safe.
Consider this:
0x23 ^ 0x34 = 0x77B352C55E214991
Thus we could "compress" the 64bit binary data to 24bit binary data by storing only 0x239434. This indicates 2x8 bit numbers separated with a mathematical symbol. This leaves 56 bits for code, etc/64 bits of data -> 87.5% of data size for code in the decompressor.
The point being that if the decompressor need only work on a single dataset there are usually cunning methods of representing the data. Such a decompressor will not be usable to compress day-to-day data (as is well known from compression theory) but the number of possibilities to find a pattern that is known to exist in a file increases the probablity of meeting such a challenge greatly.
Am I just crazy or is this a possibilty?