Failure to see applications is a failure of imagination. Also, Cassandra is highly optimized for extremely fast write rates at large scale, not read-only. Cassandra also has no SPOF, includes variable consistency guarantees, and many other great features. Specifically focusing on the fact that it can have 2 billion columns in a row glosses over a number of other things, but that hardly denormalized or non-transactional databases have few applications.
I use a branch current monitor from PowerLogic (BCM42) coupled with a Barionet-50 to perform monitoring of every circuit in the house simultaneously. Some details can be found here:
The kinds of analysis you can perform with this are tons of fun. For instance, how much does that desuperheater in my geothermal unit really help the water heater? Has a lightbulb in one of my floodlights gone out? Did somebody leave the refrigerator door open? Has the water softener stuck again causing the well pump to work overtime? This approach lets you monitor WAY more than just consumption.
I don't know that one needs to be as specific as neural prosthesis, but nearly every human on the planet is augmented in some way from the natural state, be it a pair of glasses or contacts, clothes to augment the insulating properties of skin, umbrellas to keep the rain off our heads, or air conditioning to stay cooler in warm climates. It seems that many people who refer to cyborg-like physical augmentation don't have a clear distinction between what would be considered a 'tool' and what would be considered an augmentation.
My personal opinion on the matter is that it is both impossible to define the distinction without being able to come up with exceptions on either side and not currently a relevant question to ask. I believe that if technology becomes so advanced that it is impossible to distinguish between a human with machine bits added and a machine with human bits added, the distinction will either not matter, or be readily apparent.
From the announcement e-mail on the -CURRENT mailing list.
[...] - many improvements and fixes to the ATA driver - new kdeadmin3 package to address the 'KUser' problem - fixes to several network drivers, IPSec, NFSv4, and NNS. - fix for the cd bootloader code to handle USB cdrom drives. [...]
As you see, most of the above fixes do not apply to ports/packages as they are in the base system.
Greylisting is fantastic. We are using it at one client who received ~15000k spams/month, and now they receive about 12 spams/week with nearly no false positives. Best spam deterrent yet (including Bayes, which solves the wrong problem)
An interesting alternative that was mentioned on slashdot a while back was graylisting. Having implemented a modified version of it on several mail servers, my experience has been massively positive. In a company of about 70 people, graylisting has allowed through maybe 10 pieces of spam in nearly a month (where before 70% of all e-mail traffic was spam) and the only false positives are from large mailing list clusters like groups.yahoo and the like, which are easily whitelisted. While this sort of maintenance on the server side may seem cumbersome to some administrators, it's definitely worth it, costs the spammer storage/time, and wastes only the bandwidth/storage of the token on the receiving side.
Also, unlike many spam filters (Paul Graham's new active method), it doesn't need to be used by everybody to work it's magic.
One of my company's clients is a broker/dealer who is crazy about compliance and has been logging all IM conversations for months now. They have an OpenBSD firewall running a little ruby+pf+ethereal+snort script to detect all IM activity and log it to sorted files, but you could probably do this sort of thing on any box. The only thing it can't log thus far is SSL'd jabber. (Which can log on it's own)
I had linux running on a dual pentium-233 on a Tyan Tomcat IV motherboard my sophomore year of college, which began August of '98. I remember being very exciting every time a new dev kernel was released because SMP was under heavy dev by Alan Cox then. (if I remember correctly).
Slashdot Mirror
Failure to see applications is a failure of imagination. Also, Cassandra is highly optimized for extremely fast write rates at large scale, not read-only. Cassandra also has no SPOF, includes variable consistency guarantees, and many other great features. Specifically focusing on the fact that it can have 2 billion columns in a row glosses over a number of other things, but that hardly denormalized or non-transactional databases have few applications.
I use a branch current monitor from PowerLogic (BCM42) coupled with a Barionet-50 to perform monitoring of every circuit in the house simultaneously. Some details can be found here:
http://blog.insidesystems.net/modbusrtu-via-tcp-serial-gateway-with-ruby
The kinds of analysis you can perform with this are tons of fun. For instance, how much does that desuperheater in my geothermal unit really help the water heater? Has a lightbulb in one of my floodlights gone out? Did somebody leave the refrigerator door open? Has the water softener stuck again causing the well pump to work overtime? This approach lets you monitor WAY more than just consumption.
I don't know that one needs to be as specific as neural prosthesis, but nearly every human on the planet is augmented in some way from the natural state, be it a pair of glasses or contacts, clothes to augment the insulating properties of skin, umbrellas to keep the rain off our heads, or air conditioning to stay cooler in warm climates. It seems that many people who refer to cyborg-like physical augmentation don't have a clear distinction between what would be considered a 'tool' and what would be considered an augmentation.
My personal opinion on the matter is that it is both impossible to define the distinction without being able to come up with exceptions on either side and not currently a relevant question to ask. I believe that if technology becomes so advanced that it is impossible to distinguish between a human with machine bits added and a machine with human bits added, the distinction will either not matter, or be readily apparent.
From the announcement e-mail on the -CURRENT mailing list.
[...]
- many improvements and fixes to the ATA driver
- new kdeadmin3 package to address the 'KUser' problem
- fixes to several network drivers, IPSec, NFSv4, and NNS.
- fix for the cd bootloader code to handle USB cdrom drives.
[...]
As you see, most of the above fixes do not apply to ports/packages as they are in the base system.
Greylisting is fantastic. We are using it at one client who received ~15000k spams/month, and now they receive about 12 spams/week with nearly no false positives. Best spam deterrent yet (including Bayes, which solves the wrong problem)
An interesting alternative that was mentioned on slashdot a while back was graylisting. Having implemented a modified version of it on several mail servers, my experience has been massively positive. In a company of about 70 people, graylisting has allowed through maybe 10 pieces of spam in nearly a month (where before 70% of all e-mail traffic was spam) and the only false positives are from large mailing list clusters like groups.yahoo and the like, which are easily whitelisted. While this sort of maintenance on the server side may seem cumbersome to some administrators, it's definitely worth it, costs the spammer storage/time, and wastes only the bandwidth/storage of the token on the receiving side.
Also, unlike many spam filters (Paul Graham's new active method), it doesn't need to be used by everybody to work it's magic.
One of my company's clients is a broker/dealer who is crazy about compliance and has been logging all IM conversations for months now. They have an OpenBSD firewall running a little ruby+pf+ethereal+snort script to detect all IM activity and log it to sorted files, but you could probably do this sort of thing on any box. The only thing it can't log thus far is SSL'd jabber. (Which can log on it's own)
Daunting, pushaw.
I had linux running on a dual pentium-233 on a Tyan Tomcat IV motherboard my sophomore year of college, which began August of '98. I remember being very exciting every time a new dev kernel was released because SMP was under heavy dev by Alan Cox then. (if I remember correctly).