There is something odd about the size of the difference you are finding. For comparision, here are the results I obtained running the same commands on a 1.5GHz G4 powerbook running OS X 10.4.4.
$ find . | wc 42851 90383 3169155 $ du |tail -n1 7528916. $ time gtar 2>/dev/null cf/dev/null. real 0m18.066s user 0m0.980s sys 0m4.581s
Is HFS+ journaling turned on? Do you have journaling enabled on the FreeBSD machine? Comparable system buses? Hard drive performance differences: rotaional speed, hard drive caching, transfer rates? What about file locality on each system?
There are plenty of unknown factors that may have skewed your test results.
The UK uses a combination of imperial and metric units. They switched from using pounds to kilograms at the end of 1995. However roads are still marked in miles and if I recall correctly they use centigrade for temperature.
The original poster is almost right. Windows as in Windows 3.0/3.1/95/98 and so on (including earlier unsuccessful releases) were not designed using a micro-kernel approach. Many would argue that they were not designed at all.
Windows NT 3.5 however was designed with micro-kernel architecture. Well, as close to a micro-kernel as Redmond has been able to achieve. However NT 3.5 suffered from significant performance issues.
Some of these issues were seen to be a result of the layered microkernel approach. To address this, NT 4.0 moved many of the upper layers down into layers 0 and 1 and provided the ability for driver writers to install services into the lower layers. This was great for performance reasons since now drivers could make direct calls rather than message passing to the kernel services. However removing the separation started the rot of stuffing more and more things into the lower layers effectively transforming a micro-kernel into a monolithic kernel. The fact that these extra things stuffed into the kernel space have varying levels of QA and compatibility testing caused a trend to reduced stability in the resulting OS.
If you read through the details on the irs tax freedom website you'll find that the legislation threat that is driving these free filing programs only requires a certain number of to be able to file freely. The aim was as you suggest that people with lower incomes would be the ones to benefit from this program.
However, one of the companies participating in the program decided that it would be too much work to implement income testing so decided to open it up for everyone. Once this happened, basically all of the companies involved in the scheme decided to make it available to everyone.
This whole free-filing scheme is a way for the tax preparation industry to play nice and thus avoid having legislation imposed on the industry. Thus they are not advertising the program widely, but they are making it available to everyone.
This was certainly true two weeks ago when I filed my return.
Slashdot Mirror
gnutar as installed by fink on my powerbook is version 1.15.1. (On my system gtar is a link to the gnutar binary.)
/usr/bin/gtar as installed as part of OS X 10.4.4 reports as version 1.14.
gtar is installed on my system as part of the fink package system.
There is something odd about the size of the difference you are finding. For comparision, here are the results I obtained running the same commands on a 1.5GHz G4 powerbook running OS X 10.4.4.
$ find . | wc . /dev/null .
42851 90383 3169155
$ du |tail -n1
7528916
$ time gtar 2>/dev/null cf
real 0m18.066s
user 0m0.980s
sys 0m4.581s
Is HFS+ journaling turned on? Do you have journaling enabled on the FreeBSD machine? Comparable system buses? Hard drive performance differences: rotaional speed, hard drive caching, transfer rates? What about file locality on each system?
There are plenty of unknown factors that may have skewed your test results.
The UK uses a combination of imperial and metric units. They switched from using pounds to kilograms at the end of 1995. However roads are still marked in miles and if I recall correctly they use centigrade for temperature.
The original poster is almost right. Windows as in Windows 3.0/3.1/95/98 and so on (including earlier unsuccessful releases) were not designed using a micro-kernel approach. Many would argue that they were not designed at all. Windows NT 3.5 however was designed with micro-kernel architecture. Well, as close to a micro-kernel as Redmond has been able to achieve. However NT 3.5 suffered from significant performance issues. Some of these issues were seen to be a result of the layered microkernel approach. To address this, NT 4.0 moved many of the upper layers down into layers 0 and 1 and provided the ability for driver writers to install services into the lower layers. This was great for performance reasons since now drivers could make direct calls rather than message passing to the kernel services. However removing the separation started the rot of stuffing more and more things into the lower layers effectively transforming a micro-kernel into a monolithic kernel. The fact that these extra things stuffed into the kernel space have varying levels of QA and compatibility testing caused a trend to reduced stability in the resulting OS.
If you read through the details on the irs tax freedom website you'll find that the legislation threat that is driving these free filing programs only requires a certain number of to be able to file freely. The aim was as you suggest that people with lower incomes would be the ones to benefit from this program.
However, one of the companies participating in the program decided that it would be too much work to implement income testing so decided to open it up for everyone. Once this happened, basically all of the companies involved in the scheme decided to make it available to everyone.
This whole free-filing scheme is a way for the tax preparation industry to play nice and thus avoid having legislation imposed on the industry. Thus they are not advertising the program widely, but they are making it available to everyone.
This was certainly true two weeks ago when I filed my return.