My father and I are together on three different banks and they all support Linux and different browsers. And this country I live in (Belgium) is rather small.
I think that you should try to find the short story "Assault on a city" from Jack Vance (and no, this is not OT, I read it a few days ago and he basically makes the same statement, but the story is from 1966).
Just to show that it depends upon the application you need to run.
Now, you will not hear me say that you should not use ReiserFS, for a desktop it is probably the best choice, but for servers, please think again.
In addition to that, my own experiences with ReiserFS are mostly positive, especially on my 233 Mhz laptop, but I have also a big system with a Promise SX6000 raid controller, where I had a partition with ReiserFS, ext3 and JFS using Red Hat 9. Everytime I did file operations using ReiserFS I got problems with the consistency of my RAID 5 array, so I scrapped ReiserFS and only kept ext3 and JFS, which give me no problems anymore.
Well, I think it has become more like "make each library do one thing and do it well".
A a certain point in time, you will have to integrate those tools into an application. Now if you do this using separate programs and a script, or separate libraries and a program, what is the difference ?
In addition to the European software patent legislation, I think I know why the current chairman tries to push software patents.
Ireland is since a long time the base for much software corporations. I think that he wants software patents to make Ireland the center of software development in Europe and thus create a constant revenue stream for one country on everything that has to do with software in Europe.
It seems that the original thread was not about swapping in or out, but about the amount of cache that is used by the kernel.
I just have the same problem. I have 2G RAM, and I run my KDE desktop, some standard server programs and some UML instances.
When I create UML instances (eg. 8G image) then my memory gets full and is not easily reclaimed.
I agree with the philosophy of the buffers and the cache, to speed up IO operations for recently accessed files, but I do not agree with the time that they are in memory.
I do not know if the VM subsystem knows about the three kinds of memory in use, but I think it should, and I also think that it should first of all free up buffers and cache memory before starting to swap out applications.
For me this just means that if you look at memory usage with any application, that after some time you should see the amount of memory used as filesystem cache gradually decrease over time (ie. when not written to, the system should just mark those pages as free).
What is not generally known, is that somewhere in the nineties an ancient mosaic was found wich showed the harbour city of Atlantis in the old volcano before it exploded.
I haven't read Plato, but I surmise that his description of Atlantis correlates with this picture.
If anybody has read 'Dancer From Atlantis' from Poul Anderson, he uses the same description.
And all this leads to the conclusion that Atlantis was the mainstay of the fleet of Crete, around 1600 BC.
It is not because the official installation of Debian is four years behind that you have to run older software.
Production setting : keep a local mirror of sarge and update it every three months. Test the updated software, upgrade production servers from the local mirror.
:-) I run Debian over Red Hat on my flashy new server, because Red Hat is able to control my Promise SX6000 controller, but Debian isn't.
I bought for my father the cheapest PC I could find, I installed Libranet 2.7.
You know what ? Everything worked fine from step 1 : sound, USB, video. Afterwards I had to install his printer : a HP 710C colour printer, which is a Winprinter. Works flawless.
Along similar usability lines, I am currently running Xandros as a pilot to see if it is "mom" (or grandmom) ready as it is advertised as "very user friendly". In the process of this pilot, I've done detailed documentation of every step I've done to get my Xandros fully working. My hope is that these step-by-step instructions will help my various friends who seem interested, but are a bit scared of Linux.
I did this already with my father. Although a little bit memory hungry, I opted for KDE on his 256Mb, 2.0Ghz white box, installed Xandros, used the Debian CD's to have a larger list of applications and installed StarOffice 7.0 for him.
What can he do ? Everything he needs to, but I need more documentation for him.
Camera mounting via a shortcut on his desktop
Printing on his HP 710C colour printer
Using StarOffice, GIMP, etc...
Use Sylpheed for mail
Added Mozilla 1.6
Surfing via modem
I also connected his old PC using Samba for those things that are really not possible on his new PC : he has an unsupported scanner, he has a dia scanner on loan, but the SCSI card is ISA, he has an organiser on his serial port (maybe the software can run under Wine, but I have no time to learn Wine).
The microsoft patents aren't about all of the FAT file system. Rather the method of storing short and long file names in the FAT file system.
IIRC, then something like this was already done in OS/2 3.0, of which I bought a copy in 1994. OS/2 could work on HPFS or FAT, and on the FAT system you had the same possibilities (albeit slower) like on the HPFS system.
I also seem to remember that the files which store the extended attributes had the same names on OS/2, as on Win95.
Saturday 27 March 2004 I created such a setup for my father. Most of the things that he wants to do know, he can do on Linux (I use Libranet + Debian updates), but he has a scanner which is not supported under Linux, a simple organizer which must exchange data through a Windows program, and a slide scanner, which uses an old Adaptec SCSI card in an ISA slot.
I set up his Windows computer headless, Win98, and you must use TweakUI for the system to automatically logon on the network. You must also disable the Stop on No Keyboard in the BIOS.
I installed tightVNC, which automatically starts up as a service. I also setup Samba to export the home directory, so that everything that is being saved on the Windows computer, must always be done on the mapped drive.
most modern COBOL compilers can relax on the positioning requirements
most modern COBOL compilers can also take it easier on the red tape
COBOL is still the best language to make computations on money, because it has integrated fixed point types.
It seems that ADA maybe also has that last possibility, but I do not know if their fixed-point data types are based upon a float library or a real integer library. In COBOL these are absolutely based upon integer computations.
All other Unix/C derived languages do not have the possibility to make financial computations like COBOL does.
That is because most managers have never made an effort to compute what it would cost them if they should lose their data, and also do not have a clue what a downtime costs them.
I have worked with someone who had done these figures, and he was never happy with downtimes, to say the least.
In general, though, the more 'realistic' it is, the less spectacular it's going to be. *shrug*
Probably, because movie- and seriesmakers will more go for effect than realism.
I think that the most realistic space wars have probably been written by Larry Niven (Protector), CJ Cherryh (Downbelow station), Poul Anderson (The Space Fox), Joe Haldeman (The Forever War).
The scale of these things is too large and too slow for a movie, but they provide fine strategic thinking, which is mostly absent in movies or series.
If you take and use Gauss theory on the bell curve and apply it, then after cutting out the dead wood and replacing them, you should get a finer and higher curve.
I think this is typical for people who do not know anything about math or science. The bell curve is something that you derive from measurements, doing the inverse, putting your bell curve upon your measurements does not make sense.
The algorithm to use is to first do your measurements, compute your curve and then discard the measurements which fall outside your range. But normally you would also discard measurements which are too high, which would mean that you should discard also someone who has too high a productivity. As this make no sense, measurements should be redone.
You can use it on people, and if you apply this correctly then you will have a bell curve, and maybe there will be people who fall behind. But if you replace these with others who ranks more near the middle, then your bell curve will contract and shift to the right.
Again, this way of working is purely devised by evil Catbert people, who do not understand anything of mathematics.
Slashdot Mirror
That's rubbish.
My father and I are together on three different banks and they all support Linux and different browsers. And this country I live in (Belgium) is rather small.
I think that you should try to find the short story "Assault on a city" from Jack Vance (and no, this is not OT, I read it a few days ago and he basically makes the same statement, but the story is from 1966).
Jurgen
Enlightenment with the Aqua Theme ! Fantastic.
Talking about Athens, maybe we should setup a campaign to reintroduce ostracism ?
Hey, hey, I thought that we Belgians were the number 1 XTC exporters !!
Yeah, I am heating my attic with one.
Yeah, next step : M$ patents glass!
Please read this
Just to show that it depends upon the application you need to run.
Now, you will not hear me say that you should not use ReiserFS, for a desktop it is probably the best choice, but for servers, please think again.
In addition to that, my own experiences with ReiserFS are mostly positive, especially on my 233 Mhz laptop, but I have also a big system with a Promise SX6000 raid controller, where I had a partition with ReiserFS, ext3 and JFS using Red Hat 9. Everytime I did file operations using ReiserFS I got problems with the consistency of my RAID 5 array, so I scrapped ReiserFS and only kept ext3 and JFS, which give me no problems anymore.
Well, I think it has become more like "make each library do one thing and do it well".
A a certain point in time, you will have to integrate those tools into an application. Now if you do this using separate programs and a script, or separate libraries and a program, what is the difference ?
In addition to the European software patent legislation, I think I know why the current chairman tries to push software patents.
Ireland is since a long time the base for much software corporations. I think that he wants software patents to make Ireland the center of software development in Europe and thus create a constant revenue stream for one country on everything that has to do with software in Europe.
Well, I think I read that about the same time in other electronics magazines.
It seems that the original thread was not about swapping in or out, but about the amount of cache that is used by the kernel.
I just have the same problem. I have 2G RAM, and I run my KDE desktop, some standard server programs and some UML instances.
When I create UML instances (eg. 8G image) then my memory gets full and is not easily reclaimed.
I agree with the philosophy of the buffers and the cache, to speed up IO operations for recently accessed files, but I do not agree with the time that they are in memory.
I do not know if the VM subsystem knows about the three kinds of memory in use, but I think it should, and I also think that it should first of all free up buffers and cache memory before starting to swap out applications.
For me this just means that if you look at memory usage with any application, that after some time you should see the amount of memory used as filesystem cache gradually decrease over time (ie. when not written to, the system should just mark those pages as free).
Yep, it is the island of Santorini.
What is not generally known, is that somewhere in the nineties an ancient mosaic was found wich showed the harbour city of Atlantis in the old volcano before it exploded.
I haven't read Plato, but I surmise that his description of Atlantis correlates with this picture.
If anybody has read 'Dancer From Atlantis' from Poul Anderson, he uses the same description.
And all this leads to the conclusion that Atlantis was the mainstay of the fleet of Crete, around 1600 BC.
Install Debian Woody
/etc/apt/sources.list -> testing
apt-get update
apt-get upgrade
It is not because the official installation of Debian is four years behind that you have to run older software.
Production setting : keep a local mirror of sarge and update it every three months. Test the updated software, upgrade production servers from the local mirror.
:-) I run Debian over Red Hat on my flashy new server, because Red Hat is able to control my Promise SX6000 controller, but Debian isn't.
Finding documentation and training about it is rather hard, but you can create forms, queries and reports in OOo.
For the reports you need to have Java 1.4 installed.
Yep.
I bought for my father the cheapest PC I could find, I installed Libranet 2.7.
You know what ? Everything worked fine from step 1 : sound, USB, video. Afterwards I had to install his printer : a HP 710C colour printer, which is a Winprinter. Works flawless.
SAR!
Along similar usability lines, I am currently running Xandros as a pilot to see if it is "mom" (or grandmom) ready as it is advertised as "very user friendly". In the process of this pilot, I've done detailed documentation of every step I've done to get my Xandros fully working. My hope is that these step-by-step instructions will help my various friends who seem interested, but are a bit scared of Linux.
I did this already with my father. Although a little bit memory hungry, I opted for KDE on his 256Mb, 2.0Ghz white box, installed Xandros, used the Debian CD's to have a larger list of applications and installed StarOffice 7.0 for him.
What can he do ? Everything he needs to, but I need more documentation for him.
I also connected his old PC using Samba for those things that are really not possible on his new PC : he has an unsupported scanner, he has a dia scanner on loan, but the SCSI card is ISA, he has an organiser on his serial port (maybe the software can run under Wine, but I have no time to learn Wine).
Regards,
Jurgen
The microsoft patents aren't about all of the FAT file system. Rather the method of storing short and long file names in the FAT file system.
IIRC, then something like this was already done in OS/2 3.0, of which I bought a copy in 1994. OS/2 could work on HPFS or FAT, and on the FAT system you had the same possibilities (albeit slower) like on the HPFS system.
I also seem to remember that the files which store the extended attributes had the same names on OS/2, as on Win95.
I think that women's lib in the past 30 year have put that idea in everybodies mind.
Saturday 27 March 2004 I created such a setup for my father. Most of the things that he wants to do know, he can do on Linux (I use Libranet + Debian updates), but he has a scanner which is not supported under Linux, a simple organizer which must exchange data through a Windows program, and a slide scanner, which uses an old Adaptec SCSI card in an ISA slot.
I set up his Windows computer headless, Win98, and you must use TweakUI for the system to automatically logon on the network. You must also disable the Stop on No Keyboard in the BIOS.
I installed tightVNC, which automatically starts up as a service. I also setup Samba to export the home directory, so that everything that is being saved on the Windows computer, must always be done on the mapped drive.
Jurgen
Yes, COBOL is verbose, but...
It seems that ADA maybe also has that last possibility, but I do not know if their fixed-point data types are based upon a float library or a real integer library. In COBOL these are absolutely based upon integer computations.
All other Unix/C derived languages do not have the possibility to make financial computations like COBOL does.
That is because most managers have never made an effort to compute what it would cost them if they should lose their data, and also do not have a clue what a downtime costs them.
I have worked with someone who had done these figures, and he was never happy with downtimes, to say the least.
In general, though, the more 'realistic' it is, the less spectacular it's going to be. *shrug*
Probably, because movie- and seriesmakers will more go for effect than realism.
I think that the most realistic space wars have probably been written by Larry Niven (Protector), CJ Cherryh (Downbelow station), Poul Anderson (The Space Fox), Joe Haldeman (The Forever War).
The scale of these things is too large and too slow for a movie, but they provide fine strategic thinking, which is mostly absent in movies or series.
If you take and use Gauss theory on the bell curve and apply it, then after cutting out the dead wood and replacing them, you should get a finer and higher curve.
I think this is typical for people who do not know anything about math or science. The bell curve is something that you derive from measurements, doing the inverse, putting your bell curve upon your measurements does not make sense.
The algorithm to use is to first do your measurements, compute your curve and then discard the measurements which fall outside your range. But normally you would also discard measurements which are too high, which would mean that you should discard also someone who has too high a productivity. As this make no sense, measurements should be redone.
You can use it on people, and if you apply this correctly then you will have a bell curve, and maybe there will be people who fall behind. But if you replace these with others who ranks more near the middle, then your bell curve will contract and shift to the right.
Again, this way of working is purely devised by evil Catbert people, who do not understand anything of mathematics.