Sorry for the incomplete nature of this reply. IIRC that a year or so ago there was some discussion about the IDE driver in Linux. At one point it was a kludge. Then someone redid it as a kind of SCSI interface that really cleaned it up and made it much more maintainable. I think the reason the original one was inadequate had to do with trying to maintain compatibility with just the kind of Windows 'exceptions' you are referring to. You might want to hunt up some of those old threads on the kernel mailing lists and see if they give you an insight into any of the issues. Sorry I can't tell you more, this is not really my area. S
Since patent infringement and defenses against suits related to it, seem to be the business model of the future for software firms, how about a Law degree?
I am not at ALL sure you can 'get the facts anywhere'. I certainly can't. And even the 'public record' can be censored after the fact depending on the settlement of the case. If you could get the facts anywhere, there would be no such thing as FUD. I agree with another poster. Objective, and Balanced, and Neutral are not always the same thing. When it comes to facts, I think Groklaw is objective. They present what is there. This is NOT true of most media, especially mainstream media. When it comes to opinion, they are honest; they disclose their biases up front. This is also not true of most media, especially mainstream media.
From personal experience a P-III with 256MB of RAM should have no trouble running almost any Distro. (MDK 10.1, FC3, Debian, Gentoo....) I would take a look at what sort of apps you would need, what sort of support you would need (Mandrake have just merged with Connectiva which I understand are well established in S.A.) and if language support (spanish?) is a big issue for you.
I sysadmin Solaris boxes at work, so I am not totally clueless when it comes to Linux. But Linux installs are definitely not easy. And installing apps afterwards is even worse. Here is what Fedora Core 3 did NOT do when I tried to install it last week. 1) Boot from the CD. (RH 6.0 CD, & Knoppix CD boot fine) (No error message from FC3 either.) 2) Detect the LAN card. (Had to modprobe the correct module) 3) There was a documented and fixed bug in the X.org server that crashed the install till I changed the graphics card. FC3 did not include the fixed version of X.org for some reason. Again Knoppix and RH 6 had no trouble with the card. (Using an older version of XFree) 4) Detect the sound card. Had to modprobe etc. 5) Play a CD, or give a hint. (Had to open the CPU and connect up the Analog audio cable). 6) Preserve my previous partitions. MDK tells me that it needs to reformat swap and / but it will let me keep/home. FC3 blows everything away. Lest you think FC3 is unique in this regard it isn't. MDK 10.1 was a lilttle better (it detected the LAN card) but otherwise experienced exactly the same problems. I haven't even tried digital cameras, network printing, Samba access to my Windows machines, or the scanner yet. Any bets on how smoothly THAT will go?
At the 2004 OsCON I heard about a Portland Ore. group called 'Free Geek'. From what I understand they take old computers, and reconfigure them for novices then give them away. They spend up to 12 hours helping the novice plug in cards, and IDE cables, load Linux etc. When they are finished the novice is at lest minimally competent. In exchange the novice volunteers some of his time to help the next group and so on. They pay for their rent, and the one full time 'employee' with the money they get from the gold and copper in the motherboards and parts that they cannot use, and must send to the scrap heap. Seems like a much better approach than another layer of bureaucracy.
Perhaps you should ask the project lead for Mozilla, who is I believe, female. At the 2004 OsCON they had a panel about FOSS volunteers. The Apache lead mentioned that there were two groups of people who tended to drop out early: women, and Japanese. In his opinion, they were driven away by the flame wars on the mailing lists. Women, and Japanese, were just not comfortable with the level of 'discussion'. The Apache head was not happy about this, but there doesn't seem much he can do.
St. Thomas Aquinas, considered one of the greatest minds of the last 1000 years, (I forget the exact poll) had similar postulates. He needed as a starting point that you grant that : "There is an IS". And of course DesCartes famous dictum: "I think therefore I am" is somewhat in the same vein. So being an atheist (or not) doesn't have much to do with it.
Re:The GPL should be a little friendler.
on
Revising the GPL
·
· Score: 1
The previous replies are too hard on the poster. My reading of his post is that he does not want others to derive profit from his work, unless he is going to get a share. So long as everyone keeps the derived products open, I think he is happy to release his patented material for free. I think this is exactly what the GPL is trying to accomplish. I think the poster is worried about what happens to his code/ideas if some later party circumvents the GPL. As to patenting math, one good reason for doing so is so that no one else can patent it. That gives the patent holder the right to make sure that the idea stays free. A large firm like IBM might adopt this strategy for example.
"Begin Sarcasm Type=Extreme " AHA!! I note that even an Open Source Company ADMITS that it takes 3 times as much money for consultants to support Linux as Windows. This "proves" that Linux is trouble prone! "End Sarcasm " I wonder how long it will be before the above version of the study appears on M$ 'Get the facts' page. After all if they can pervert PJ's (now terminated) involvement in OSRM, they can put spin on anything.
A good summary of how to do this can be found in ESR's article "The Magic Cauldron". In it he explains the difference between 'sale value' which is usually realized by selling proprietary bits, and 'use value'. Also why even though it is theoretically possible to derive 'sale value' from OSS, it is practically never done. The fact is that the real value in S/W (not just OSS) is in the 'use value', and that value is mostly derived from support, upgrades, training etc. etc. The software business is really a service business, not a manufacturing business, and OSS forces s/w mfg's to confront this fact.
Thanks to previous posters we now know the Beowulf flavor of clustering (supercomputing) and the load balancing flavor of clustering (mostly web servers) and also the High Availability flavor of clustering (duplicated servers with or without load balancing.) There is one area I have not seen addressed and that is disk clustering where the disk farm (usually a RAID array) has (hardware) access by two or more machines. Microsoft call their version MSCS. It is useful for database servers because you do not need two copies of the database. One DB copy on disk can be read/written by more than one machine directly over a scsi ( or fiber) channel. (Disk redundancy is provided by the RAID-5).
Is there such a thing for Linux? (Not RAID-5. It works fine right out of the box. It's the dual porting software I have not seen).
Slashdot Mirror
Sorry for the incomplete nature of this reply.
IIRC that a year or so ago there was some discussion about the IDE driver in Linux. At one point it was a kludge. Then someone redid it as a kind of SCSI interface that really cleaned it up and made it much more maintainable.
I think the reason the original one was inadequate had to do with trying to maintain compatibility with just the kind of Windows 'exceptions' you are referring to.
You might want to hunt up some of those old threads on the kernel mailing lists and see if they give you an insight into any of the issues.
Sorry I can't tell you more, this is not really my area.
S
Since patent infringement and defenses against suits related to it, seem to be the business model of the future for software firms, how about a Law degree?
I am not at ALL sure you can 'get the facts anywhere'. I certainly can't. And even the 'public record' can be censored after the fact depending on the settlement of the case. If you could get the facts anywhere, there would be no such thing as FUD.
I agree with another poster. Objective, and Balanced, and Neutral are not always the same thing. When it comes to facts, I think Groklaw is objective. They present what is there. This is NOT true of most media, especially mainstream media.
When it comes to opinion, they are honest; they disclose their biases up front. This is also not true of most media, especially mainstream media.
From personal experience a P-III with 256MB of RAM should have no trouble running almost any Distro. (MDK 10.1, FC3, Debian, Gentoo ....)
I would take a look at what sort of apps you would need, what sort of support you would need (Mandrake have just merged with Connectiva which I understand are well established in S.A.) and if language support (spanish?) is a big issue for you.
I sysadmin Solaris boxes at work, so I am not totally clueless when it comes to Linux. But Linux installs are definitely not easy. And installing apps afterwards is even worse. /home. FC3 blows everything away.
Here is what Fedora Core 3 did NOT do when I tried to install it last week.
1) Boot from the CD. (RH 6.0 CD, & Knoppix CD boot fine) (No error message from FC3 either.)
2) Detect the LAN card. (Had to modprobe the correct module)
3) There was a documented and fixed bug in the X.org server that crashed the install till I changed the graphics card. FC3 did not include the fixed version of X.org for some reason. Again Knoppix and RH 6 had no trouble with the card. (Using an older version of XFree)
4) Detect the sound card. Had to modprobe etc.
5) Play a CD, or give a hint. (Had to open the CPU and connect up the Analog audio cable).
6) Preserve my previous partitions. MDK tells me that it needs to reformat swap and / but it will let me keep
Lest you think FC3 is unique in this regard it isn't. MDK 10.1 was a lilttle better (it detected the LAN card) but otherwise experienced exactly the same problems.
I haven't even tried digital cameras, network printing, Samba access to my Windows machines, or the scanner yet. Any bets on how smoothly THAT will go?
At the 2004 OsCON I heard about a Portland Ore. group called 'Free Geek'. From what I understand they take old computers, and reconfigure them for novices then give them away. They spend up to 12 hours helping the novice plug in cards, and IDE cables, load Linux etc. When they are finished the novice is at lest minimally competent. In exchange the novice volunteers some of his time to help the next group and so on.
They pay for their rent, and the one full time 'employee' with the money they get from the gold and copper in the motherboards and parts that they cannot use, and must send to the scrap heap.
Seems like a much better approach than another layer of bureaucracy.
Perhaps you should ask the project lead for Mozilla, who is I believe, female. At the 2004 OsCON they had a panel about FOSS volunteers. The Apache lead mentioned that there were two groups of people who tended to drop out early: women, and Japanese. In his opinion, they were driven away by the flame wars on the mailing lists. Women, and Japanese, were just not comfortable with the level of 'discussion'. The Apache head was not happy about this, but there doesn't seem much he can do.
St. Thomas Aquinas, considered one of the greatest minds of the last 1000 years, (I forget the exact poll) had similar postulates. He needed as a starting point that you grant that : "There is an IS".
And of course DesCartes famous dictum: "I think therefore I am" is somewhat in the same vein.
So being an atheist (or not) doesn't have much to do with it.
The previous replies are too hard on the poster. My reading of his post is that he does not want others to derive profit from his work, unless he is going to get a share. So long as everyone keeps the derived products open, I think he is happy to release his patented material for free.
I think this is exactly what the GPL is trying to accomplish. I think the poster is worried about what happens to his code/ideas if some later party circumvents the GPL.
As to patenting math, one good reason for doing so is so that no one else can patent it. That gives the patent holder the right to make sure that the idea stays free. A large firm like IBM might adopt this strategy for example.
"Begin Sarcasm Type=Extreme "
AHA!! I note that even an Open Source Company ADMITS that it takes 3 times as much money for consultants to support Linux as Windows. This "proves" that Linux is trouble prone!
"End Sarcasm "
I wonder how long it will be before the above version of the study appears on M$ 'Get the facts' page. After all if they can pervert PJ's (now terminated) involvement in OSRM, they can put spin on anything.
A good summary of how to do this can be found in ESR's article "The Magic Cauldron".
In it he explains the difference between 'sale value' which is usually realized by selling proprietary bits, and 'use value'. Also why even though it is theoretically possible to derive 'sale value' from OSS, it is practically never done. The fact is that the real value in S/W (not just OSS) is in the 'use value', and that value is mostly derived from support, upgrades, training etc. etc. The software business is really a service business, not a manufacturing business, and OSS forces s/w mfg's to confront this fact.
As Director of marketing for Linux, what is your estimate of the number of people who would be prepared to pay for applications written for Linux?
Thanks to previous posters we now know the Beowulf flavor of clustering (supercomputing) and the load balancing flavor of clustering (mostly web servers) and also the High Availability flavor of clustering (duplicated servers with or without load balancing.) There is one area I have not seen addressed and that is disk clustering where the disk farm (usually a RAID array) has (hardware) access by two or more machines. Microsoft call their version MSCS. It is useful for database servers because you do not need two copies of the database. One DB copy on disk can be read/written by more than one machine directly over a scsi ( or fiber) channel. (Disk redundancy is provided by the RAID-5). Is there such a thing for Linux? (Not RAID-5. It works fine right out of the box. It's the dual porting software I have not seen).