I have been watching the browser wars with much interest in the last several months. This is because I manage a project developing Java applet based database clients. I find both browsers very buggy and don't particularly like either. At the moment there just isn't anything better. Unfortunately because of the slow movement at Netscape for the last year, I see Microsoft gaining quickly in performance and features, and will probably blow by Netscape in the next year.
I see IE as the biggest threat to Linux. The reason for that is I see the browser becoming the desktop of the future. If IE is the only real browser left, then Micrsoft will have an even bigger and stronger monopoly on "desktops" than they have now. Microsoft isn't about to make a Linux version of IE. And with out a good browser, Linux will never make the transition to the desktop from the server. (Taking over the server market I see as just a matter of time.) Part of my assumption here is that the next killer App will be built on top of a browser. And if IE is the only serious browser in town, then Microsoft still holds all the cards, (and a couple of spare Ace's).
So from my point of view Mozilla is more important to linux than gnome/kde. Having said all of this, I have a question. A freind of mine and myself have talked about putting together a Mozilla distribution CD that contains "up to the week" source code, and the latest Milestone binaries currently found on mozilla.org. If you could buy one of these CDs, would you buy it? Would you report bugs or help with the Mozilla development? The only problem is that because of the rapid change in code and binaries, glass mastered CD's are out, it takes too long to have that done. And quick turn around for CD-R's is a bit higher per disk ($5-6). If you would buy and use a Mozilla distribution CD , mail me at Noble. Also we need people who can help set up the distribution for others OS's (Windows and Mac). I think we have Linux (and most unix) covered. E-mail me if you have time and knowledge to help with that.
But most importantly, help Mozilla anyway you can.
For the younger generation. Read Clifford Stoll's book "The Cuckoo's Egg". It will shed a bit of light on the abilities (or lack there of) of the US government.
Sun isn't giving away the OS. They are just offering the Source under "yet another license."
For Everyone !
Who cares? On hardware less than $500,000 linux is better than Solaris. It has more device driver support, it has much better virtual memory management, it is easier to install, port to, and use.
To quote a freind upon hearing the news. "Good maybe Solaris will improve."
I don't think Sun did this to compete with MS. I think they are scared of Linux. And they have reason to be. Sun will get pushed to the high end market and be competeing more with Cray than with MS, if Linux continues to get market share in the low and mid-range Unix markets. And we know how hard it is to make money in the super computer market.
I'm more worried about what sun does with Java. But that a whole other thread.
First let me say I went to NC State University 7 years ago. There I met many student who had gone to VA Tech for Computer Science, and left for NCSU because the VA Tech program "sucked."
On the other hand, when I was at NCSU I would say only about 50% of the faculty knew any sort of modern programing language (like one I had heard of). Tenured (SP?) professors can be VERY lazy.
Even if Episode 1 is the worst movie of the year, it will still be the top grossing movie of the decade. My freinds think of me as a StarWars fan, but compared to REAL fans, I'm just a guy that enjoyed the movies. (In high school I once wrote that "it is possible that George Washington existed, the revolutionary war could have really been faught, but Yoda IS real, and Star Wars DID happen.") I really expect Episode 1 to SUCK badly. But I still will see it even if its the only movie I see for the rest of the year. And even if I'm right, and people all go around talking about how bad the movie is, and there is as much "anti-hype" as there has been hype, this movie will still sale more tickets than anything before it. Titanic had an incredible draw. I remember seeing people sitting on the floor when I saw it 2 months after it came out. But Titanic will look small compared to E1. Bank on it, George is...
First I think the placement of all RPMS in/usr/bin is a real problem. I no longer know the properway to partition my harddrives. I think core functions should go in/usr/bin and others somewhere else. Where you ask? Well if not/usr/local/bin or/usr/local/RH/bin, how about a user definable location? Even Microsoft (most of the time) gives you a choice on where to install (most) of the files.
Second problem, is the official lack of useful support. I don't expect alot of support, but going through official channels I wouldn't trust RH to do anything. Through un-official channels, RH employees are good about helping people. They now how support to buisnesses, but how about 1-900 support for home users or people in a hurry.
I think there will be more money to be made in spreadsheets and wordprocessors than in OS's for a long time coming. I have yet to see a "free" (read non-commercial, open-source-software (OSS)) spreadsheet or wordprocessor that was acceptable.
Further more I'm convinced that in the future (say 10 to 20 years) money in software will only be made on HIGH priced, but small market products. Things where the user base isn't big enough to support a good OSS project. Or in very limited time projects (Tax software, with a 3 month window of usefullness).
There are tuning param's in the Kernal for this. I know because at my last job we played with a Linux box with 1 gig of memory. We noticed that it was doing lots of IO when it was only using 400Meg of memory. E-mail to Linus got a resonce with some var to play with in real time (writing to files in/proc) that fixed things up nicely. These tricks probably need to be documented somewhere.
As Linux becomes used for bigger jobs in buisness a high quality Kernal Tuning HOWTO would be good. Even if it were a published book.
If you can put together a reliable Fax on Demand system under linux for less than $3000 up front and less than $1200 a year in support contracts then you have a viable chance to start your own company. And I'm talk about SOFTWARE prices only this excludes the $4000 to $15,000 in hardware costs!!
Reliable is defined as your software crashes less often than Windows-NT.
1. Redhat does charge for their OS. Although you can down load it for free, they do sale somewhere around 250,000 copies a year. (I have inside knowledge of this magic, if arbitrary sounding number.) Even at $10 a copy thats not bad.
2. Microsoft pays many people to develope their OSes. While Redhat does contribute and pay for some Linux development, their cost is small compared to Microsoft's.
In my case I am the MIS department, and I consider myself at least a budding "linux hacker." And I do tend to scroll through news groups (or ask on Slashdot) for answers. Infact I honestly believe support is mostly for those too lazy to read the manual. However my manager (company president) would feel much more confortable if we had a support contract for linux. Or atleast if there were a company that charged $300 an issue to help resolve problems Microsoft does this at $200 an issue, and their support is reasonably good, as long as its not about a real bug in their software. Then they just refund your $200 and say "Next." If there is such a company, let me know.
I too find it worrying. But I take consolation in the fact that Linux was cool before RedHat and will still be cool after RedHat.:-)
Has anyone noticed that Redhat's commercial support prices make Microsoft's support seem down right cheap?! It is almost $1000 for 3 issues that expires in 90 days!
This reaction isn't suprising. A basic instinct is for people to be afraid of what they don't understand. The vast majority of the population doesn't understand computers, much less hackers. So the reaction is total fear.
This doesn't excuse the reaction. I generally feel that what makes humans human is the ability to react AGAINST our basic instincts!
Mellisa was just the "internet worm" for 1999. (I still wonder if I saw the author of the worm at last years Linux Expo. The name on the name tag was right, as was his apparent age.) It wasn't a big deal. But some people are still afraid of the dark.
For more info on the internet worm, read http://www.alw.nih.gov/Securi ty/FIRST/papers/virus/gao.txt
Ok so can somebody list the security holes for default RH5.2.
I administer a Website running linux (RH 5.2) and have been worried about security. I know this is off topic, but it is probably very important for a large number of/. readers.
I'm not going to pass judgment on Mozilla.org, but Linux does need a Good browser. I like Netscape 3.01 better than any browser out there. And I know a few who agree with me.
I hope people do get behind Mozilla or start a new browser project for Linux. Is it legal to take raptor an run with that?
I agree that MS hasn't done much against Oracle. But remember most MS products suck, and that hasn't stopped them before.
I see Oracle on Linux eventually being the defacto small to medium sized database platform. The stability of SQL server and Win-NT make this very likely. We can only hope.
I'm curious as to what happened/happens next? Does your freind continue to press charges? So the detective doesn't believe it was rape. His opinion is realtively unimportant. You can press charges anyway. The judge, jury, Attornies opinions are the ones that are important.
I hope your freind didn't give up just because of one very miss informed detective!
Unfortunately I have to administer Windows NT boxes along side the Linux boxes I maintain. At work the users are all on Win NT. I would love to know the names of some of the Software packages for NT that allow remote install of software packages. The network has 20 to 25 users. So we can't afford to spend a fortune, but it would be worth $2000 or so.
This is my reply to Lewis's post on security issues.
Lewis A. Mettler,
You have once again gained quiet a bit of notoriety on Slashdot. And while the charge that Slashdotters are arrogant, might be true for many, I wish to take you to task for your original post on Open Source and security issues.
In your document (http://freehosting.at.webjump.com/le/lewismettler -software-engineering/OpenSource.html) you talk about the security risk on open source operating systems of someone changing the OS source code and then compiling a new kernel to give them as much access as they want.
However you gloss over a few issues and I will explain how they are BIG details you leave unmentioned.
1) You fail to mention that on these systems you must already have root ("top dog") access to the machine to edit the source code in the default installation configuration. And if you already have root access, there is not much need for you to recompile the kernel to do pretty much anything you want to do. I feel that this already gives Unix (or Unix like) operating systems more security than most PC operating system environments where there is very little any user sitting at the machine can't do.
Now I ask myself is there a way around this issue? Why yes you could build your own kernel on floppy disk (or burn it on a CD) and reboot the machine if you have physical access to it. Then you could effectively install your kernel on the machine. However this is not really any different from reboot the machine and reinstalling ANY operation system. Allow me to repeat a story told to me a few years ago by a friend of mine.
He was at a company party with a friend. During the middle of the party they went into their manager's office with an old MS-DOS floppy disk. They inserted the disk into their manager's PC and rebooted it. It booted off the floppy and into MS-DOS which most computer literate people know has no concept of security. They walked back out of the office with a spreadsheet containing the salaries for their division. What division might that have been, you may ask? My friend worked in the computer system security audit division of one of the US's largest Banks!!
Does having a proprietary closed source operating system give people a false sense of security? Based on the above story I would say that yes it does, even among those that should be very paranoid about security.
2) You fail to mention or perhaps are unaware of the fact that the security issue you describe has been taken very seriously. And that almost 10 years ago solutions to it for large networks of computers in a corporate setting or even worse, a university setting. Now you may say hold on a University is academia, that not the real world. But let me tell you when it comes to computer security academia is worse than the real world. Because in the university setting the administrators can be absolutely sure that some of the students will be VERY creative about trying to break into the computing systems.
Starting probably 10 or more years ago, CMU and MIT were looking into very much the same security issues you raise. Namely how to set up a large functioning network of computers when you can not trust the authenticity of the user or the OS running on machines on the network. The result of that research is AFS and Kerberos.
It has been more than 6 years since I have paid much attention to such security issues, but you and your readers could learn quiet a bit spending an afternoon surfing on these two topics. And they will allow a company to set up a network of computers with a reasonably high degree of security from the type of attacks you describe.
This is a very good point. It is rather ironic, that this is happening.
To me this fight amoung factions of the open source movement is the greatest threat facing open source. Once they had someone "sign off" on it Apple probably thought they had a good License. I beleive this whole thing is a big surise to them.
What Apple does next is going to be the big question. Will they be scared away from Open Source? Jobs is widely blamed for not wanting Mac clones made; so I just can't see him being very warm to OS. Time will tell.
I hope this works. I started reading/. back when I could read ALL the comments on Most of the stories. Now I rarely read all the stories much less the comments. Maybe this will make it worth looking at comments again.:-)
Anyway, the only thing I could add is "Would it work to let everyone vote on comments and rank them automatically by positive and negative votes?" Of course this assumes that they are more "normal" users than flame mongers.
I would not know Linux World exists if it haddn't been posted to/. Further I visit/. 4 times more often than I visit any other site on the web, including my own.
Granted I don't need slashdot, but since it is here, it is almost my exclusive link to the linux world. And I find/. more valuble than Linux world.
The linux community is very different from any other business community in existance. LW is making prohibatively high access fees to this conference. $1600 for 3 IP addresses? How is that even vaguely related to the cost of the service?
Slashdot Mirror
I have been watching the browser wars with much interest in the last several months. This is because I manage a project developing Java applet based database clients. I find both browsers very buggy and don't particularly like either. At the moment there just isn't anything better. Unfortunately because of the slow movement at Netscape for the last year, I see Microsoft gaining quickly in performance and features, and will probably blow by Netscape in the next year.
I see IE as the biggest threat to Linux . The reason for that is I see the browser becoming the desktop of the future. If IE is the only real browser left, then Micrsoft will have an even bigger and stronger monopoly on "desktops" than they have now. Microsoft isn't about to make a Linux version of IE. And with out a good browser, Linux will never make the transition to the desktop from the server. (Taking over the server market I see as just a matter of time.) Part of my assumption here is that the next killer App will be built on top of a browser. And if IE is the only serious browser in town, then Microsoft still holds all the cards, (and a couple of spare Ace's).
So from my point of view Mozilla is more important to linux than gnome/kde. Having said all of this, I have a question. A freind of mine and myself have talked about putting together a Mozilla distribution CD that contains "up to the week" source code, and the latest Milestone binaries currently found on mozilla.org. If you could buy one of these CDs, would you buy it? Would you report bugs or help with the Mozilla development? The only problem is that because of the rapid change in code and binaries, glass mastered CD's are out, it takes too long to have that done. And quick turn around for CD-R's is a bit higher per disk ($5-6). If you would buy and use a Mozilla distribution CD , mail me at Noble. Also we need people who can help set up the distribution for others OS's (Windows and Mac). I think we have Linux (and most unix) covered. E-mail me if you have time and knowledge to help with that.
But most importantly, help Mozilla anyway you can.
For the younger generation. Read Clifford Stoll's book "The Cuckoo's Egg". It will shed a bit of light on the abilities (or lack there of) of the US government.
Sun isn't giving away the OS. They are just offering the Source under "yet another license."
For Everyone !
Who cares? On hardware less than $500,000 linux is better than Solaris. It has more device driver support, it has much better virtual memory management, it is easier to install, port to, and use.
To quote a freind upon hearing the news. "Good maybe Solaris will improve."
I don't think Sun did this to compete with MS. I think they are scared of Linux. And they have reason to be. Sun will get pushed to the high end market and be competeing more with Cray than with MS, if Linux continues to get market share in the low and mid-range Unix markets. And we know how hard it is to make money in the super computer market.
I'm more worried about what sun does with Java. But that a whole other thread.
There are many more professional level engineering packages for Unix than for NT.
First let me say I went to NC State University 7 years ago. There I met many student who had gone to VA Tech for Computer Science, and left for NCSU because the VA Tech program "sucked."
On the other hand, when I was at NCSU I would say only about 50% of the faculty knew any sort of modern programing language (like one I had heard of). Tenured (SP?) professors can be VERY lazy.
It was pretty easy.
They have interfaces in c and perl ammount many other langs.
Even if Episode 1 is the worst movie of the year, it will still be the top grossing movie of the decade. My freinds think of me as a StarWars fan, but compared to REAL fans, I'm just a guy that enjoyed the movies. (In high school I once wrote that "it is possible that George Washington existed, the revolutionary war could have really been faught, but Yoda IS real, and Star Wars DID happen.")
I really expect Episode 1 to SUCK badly. But I still will see it even if its the only movie I see for the rest of the year. And even if I'm right, and people all go around talking about how bad the movie is, and there is as much "anti-hype" as there has been hype, this movie will still sale more tickets than anything before it. Titanic had an incredible draw. I remember seeing people sitting on the floor when I saw it 2 months after it came out. But Titanic will look small compared to E1. Bank on it, George is...
First I think the placement of all RPMS in /usr/bin is a real problem. I no longer know the properway to partition my harddrives. I think core functions should go in /usr/bin and others somewhere else. /usr/local/bin or /usr/local/RH/bin, how about a user definable location? Even Microsoft (most of the time) gives you a choice on where to install (most) of the files.
Where you ask? Well if not
Second problem, is the official lack of useful support. I don't expect alot of support, but going through official channels I wouldn't trust RH to do anything. Through un-official channels, RH employees are good about helping people. They now how support to buisnesses, but how about 1-900 support for home users or people in a hurry.
I think there will be more money to be made in spreadsheets and wordprocessors than in OS's for a long time coming. I have yet to see a "free" (read non-commercial, open-source-software (OSS)) spreadsheet or wordprocessor that was acceptable.
Further more I'm convinced that in the future (say 10 to 20 years) money in software will only be made on HIGH priced, but small market products. Things where the user base isn't big enough to support a good OSS project. Or in very limited time projects (Tax software, with a 3 month window of usefullness).
These tricks probably need to be documented somewhere.
As Linux becomes used for bigger jobs in buisness a high quality Kernal Tuning HOWTO would be good. Even if it were a published book.
If you can put together a reliable Fax on Demand system under linux for less than $3000 up front and less than $1200 a year in support contracts then you have a viable chance to start your own company. And I'm talk about SOFTWARE prices only this excludes the $4000 to $15,000 in hardware costs!!
Reliable is defined as your software crashes less often than Windows-NT.
Did you catch the note about Burlington Coat Factory ordering 1,250 Linux desktop machines?
9 9/9904/6a.htm
To me thats the big news, not Dell's investment.
Read it here http://www.dell.com/corporate/media/newsreleases/
Dell press release.
1. Redhat does charge for their OS. Although you can down load it for free, they do sale somewhere around 250,000 copies a year. (I have inside knowledge of this magic, if arbitrary sounding number.) Even at $10 a copy thats not bad.
2. Microsoft pays many people to develope their OSes. While Redhat does contribute and pay for some Linux development, their cost is small compared to Microsoft's.
In my case I am the MIS department, and I consider myself at least a budding "linux hacker." And I do tend to scroll through news groups (or ask on Slashdot) for answers. Infact I honestly believe support is mostly for those too lazy to read the manual. However my manager (company president) would feel much more confortable if we had a support contract for linux. Or atleast if there were a company that charged $300 an issue to help resolve problems Microsoft does this at $200 an issue, and their support is reasonably good, as long as its not about a real bug in their software. Then they just refund your $200 and say "Next." If there is such a company, let me know.
A local company I interviewed with implied they helped set up the CNN site. Have you heard of Persimon IT?
I too find it worrying. But I take consolation in the fact that Linux was cool before RedHat and will still be cool after RedHat. :-)
Has anyone noticed that Redhat's commercial support prices make Microsoft's support seem down right cheap?! It is almost $1000 for 3 issues that expires in 90 days!
This doesn't excuse the reaction. I generally feel that what makes humans human is the ability to react AGAINST our basic instincts!
Mellisa was just the "internet worm" for 1999. (I still wonder if I saw the author of the worm at last years Linux Expo. The name on the name tag was right, as was his apparent age.) It wasn't a big deal. But some people are still afraid of the dark.
For more info on the internet worm, read
http://www.alw.nih.gov/Securi ty/FIRST/papers/virus/gao.txt
I administer a Website running linux (RH 5.2) and have been worried about security. I know this is off topic, but it is probably very important for a large number of /. readers.
I'm not going to pass judgment on Mozilla.org, but Linux does need a Good browser. I like Netscape 3.01 better than any browser out there. And I know a few who agree with me.
I hope people do get behind Mozilla or start a new browser project for Linux. Is it legal to take raptor an run with that?
I agree that MS hasn't done much against Oracle. But remember most MS products suck, and that hasn't stopped them before.
I see Oracle on Linux eventually being the defacto small to medium sized database platform. The stability of SQL server and Win-NT make this very likely. We can only hope.
I'm curious as to what happened/happens next? Does your freind continue to press charges? So the detective doesn't believe it was rape. His opinion is realtively unimportant. You can press charges anyway. The judge, jury, Attornies opinions are the ones that are important.
I hope your freind didn't give up just because of one very miss informed detective!
Unfortunately I have to administer Windows NT boxes along side the Linux boxes I maintain. At work the users are all on Win NT. I would love to know the names of some of the Software packages for NT that allow remote install of software packages. The network has 20 to 25 users. So we can't afford to spend a fortune, but it would be worth $2000 or so.
This is my reply to Lewis's post on security issues.
r -software-engineering/OpenSource.html) you talk about the security risk on open source operating systems of someone changing the OS source code and then compiling a new kernel to give them as much access as they want.
Lewis A. Mettler,
You have once again gained quiet a bit of notoriety on Slashdot. And while the charge that Slashdotters are arrogant, might be true for many, I wish to take you to task for your original post on Open Source and security issues.
In your document (http://freehosting.at.webjump.com/le/lewismettle
However you gloss over a few issues and I will explain how they are BIG details you leave unmentioned.
1) You fail to mention that on these systems you must already have root ("top dog") access to the machine to edit the source code in the default installation configuration. And if you already have root access, there is not much need for you to recompile the kernel to do pretty much anything you want to do. I feel that this already gives Unix (or Unix like) operating systems more security than most PC operating system environments where there is very little any user sitting at the machine can't do.
Now I ask myself is there a way around this issue? Why yes you could build your own kernel on floppy disk (or burn it on a CD) and reboot the machine if you have physical access to it. Then you could effectively install your kernel on the machine. However this is not really any different from reboot the machine and reinstalling ANY operation system. Allow me to repeat a story told to me a few years ago by a friend of mine.
He was at a company party with a friend. During the middle of the party they went into their manager's office with an old MS-DOS floppy disk. They inserted the disk into their manager's PC and rebooted it. It booted off the floppy and into MS-DOS which most computer literate people know has no concept of security. They walked back out of the office with a spreadsheet containing the salaries for their division. What division might that have been, you may ask? My friend worked in the computer system security audit division of one of the US's largest Banks!!
Does having a proprietary closed source operating system give people a false sense of security? Based on the above story I would say that yes it does, even among those that should be very paranoid about security.
2) You fail to mention or perhaps are unaware of the fact that the security issue you describe has been taken very seriously. And that almost 10 years ago solutions to it for large networks of computers in a corporate setting or even worse, a university setting. Now you may say hold on a University is academia, that not the real world. But let me tell you when it comes to computer security academia is worse than the real world. Because in the university setting the administrators can be absolutely sure that some of the students will be VERY creative about trying to break into the computing systems.
Starting probably 10 or more years ago, CMU and MIT were looking into very much the same security issues you raise. Namely how to set up a large functioning network of computers when you can not trust the authenticity of the user or the OS running on machines on the network. The result of that research is AFS and Kerberos.
It has been more than 6 years since I have paid much attention to such security issues, but you and your readers could learn quiet a bit spending an afternoon surfing on these two topics. And they will allow a company to set up a network of computers with a reasonably high degree of security from the type of attacks you describe.
This is a very good point. It is rather ironic, that this is happening.
To me this fight amoung factions of the open source movement is the greatest threat facing open source. Once they had someone "sign off" on it Apple probably thought they had a good License. I beleive this whole thing is a big surise to them.
What Apple does next is going to be the big question. Will they be scared away from Open Source? Jobs is widely blamed for not wanting Mac clones made; so I just can't see him being very warm to OS. Time will tell.
I hope this works. I started reading /. back when I could read ALL the comments on Most of the stories. Now I rarely read all the stories much less the comments. Maybe this will make it worth looking at comments again. :-)
Anyway, the only thing I could add is "Would it work to let everyone vote on comments and rank them automatically by positive and negative votes?" Of course this assumes that they are more "normal" users than flame mongers.
Good Luck Rob.
I would not know Linux World exists if it haddn't been posted to /. Further I visit /. 4 times more often than I visit any other site on the web, including my own.
/. more valuble than Linux world.
Granted I don't need slashdot, but since it is here, it is almost my exclusive link to the linux world. And I find
The linux community is very different from any other business community in existance. LW is making prohibatively high access fees to this conference. $1600 for 3 IP addresses? How is that even vaguely related to the cost of the service?