"It is by caffeine alone I set my mind in motion, it is by the beans of java the thoughts acquire speed, the hands acquire shaking, the shaking becomes a warning, it is by caffeine alone I set my mind in motion."
Tim makes a great point and I'm glad he said this. Windows will not be talked to death, but more importantly, it will not be *imitated* to death. My fellow former OS/2 users can cringe with me as I recite the mantra, "a better DOS than DOS, a better Windows than Windows!"
Whether we're talking about the OS or the applications, the traditional Microsoft architecture and the assumptions built into them are wholly and irrevocably incompatible with a *truly* networked system typified by Unix or the Internet. As annoyed as I might be with Microsoft's quality or ethics, it is ultimately their archaic single-user, local hard drive mentality that I reject. It served the purpose of bringing computer technology into the mainstream of society and providing an economy of scale, but its day is gone, as evidenced by the hilarious and/or pathetic attempts to integrate networking and multi-user capabilities into the Microsoft legacy.
Although nominally a critique of CatB, the author quickly strays into much speculative and anecdotal critisism of Open Source itself, with Linux as a favorite subject. Strange that the article managed to quote just about everybody but Joe down at the pub, while not quoting CatB at all.
"Um, someone I know who wants to remain anonymous says Linux is insecure and unstable." Excuse me, but what are Jesse Berst's and Fred Moody's qualifications to address software development models and economics?
I had heard that First Monday's articles were 'peer reviewed'. In its published state, this would not have survived any peer review I'm familiar with.
> I could open up 4 xterms and display more informaiton that you can with just a dumb text terminal.
At a cost of maybe 30x the RAM:-). Seriously, GNU screen or virtual consoles would be better in this example, let X Window be X Window and serve apps where graphics are a fundamental and essential component.
> FLT. Flashing Led Terminal.
Heh. Where can I get one? Mmmmmm...das blinkenlights...
>...aren't most LAN's a combination of both stand-alone computers and main servers?
A LAN's *hardware environment* is a combination of stand-alone computers and main servers. The software environment of traditional PC's (DOS, Windows, OS/2 and their applications) is very much dedicated to the 'single-user, local configuration in real time' architecture, making any attempt to apply effective networking strategies to them a failure from the start.
> Most have a main file server where everyone's homedir is, but everyone's desktop machine has the ability to live apart from the network.
And do what? Access their files, access shared templates, read e-mail, print to a network printer? Server downtime is a catastrophe in any environment, but it is much easier to implement server redundancy than to maintain reduncancy throughout the desktops of an enterprise. ---------------------
You're correct that any average Unix/Linux system is well equipped to offer the major benefits of Network Computing. I see NC's as a way to overcome the maintenance and logistical mess that networking traditional DOS/Windows/OS2 involves, primarily from their single-user software architecture. In my wired home, I practice the NC-esque "Monster-in-the-Closet" approach, where nearly all computer resources (and thus expenses) are concentrated in one master server, while the rest of the network are little more that SSH and/or X clients. Even an old and very feeble notebook of mine runs DOS and Kermit's Telnet client. In such an environment, NFS is unnecessary and in fact becomes a maintenance and security burden. My example lab is too small to gain much from NIS, but a larger network would implement either NIS or some means of sharing relevant configuration data. Kerberos, done right (dedicated server, etc), can be useful for environments large enough to justify its requirements, but keep in mind that it requires client side support, which means custom versions of commercial software or custom compiles of Open Source packages. Much of Kerberos' functionality is available through SSH, anyway. In a larger environment, I would implement multiple server-class systems for performance reasons, using either failover or individual servers dedicated to certain functions, with mirroring between them. In short, the main benefits of NC architecture are available to the standard Unix and Linux system. NC's have been deprived of success in the market largely for 2 reasons: first, their market (DOS/Windows/OS2) can't grasp how a centralized network is *more* reliable that their current situation, second was the way NC's were prematurely associated with Java and Web browsers, neither of which were ready to serve up the goods at the time that the NC idea was first being promoted. ---------------------
I've been carrying Victorinox Swiss Army knives for longer that I've been working on PC's, and they are a godsend. Lately, I prefer the SwissChamp, although it's large and heavy enough to really require a belt holster instead of a pocket.
The Cybertool looks interesting if only for the Torx drivers, my SwissChamp has everything else and more. ---------------------
This decision, and the corresponding drop in SDRAM prices, was made inevitable by my purchase of a 128 meg upgrade for my server *just yesterday*. ---------------------
This is an interesting point, and perhaps it was a motivating factor behind Red Hat's pre-IPO offers to Linux developers. The more 'owners' they could gather who were knowledgable and sympathetic to Free/Open software, the less likely they could be subject to 'owners' who would demand a maximum-growth-and-profit-or-we'll-sue-you-to-deat h approach.
I'm more optimistic about SuSE, though, I think they'll survive this little "invasion" just fine...:-) ---------------------
Turbo Linux started as a fork of the Red Hat distribution, much as Mandrake and so many others have done since. In other words, Red Hat do not 'assimilate' like the Borg, they *disseminate* to any and all who want to make use of their work.
Those who would promote their preferred distribution by ineptly riduculing another do a disservice to both. ---------------------
Well, releasing 'pre' versions is a habit they picked up from Red Hat:). As long as they follow Red Hat's practice of announcing the 'final' release as a free download, it shouldn't present a problem.
I'm just glad they didn't jump on glibc 2.1.2 right away. --------------------- Most Annoying/. Comment Ever: "If I use Red$at Gnulix to make a Beowulf,
Not to be pedantic, Aaron, but the default RPM build target for Red Hat is the 486. The 'i386' that appears in the names of the packages refers to the architecture, which is a different parameter in RPM and doesn't necessarily reflect the actual compiler flags being used for the binary.
I just wanted to clarify this, since many people see the i386 in Red Hat's RPMs and think RH binaries are the opposite of optimized (pessimized?:). --------------------- Most Annoying/. Comment Ever: "If I use Red$at Gnulix to make a Beowulf,
Garfinkel phrased his statement in such a way as to make it difficult to tell if he meant that Linux was a favorite target of crackers or a favorite *tool* of crackers. What part of that doesn't qualify as Fear, Uncertainty and Doubt? If this had been written by one of our beloved Ziff-Davis comlumnists, the answer would be assumed obvious.
BTW, you've just pushed another of my buttons, namely the whole 'opinion/critisism' defense. There is a common misunderstanding in forums such as this that 'everyone has a right to an opinion' and that 'any critisism is valid as long as it's protected by the 1st Amendment'. Balls to that! People only have a right to an INFORMED opinion, and critisism is valid ONLY to the extent that it can survive countering arguments. Neither Garfinkel nor you (nor I) can assume that we can express any old buggered-up idea and have it stand unchallenged. --------------------- the SlashDot spellchecker:
Garfinkel took a few FUDdy cheap shots (Linux as 'a cracker's OS', too commercial, etc.) but the article was pretty good when he stuck to discussing the merits of the BSD's. In a way, he does everyone a favor by pointing out in a mainstream publication that Linux is a kernel, and that the components of a Linux distribution also come from GNU, BSD, et al. To listen to other articles in mundane media, one might think that Linus wrote the entire OS himself. --------------------- the SlashDot spellchecker:
Only if all the kernel authors would get together and re-release Linux under the GPL (Generic Planet License). Maybe if the planet has a few moons, it would make a neat Beowulf...
>> Frankly, I think Red Hat would do consumers a favour, by unbundling the support service from the CD by itself....
Actually, Red Hat does this. From their web site, you can purchase the Red Hat Core package for $40US, which are the binary and source CD's but without the Applications Demo disk, printed manuals, support, etc. They also offer various separate support programs ranging from a few hundred dollars for per-incident to well into five figures for 24/7 unlimited support.
>> Are they waiting for enough market share and volume to simply fork from Linux, and produce some kind of semi-proprietary version, in which free and non-free software are inextricable mixed, making the free content unfree? After which they can capitalize on the market share Red Hat name, built using free resources, to disallow the freely copying of the CD, and in effect insisting on a license fee per copy?...
Red Hat once mixed proprietary code with Free Software, but have been moving away from this since 5.0. In 6.0, the only proprietary package left is Netscape. What you're suggesting of Red Hat more appropriately describes the practices of other commercial Linux distributors.
>> Therefore, I think the GPL should be adapted quickly, to prohibit that practice of putting GPLed software on a single medium together with proprietary software, in effect removing from the buyer the freedom to copy the free content, because he is not allowed to copy the proprietary content....
That sounds extraordinarily bad, and similar to the problem with version 1 of the GPL which version 2 and the LGPL were intended to correct.
Older notebooks make excellent terminals for a home LAN. I use two 486 notebooks, one running a lightweight install of Red Hat 6.0 w/ SSH and an X server. The other one has only a 486SLC (no math coprocessor) and an 80 meg HD and won't run Linux comfortably, so I use DOS-based Kermit and a TCP/IP packet driver for a Telnet session. Prior to Kermit, I used a serial link w/ Procomm Plus. This is what I love about Linux, and Unix in general, you don't need to maintain and synchronize 5+ computers, just have access to one computer from everywhere. --------------------- the SlashDot spellchecker:
The idea that anyone would steal a computer and then operate said computer on a public network without reformatting the drive or otherwise replacing all identifying parameters... well, I don't know whether to laugh or cry.
Serves them right, of course. It just goes to show that good wins because evil is generally stupid. --------------------- the SlashDot spellchecker:
Before we begin tearing Red Hat apart for all the real and imaginary wickedness they've perpetrated against Free Software, quality and choice, and reading through all the redundant posts that defend them, I propose the following solution that should at least make this latest flamefest worthwhile.
A voluntary Red Hat Bashing Tax, payable, for want of a better metric, in lines of code contributed to Free Software and/or Open Source projects. Say for instance that someone feels compelled to toss around unfunny parodies of Red Hat's name, like "FUD Hat", "Red$at" or "RootHat". Each occurance would oblige the posting individual to submit 10 lines of acceptable code to the project of their choice. More extreme complaints, such as Red Hat "forcing" people to use buggy software or not allowing the posting individual to configure their system without a GUI, would require 100+ lines of code submitted. Complaints about RPM could be a special case, in which the complaining party could submit code that corrects the alleged deficiency in RPM, regardless of line count.
I believe such a policy would greatly benefit the Free Software community by generating a virtually infinite supply of new code while letting people vent all their frustrations and fears against a favorite target. ---------------------
...the last company to abandon disk-based copy protection, which did virtually nothing to hinder illegal copying but routinely punished paying customers, ...asinine 'look-and-feel' lawsuits against VIP, Paperback and ultimately Borland. The first two were driven from the market, QuattroPro operated under a legal cloud for years while Windows and Excel took the market. Lotus had the gall to complement Microsoft during the trial, for using macro translators instead of direct menu and macro compatibility, ...OS/2 ports that were so horrid and dysfunctional as to beggar the imagination, ...IBM's OS/2 support drying up and disappearing almost to the day that they acquired Lotus. No more Visualizer clients for OS/2 DB/2, the new DB/2 client became Approach...for Windows95. WTF?
Lotus has an extensive history for being far more trouble than they're worth. Leave them to be the bottom feeders in the Microsoft owned market they vigorously helped create. ---------------------
Slashdot Mirror
With apologies to Frank Herbert...
"It is by caffeine alone I set my mind in motion,
it is by the beans of java the thoughts acquire speed,
the hands acquire shaking, the shaking becomes a warning,
it is by caffeine alone I set my mind in motion."
---------------------
Tim makes a great point and I'm glad he said this. Windows will not be talked to death, but more importantly, it will not be *imitated* to death. My fellow former OS/2 users can cringe with me as I recite the mantra, "a better DOS than DOS, a better Windows than Windows!"
Whether we're talking about the OS or the applications, the traditional Microsoft architecture and the assumptions built into them are wholly and irrevocably incompatible with a *truly* networked system typified by Unix or the Internet. As annoyed as I might be with Microsoft's quality or ethics, it is ultimately their archaic single-user, local hard drive mentality that I reject. It served the purpose of bringing computer technology into the mainstream of society and providing an economy of scale, but its day is gone, as evidenced by the hilarious and/or pathetic attempts to integrate networking and multi-user capabilities into the Microsoft legacy.
---------------------
Although nominally a critique of CatB, the author quickly strays into much speculative and anecdotal critisism of Open Source itself, with Linux as a favorite subject. Strange that the article managed to quote just about everybody but Joe down at the pub, while not quoting CatB at all.
"Um, someone I know who wants to remain anonymous says Linux is insecure and unstable." Excuse me, but what are Jesse Berst's and Fred Moody's qualifications to address software development models and economics?
I had heard that First Monday's articles were 'peer reviewed'. In its published state, this would not have survived any peer review I'm familiar with.
---------------------
> I could open up 4 xterms and display more informaiton that you can with just a dumb text terminal.
:-). Seriously, GNU screen or virtual consoles would be better in this example, let X Window be X Window and serve apps where graphics are a fundamental and essential component.
At a cost of maybe 30x the RAM
> FLT. Flashing Led Terminal.
Heh. Where can I get one? Mmmmmm...das blinkenlights...
---------------------
> ...aren't most LAN's a combination of both stand-alone computers and main servers?
A LAN's *hardware environment* is a combination of stand-alone computers and main servers. The software environment of traditional PC's (DOS, Windows, OS/2 and their applications) is very much dedicated to the 'single-user, local configuration in real time' architecture, making any attempt to apply effective networking strategies to them a failure from the start.
> Most have a main file server where everyone's homedir is, but everyone's desktop machine has the ability to live apart from the network.
And do what? Access their files, access shared templates, read e-mail, print to a network printer? Server downtime is a catastrophe in any environment, but it is much easier to implement server redundancy than to maintain reduncancy throughout the desktops of an enterprise.
---------------------
You're correct that any average Unix/Linux system is well equipped to offer the major benefits of Network Computing. I see NC's as a way to overcome the maintenance and logistical mess that networking traditional DOS/Windows/OS2 involves, primarily from their single-user software architecture.
In my wired home, I practice the NC-esque "Monster-in-the-Closet" approach, where nearly all computer resources (and thus expenses) are concentrated in one master server, while the rest of the network are little more that SSH and/or X clients. Even an old and very feeble notebook of mine runs DOS and Kermit's Telnet client.
In such an environment, NFS is unnecessary and in fact becomes a maintenance and security burden. My example lab is too small to gain much from NIS, but a larger network would implement either NIS or some means of sharing relevant configuration data. Kerberos, done right (dedicated server, etc), can be useful for environments large enough to justify its requirements, but keep in mind that it requires client side support, which means custom versions of commercial software or custom compiles of Open Source packages. Much of Kerberos' functionality is available through SSH, anyway.
In a larger environment, I would implement multiple server-class systems for performance reasons, using either failover or individual servers dedicated to certain functions, with mirroring between them.
In short, the main benefits of NC architecture are available to the standard Unix and Linux system. NC's have been deprived of success in the market largely for 2 reasons: first, their market (DOS/Windows/OS2) can't grasp how a centralized network is *more* reliable that their current situation, second was the way NC's were prematurely associated with Java and Web browsers, neither of which were ready to serve up the goods at the time that the NC idea was first being promoted.
---------------------
I've been carrying Victorinox Swiss Army knives for longer that I've been working on PC's, and they are a godsend. Lately, I prefer the SwissChamp, although it's large and heavy enough to really require a belt holster instead of a pocket.
The Cybertool looks interesting if only for the Torx drivers, my SwissChamp has everything else and more.
---------------------
This decision, and the corresponding drop in SDRAM prices, was made inevitable by my purchase of a 128 meg upgrade for my server *just yesterday*.
---------------------
This is an interesting point, and perhaps it was a motivating factor behind Red Hat's pre-IPO offers to Linux developers. The more 'owners' they could gather who were knowledgable and sympathetic to Free/Open software, the less likely they could be subject to 'owners' who would demand a maximum-growth-and-profit-or-we'll-sue-you-to-deat h approach.
:-)
I'm more optimistic about SuSE, though, I think they'll survive this little "invasion" just fine...
---------------------
Turbo Linux started as a fork of the Red Hat distribution, much as Mandrake and so many others have done since. In other words, Red Hat do not 'assimilate' like the Borg, they *disseminate* to any and all who want to make use of their work.
Those who would promote their preferred distribution by ineptly riduculing another do a disservice to both.
---------------------
Well, releasing 'pre' versions is a habit they picked up from Red Hat :). As long as they follow Red Hat's practice of announcing the 'final' release as a free download, it shouldn't present a problem.
/. Comment Ever:
I'm just glad they didn't jump on glibc 2.1.2 right away.
---------------------
Most Annoying
"If I use Red$at Gnulix to make a Beowulf,
Not to be pedantic, Aaron, but the default RPM build target for Red Hat is the 486. The 'i386' that appears in the names of the packages refers to the architecture, which is a different parameter in RPM and doesn't necessarily reflect the actual compiler flags being used for the binary.
:) . /. Comment Ever:
I just wanted to clarify this, since many people see the i386 in Red Hat's RPMs and think RH binaries are the opposite of optimized (pessimized?
---------------------
Most Annoying
"If I use Red$at Gnulix to make a Beowulf,
SlashDot in a 'quiet period'... the mind boggles. /. Comment Ever:
---------------------
Most Annoying
"If I use Red$at Gnulix to make a Beowulf,
Wow! What color is the sky in *your* world, AC?
Garfinkel phrased his statement in such a way as to make it difficult to tell if he meant that Linux was a favorite target of crackers or a favorite *tool* of crackers. What part of that doesn't qualify as Fear, Uncertainty and Doubt? If this had been written by one of our beloved Ziff-Davis comlumnists, the answer would be assumed obvious.
BTW, you've just pushed another of my buttons, namely the whole 'opinion/critisism' defense. There is a common misunderstanding in forums such as this that 'everyone has a right to an opinion' and that 'any critisism is valid as long as it's protected by the 1st Amendment'. Balls to that! People only have a right to an INFORMED opinion, and critisism is valid ONLY to the extent that it can survive countering arguments. Neither Garfinkel nor you (nor I) can assume that we can express any old buggered-up idea and have it stand unchallenged.
---------------------
the SlashDot spellchecker:
Garfinkel took a few FUDdy cheap shots (Linux as 'a cracker's OS', too commercial, etc.) but the article was pretty good when he stuck to discussing the merits of the BSD's. In a way, he does everyone a favor by pointing out in a mainstream publication that Linux is a kernel, and that the components of a Linux distribution also come from GNU, BSD, et al. To listen to other articles in mundane media, one might think that Linus wrote the entire OS himself.
---------------------
the SlashDot spellchecker:
"The additional 1500 are not actually being ..."
laid-they are
Sorry, that just struck me as funny.
---------------------
the SlashDot spellchecker:
Only if all the kernel authors would get together and re-release Linux under the GPL (Generic Planet License). Maybe if the planet has a few moons, it would make a neat Beowulf...
---------------------
the SlashDot spellchecker:
>> Frankly, I think Red Hat would do consumers a favour, by unbundling the support service from the CD by itself. ...
...
...
Actually, Red Hat does this. From their web site, you can purchase the Red Hat Core package for $40US, which are the binary and source CD's but without the Applications Demo disk, printed manuals, support, etc. They also offer various separate support programs ranging from a few hundred dollars for per-incident to well into five figures for 24/7 unlimited support.
>> Are they waiting for enough market share and volume to simply fork from Linux, and produce some kind of
semi-proprietary version, in which free and non-free software are inextricable mixed, making the free content unfree?
After which they can capitalize on the market share Red Hat name, built using free resources, to disallow the freely copying
of the CD, and in effect insisting on a license fee per copy?
Red Hat once mixed proprietary code with Free Software, but have been moving away from this since 5.0. In 6.0, the only proprietary package left is Netscape. What you're suggesting of Red Hat more appropriately describes the practices of other commercial Linux distributors.
>> Therefore, I think the GPL should be adapted quickly, to prohibit that practice of putting GPLed software on a single
medium together with proprietary software, in effect removing from the buyer the freedom to copy the free content,
because he is not allowed to copy the proprietary content.
That sounds extraordinarily bad, and similar to the problem with version 1 of the GPL which version 2 and the LGPL were intended to correct.
---------------------
the SlashDot spellchecker:
Older notebooks make excellent terminals for a home LAN. I use two 486 notebooks, one running a lightweight install of Red Hat 6.0 w/ SSH and an X server. The other one has only a 486SLC (no math coprocessor) and an 80 meg HD and won't run Linux comfortably, so I use DOS-based Kermit and a TCP/IP packet driver for a Telnet session. Prior to Kermit, I used a serial link w/ Procomm Plus.
This is what I love about Linux, and Unix in general, you don't need to maintain and synchronize 5+ computers, just have access to one computer from everywhere.
---------------------
the SlashDot spellchecker:
The idea that anyone would steal a computer and then operate said computer on a public network without reformatting the drive or otherwise replacing all identifying parameters... well, I don't know whether to laugh or cry.
Serves them right, of course. It just goes to show that good wins because evil is generally stupid.
---------------------
the SlashDot spellchecker:
Does this mean someone can register "www.barbarastreisand.com"?
:-D
---------------------
Uh, why doesn't this story have Red Hat's logo?
---------------------
Before we begin tearing Red Hat apart for all the real and imaginary wickedness they've perpetrated against Free Software, quality and choice, and reading through all the redundant posts that defend them, I propose the following solution that should at least make this latest flamefest worthwhile.
A voluntary Red Hat Bashing Tax, payable, for want of a better metric, in lines of code contributed to Free Software and/or Open Source projects. Say for instance that someone feels compelled to toss around unfunny parodies of Red Hat's name, like "FUD Hat", "Red$at" or "RootHat". Each occurance would oblige the posting individual to submit 10 lines of acceptable code to the project of their choice. More extreme complaints, such as Red Hat "forcing" people to use buggy software or not allowing the posting individual to configure their system without a GUI, would require 100+ lines of code submitted. Complaints about RPM could be a special case, in which the complaining party could submit code that corrects the alleged deficiency in RPM, regardless of line count.
I believe such a policy would greatly benefit the Free Software community by generating a virtually infinite supply of new code while letting people vent all their frustrations and fears against a favorite target.
---------------------
Red Hat's had this in their Knowledge Base on their web site for at least a year now.
---------------------
...the last company to abandon disk-based copy protection, which did virtually nothing to hinder illegal copying but routinely punished paying customers,
...asinine 'look-and-feel' lawsuits against VIP, Paperback and ultimately Borland. The first two were driven from the market, QuattroPro operated under a legal cloud for years while Windows and Excel took the market. Lotus had the gall to complement Microsoft during the trial, for using macro translators instead of direct menu and macro compatibility,
...OS/2 ports that were so horrid and dysfunctional as to beggar the imagination,
...IBM's OS/2 support drying up and disappearing almost to the day that they acquired Lotus. No more Visualizer clients for OS/2 DB/2, the new DB/2 client became Approach...for Windows95. WTF?
Lotus has an extensive history for being far more trouble than they're worth. Leave them to be the bottom feeders in the Microsoft owned market they vigorously helped create.
---------------------