For instance take Oracle Applications, it is nearly impossible to install it on RedHat 7.0 or any glibc 2.2 based distro since the applications were built against 2.1.x. When you install this software it tries to relink itself with the correct libraries and fails miserably.
If there are substantial glibc 2.1-> 2.2 problems it is really poor coding on the part of the vendors. The use of private (but available) glibc functions was made impossible in the changeover.
There are a few models that will work in this case. First, the older version of glibc can be included with Oracle, and set LD_LIBRARY_PATH or LD_PRELOAD to load those libraries first. Then there is no problem.
Talk to your vendor. Ultimately, if you want to pay to use their software, they have a responsibility to ensure you can use it with some ease.
They cut a CD, because it is fairly easy, and nowadays everyone is doing it.
They go nowhere.
Record company comes along, and offers to make them rock stars, if they will sign over all copyright for 10 years and guarantee 7 new CDs, with an opt out for the recording company if the band crashes and burns. This is the dilemma. The band either joins the market forces and potentially becomes rich, or tries to make it on their own, and dies poor.
Access to recording studios has been costly, but never a limiting factor. Recording studios PROMOTE artists with ways and means beyond that of any garage band.
And we all KNOW good marketing and sales beats a good product every time. Look at Iomega and Syquest. CP/M and DOS. Heck, look at ANY Microsoft product.
Napster is nice for making the bands closer to their fans, but AIRTIME and PROMOTION with lots of CASH will continue to make bands.
Besides, how efficient is it, really, to burn food and take 30min. to get to work on a bicycle, when you can get there in half the time with one of these? To do a given amount of work in a given context, animal power is not always more enviro-friendly than machine power. If you want to bitch and moan, counter that having entertaining devices like the Wheelman disincentives forcibly packing ourselves into subway cars.
Animal power is 100% renewable, since animals consume foods that are 100% renewable. With respect to CO2/O2 balance, these work on a SHORT timescale compared to fossil fuels.
Animal power is also dramatically more efficient compared to fossil fuels, since our little cellular engines are quite efficient.
However, clean energy is generally a mask for electric engines that require fossil fuel burning at a power plant instead of on the road. They help places where loads of cars congregate, but they don't help the global picture.
If a substantial portion of people use SMALLER gas powered transportation, we save a LOT of total gas and energy usage, of the sort whose CO2/O2 cycle is thousands of years long. And perhaps it would be more popular (result in more savings) than trying to get people to use their own power.
Re:Affect hardware sales?
on
OS X on x86?
·
· Score: 2
By contrast, PCs are still incredibly broken AT architecture machines.
What matters ultimately is how they run, not how well they fulfill YOUR idea of what a server architecture should be. Apple has a real problem with their CPUs that other server RISC based CPUs do not have - they are about two times too slow to keep up with the AMDs and Intels of the world.
However the PC architecture may be, linux runs quite well already. So does FreeBSD. People will NOT switch to OS X when it comes out, but some Mac people will change to x86 Mac OS X. Because ultimately, the cheaper computer beats the better architecture every time. A lesson learned the hard way by IBM in the 80s, Apple in the 90s, and now mainframe Unices are learning it. If a computer can get the job done for $500, why spend $2000 to have 'a nice architecture' that provides no real functional advantage to 99% of the users ?
Companies are beholden to the bottom-line, the dollar. The onus is not on the phamaceutical companies on this one. We should expect resistance from them. Deaths of people in third world countries do not hurt quarterly reports - but free drug giveaways do.
What is really needed is a non-profit organization being given licensing over critical patents so that the drugs may be made for distribution in places where drug company prices are absurd. $10000/yr for AIDS medications seems high but affordable in San Francisco - but patently ridiculous in some African nations.
This charge MUST be backed strongly by the president of the US. He may claim it doesn't affect American interests. He may claim the corporations have a right to their patents. Or he may do the right thing. Time will tell.
The first step will be a re-affirmation of the trade watch list - not banning companies that provide generics that work around US patents to AIDS victims.
The copyright laws in the US are designed under the US Constitutional sentence
"To promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries;"
This means that authors are guaranteed certain rights about performance of their artwork, copyright of their writings, protection from conpetition for inventions... the purpose is to encourage authorship and invention through MONEY.
It has little to do with whether something MAY be copied free. Would we even have ssh right now if the RSA patent had not existed ? Would we even have mp3 encoding without patents ? (Personally I'd drop the mp3 patent, but it clearly had a large impact on mp3 spread).
Those are the arguments that need to be made. I think you can point at bad patents, and problems with duration of both copyright and patent, and use of copyright to justify EULAs... but I find it hard to believe one would argue that all software and all digital artwork should be public domain.
The RIAA has had their day. It is time to step down. May free art prosper.
The RIAA never really had control by having control of production. What they really have is control over marketing of young artists. Almost no artists that are successful today could have made it without record contracts fronting them a lot of money for an album, marketing the album, and providing a sales and distribution network.
And you know what - if you say screw the RIAA, all that is left are artists and songs. If you feel that marketing, sales, and distribution do not make bands popular, I have two words for you: Milli Vanilli.
I understand that you are a complete guru/Linux god, and are incapable of fucking up configuration of your box. If you want to host your own DNS and mail, then use a service that allows you to do that.
To put it mildly - it is pretty difficult to screw up and leave relaying open if you are not using sendmail as your SMTP. In qmail, for example, no relaying is the default.
As for DNS, it doesn't cause an issue with relaying/spam. You only list your box, you test it forwards and backwards, and you generally will know very quickly if it doesn't work properly.
Many professional sysadmins accidently leave open smtp relays on the internet
It is uncertain they should be called professionals then, Bare minimum, the longest O'Reilly book deals with sendmail conf at length.
How do they manage to keep track of the IPs used by major ISP's mail servers? Lucky guesses? Laborous investigation?
Step 1. Reverse DNS lookup on domain name.
Step 2. Check domain name entry for IP address
for MX value
Step 3. Compare IP address and MX address. If they are not equal, bounce mail.
Your mail server should be relaying through the SMTP server that PacBell assigned to you. Period.
Why is that ? Home DSL users cannot be trusted to configure an SMTP server ? As it is, when you get a static IP address, it is MXd to pacbell anyway. You cannot relay your email server through Pacbell. You need to host your own domain name to receive email(because of the MX problem). If you want to have reverse DNS work, you also need to pay another $100 to Pacbell, the fee they charge for adding 5 lines to their named configuration files - blind highway robbery.
A much easier solution, if you can configure a box, is to host your own DNS and SMTP. You save $100.
If Pacbell has a problem with that tough. The DSL line is a monopoly, but there are about 10 providers I could use other than them.
If my ISP is concerned with spam I invite them to check me for relaying, and port scan me too. Go for it.
To those thinking Bush may change something in the handling of the trial, consider
1) The states are co-prosecutors. They do not, and many have stated they will not, withdraw.
2) The DOJ won a resounding victory in the trial. Backing off would make you look like an idiot.
3) The number one senator in favor of the monopoly ruling is Orrin Hatch, the Republican Senator from Utah who is also chairman of the Judiciary Committee. Recall Caldera is in Utah, and they've already had their case with DRDOS against Microsoft. If there is one Republican Senator that can exert an enormous effect on judicial appointments, it is Orrin Hatch. Bush cannto piss him off.
It is the mission of the Attorney General, as prosecutor, to seek the STRONGEST ruling possible, and then the settlement/punishment that is in the best interests of the state (or country). The Microsoft case will be prosecuted to the fullest extent in the appeals case.
If the appeal is less strong than the initial ruling, which seems likely, then expect the Attorney General, maybe Ashcroft, to take some pro-Microsoft action. But he must be careful to do that in a way that keeps the states on board.
The mouse buttons can be VERY easily emulated via the yaboot boot manager. Just pass it the boot varible for the adb mouse button keys, and wa-la, you're there. Besides, you've got your hands on the keyboard, so emulating the 2 mouse buttons is NOT a problem. LinuxPPC/Yellowdog/DebianPPC _all_ have support for USB mouses and keyboards, so just plug in a friggin' mouse.
It would be REALLY nice to be able to take advantage of the Apple hardware without feeling like a second hand mouse user. Some people use their mice a lot. Some people use their keyboard a lot. You, apparently, are one of the latter. Good for you.
But the ergonomic benefits of actually having an intrinsic 3 button mouse should not be lost. I have two buttons on my HP Omnibook, and I REALLY REALLY wish it had three. In fact, so much so that I was willing to pay a little more for three buttons. Unfortunately, I wasn't willing to pony up for a Thinkpad, which is the only real option given my other constraints.
So take it easy on someone who actually appreciates having three buttons on the dern laptop. It could be a benefit, despite the fact that you live happily without it.
It's really sad to see the poster of an article so poorly informed, and such is often the case with Mr.Taco's posts.
An ad hominem attack is poor taste. Say something meaningful. Don't troll.
Microsoft licenses their software on a per-machine basis.
This is a legal sham. You PURCHASE the materials. At that point they are protected by copyright. They CAN be installed without accepting the EULA. A skilled hacker could hack the assembler, at least in theory.
You only need to accept licensing terms to do things not normally allowed by copyright, such as distributing copies. Microsoft would have you believe it is legally tenable to remove MORE of your rights as an owner of copyrighted material by the act of USING their copyrighted materials.
This is a set of rights INVENTED by the Microsoft, and perpetuated to financial gain by the software industry. It has never been legally thought out or challenged in court. And it really should.
Look, our system of copyright and patents existts to encourage limited time monopolies on novel inventions and writings. EULAs do not fit in this framework.
They fit in the framework of "use our product and bend over."
How hard would it be for a manufacturer to start selling NICs with a custom MAC address (if this doesn't exist already)?
Most do this already, although the instructions for setting MAC addresses are not commonly available. Why would someone like 3COM want to manufacture all their cards differently ?
They want to make all their cards identical, and use the easiest possible technique to set the MAC address.
But anyway, this strikes at a very critical issue. When I buy as licence of Whistler, do I have an intrinsic right to use that software on any machine I feel like, whenever I feel like ? Provided of course the software runs on only one machine at a time. Or does Microsoft have the right to dictate how and when I can use their software ??
Software is protected by copyright. This is akin to saying I can use the copyright in any one place.
Suppose we apply this analogy to books. Is it possible I would have the right to read a book at home, but not anywhere else ? And that the book would magically become pixie dust if I tried to read it at work instead ?? It is really quite silly. I have the right to read that book wherever and whenever I like. I can even make copies as long as I keep them to myself. I can even give that book to my friend, as long as I give him all my copies too. That is copyright law applied to books. But for software, somehow the rights are completely different.
Microsoft is perverting copyright protection into infinite time patent protection through schemes like this one. Copyright was never designed to offer such protection, and the patent system actually expires after some time.
It is really quite silly. Microsoft cannot devise a copyright protection system that also protects consumer rights. They want to protect machine exclusive licensing (which I believe to be illegal if ever challenged). Then they will go to a strict licensing model. Then they will tell you to bend over.
I don't know enough about the law to really comment on this, but if the federal lawsuit is killed by either the court of appeals or the supreme court (which after their election ruling would be unsurprising; unfortunately the court seems to have become totally politicized), the states' chances would probably be weakened.
This is a different issue. If the courts destroy the case it is no longer an issue of Bush affecting it.
But still, the states will press onward. Bush's best bet if he wants the case dropped is to do a poor job in court (it will be tough to do worse than Microsoft though based on the first trial), and try to lose or convince the states to back off.
In all likelihood he will let the case run its course, since he doesn't have a strong hand in guiding it anyway.
First, Bush would risk looking like a complete idiot. The justice department scored a resounding victory. How does it look if he stops the prosecution when he is three touchdowns ahead at halftime ?? He looks the fool.
Second, there is the Orrin Hatch factor. Don't forget the judiciary committee's power in judicial appointments. Jesse Helms has blocked appointments of black justices on racial grounds over and over again, not to be overridden during Clinton's tenure. It is a farce. Bush cannot risk losing a Republican on the Judicary Committee, especially the chair. Come to think of it, Bush cannot risk losing any Senate Republicans for two years (until the Democrats take over).
Third, even if Ashcroft (another racist, BTW) offers a deal after the appeal, the states do not have to agree. The federal government has only the power of suggestion to the state's attorney generals, and many of them are pursuing the case as is their duty - to win, and to win as big as possible, and then to consider whatever punishment/settlement is in the public's (and in this case the consumer's) best interests.
For all these reasons the case will ride unperturbed through the appeal. At that point expect Ashcroft to try to soften the settlement. In doing so he HAS to keep the states on board, so he cannot go too far.
Bush as President means almost nothing to the Microsoft case.
Render a MS-Breakup irrelevant? Is there anybody out there who thinks that the prosecution of MS is going to continue under Bush?
It is really simple.
Orrin Hatch (R) is head of the Senate Judiciary committee. Caldera is from his home state of Utah, and they just won a lawsuit about DrDOS, and view themselves as competitors in software with Microsoft. Orrin Hatch is one of the most outspoken persons in the nation against Microsoft as a monopolist.
Now, what are the odds that a Senate Judiciary Committee can block ANY appointment Bush tries to make if it is 50% Democrat and the Chairman takes an anti-Bush stance because of his actions against Microsoft ??
Also, the other party in the antitrust action is the states, and they will continue anyway. Really, Bush has nothing to do with it. Microsoft is going to have to fight their own fight in court.
They had a low quarterly revenue. This is quite easy to rationalize, and it is not about market saturation. It is about the economy slowing down, and people no longer buying new computers. Microsoft lives and dies off new computer sales. They OEM their Office quite inexpensively, and OEM their OS. People talk about poor Windows 2000/ME acceptance, but Windows 95, 98, and NT ALWAYS generated primary revenue off OEM sales. If OEM sales drop 20%, Microsoft loses money. But make no mistake about it - they still have an insane profit margin.
To get around their revenue loss, they squeeze the people who have the most invested in Windows - IT shops. These people have integrated networks of Windows, and Microsoft has them by the balls. No more volume licensing is a sure fire short term strategy. Over the long term it could hurt them.
But marginal markets are already thinking about linux as full shops. And it is happening more and more. Some use them only for servers, some use them for desktops as well.
But this move by Microsoft is about squeezing people who have no choice for more money. That is the way a monopolist uses the monopoly to their advantage. That is the Microsoft way.
1. Why would you think that the linux softupdates implementation would be any better? If there is a significant difference, it'd be in favor of FreeBSD, as McKusick is a FreeBSD developer and all.
The phase tree algorithm will not be superior initially. But it raises the possibilities that the file system could be made algorithmically much faster by utilizing the phase tree for file searches. FFS is already log structured, so this could be a wash. The phase trees could be faster. The phase trees could always have log structuring added later. Phillips is a long long time file system/database programmer, who has decades of experience with atomic soft update algorithms.
2. Sure, you could take major parts of the FreeBSD (or any BSD) kernel and start from there, but at what point would it just be easier to take BSD and add what Linux does better?
At this point linux kicks the crap out of any of the BSD ports for SMP machines. Especially multi-CPU multi-NIC machines. And the fine grained locking is quite arduous to add and debug. Since I use an SMP machine at work regularly, this matter a lot to me. For example, Linus specs his four CPU machines at 370% of the kernels with only one CPU compiled in.
To a single user machine, I don't think there are such strong arguments for linux over BSD now, especially as FFS + soft updates is a clear advantage. Where will the future lead - I don't know. But I am pretty sure I will be using linux, and possibly also FreeBSD. Both are great. Discussing the relative merits of one shouldn't automatically prompt one to say all development should be dropped for the other.
OK what else really needs to be in the kernel, what needs to be fixed.
There are two things done by FreeBSD that are really much better than linux.
Thing 1. A filesystem with soft updates. This creates a file system without journaling overhead that has a journalled file system's protection against power loss. It uses atomic updates of groups of files - see recent/. postings on Tux2. The linux versoin ought to be superior to FFS + soft updates, although the FreeBSD version works now. Journaled file systems should be added too, although I think soft updates, or phase trees, are better conceptually.
Thing 2. Scheduling. Linux scheduling is getting worse with each new generation kernel. That is supposed to be addressed in the next series. There was even the suggestion to start with the FreeBSD scheduling algorithm, and try to improve from there. Nothing like open source in action.
No backward compatibility break between rpm-3 and rpm-4 for debs. Not an issue for at least RH and, presumably, other RPM-based distros (though I don't know, so I can't swear to it). RH just runs out and makes an rpm-4 compatible version of rpm that runs on old distros. Easy. Fun. Besides, presumably all the reinstalls you've had to do when the unstable Debian tree breaks your system add up to at least some form of "upgrade", hence lack of backward compatibility?:-)
If you don't want to have to fuss with it, simply use the stable tree - that is what it is there for.
If you don't mind the occasional manual futzing with your box, the Debian unstable tree is for you.
And if you want a real pain in the ass, use a RedHat *.0 release. Totally unstable compared to Debian unstable. And RedHat calls it a production release.
Don't get me wrong - I administer RH and Debian boxes. I find I spend much less time screwing with the Debian boxes, and the initscripts and cron jobs (slocate, man) do more intelligent things. Like, for example, wiping/tmp clean on every boot. And/var/run.
Basically, I tried Debian as an experiment. It was slink, and a royal pain to install (potato is supposed to be much easier). Then it was a pain to get everything the way I liked it, as I didn't know much of apt and dpkg. Then, I used the box. And over a few months, I realized the Debian box was running more smoothly, and upgrading more easily, than the Redhat box. And this is mainly due to clean packaging and apt and debconf.
From Debconf's own description: "It's a way to get rid of all those annoying questions Debian packages often ask when they are installed; or a way to present them to the user via a varierty of UI's.". Gee. That's just swell. Instead of auto-detecting things like most RPMs do (following a Red Hat convention) and then letting you customize things if you want something special (rpm -i blah.src.rpm, edit.spec, rpm -bb.spec), you get to manually enter stupid info that could be auto-detected 99% of the time. Lovely.
Really, it does things like
1) install XFree 4.0
2) Prompt you whether you want to make XFree 4.0 the default
3) Remind you that Debian XFree 4.0 uses the binary/usr/X11R6/bin/Xfree86 instead of/usr/X11R6/bin/X, and the new config file is/etc/X11/XF86Config-4. The new install doesn't conflict with the old one that way (since 4.0 currently doesn't work for a lot of people).
4) Ask if you want to autoconfigure XF86COnfig-4 or use the existing version.
So you can see, you have multiple options, all of which any reasonable user might choose. With redhat you only get a broken system until you can hack up a new config file.
And for another example, for gdm, the debconf entry asks if you want to keep or overwrite the gdm.conf file.
More testing. Again, not a.rpm-.deb issue. This is a distribution issue. If you're claiming that.debs inherently get more testing than.rpms, than I claim that you're full of it. To be fair, if you're slamming RH 7.0, you're on target.
Really, this is critical. Debian packages get a little testing by the maintainers, and sometimes others who download them from a web page. Then, they get placed in unstable. Thousands upon thousands download unstable (since apt works so well), and an open bug reporting system checks all the bugs. The package is changed again and again until it is assumed everyone in stable can use the package without issue. Then it is moved to stable.
Redhat is trying something similar with rawhide, but it gets a few orders of magnitude less testers.
Actually this has a ton to do with reliability. The entire point of having a distribution is that they will ensure that packaging is reliable and consistent. When third party vendors start offering Redhat RPMs, or Mandrake RPMs, or you start using contrib RPMs, package dependencies get less consistent, and the machine is just not as clean. I get everything I need at debian.org in woody non-us and non-free, all packaged by debian maintainers.
Really? Mmm... Do you also purchase only Microsoft software, avoiding those nasty "other" vendors that prevent you from having a "reliable and consistent" experience? Because you could replace "Debian" with "Microsoft" there and have a sentence that could come from any number of idiot NT admins that I've run into.
Point well taken, but the environment is completely different. Everything I need under linux is free (as in speech), and can be packaged by a single party. And the single party is not a vendor with a monetary incentive to lock me in, but a collaboration of hundreds of volunteers.
There is nothing preventing RedHat from packaging everything that is packaged as RPMs for RedHat from third party vendors. The packaging basically consists of taking the installation files, and wrapping them together, setting sane dependencies, and pre/post install/uninstall scripts where appropriate.
Put another way, in a Microsoft world one of the single largest sources of OS trouble is third party software clobbering system dlls. If Microsoft could package that software properly, Windows would work much better. Of course, in a Windows world the third party software is actually competition, so this would never work.
In free software the third party software is a welcome addition to strengthen the distro. See the difference yet ??
The.deb/.rpm battle ended a long, long time ago. RPM won. How many people do you see coming out with.debs of their software? Sure, they *exist*...but at a 1:20 ratio to RPM producers.
This may surprise you, but Debian as a distribution is still growing rapidly. And the net number of people converting from RPM based systems to deb based systems is substantial.
But that is beside the point. Debian will continue. Its volunteer packaging force is growing rapidly, as is its user and testing base. There is no monetary bottom line affecting the distribution - only the joy of having free software that works the way it should. As long as the packagers keep packaging, the distro will continue. As long as the userbase keeps expanding, it will grow stronger and stronger.
Slashdot Mirror
For instance take Oracle Applications, it is nearly impossible to install it on RedHat 7.0 or any glibc 2.2 based distro since the applications were built against 2.1.x. When you install this software it tries to relink itself with the correct libraries and fails miserably.
If there are substantial glibc 2.1-> 2.2 problems it is really poor coding on the part of the vendors. The use of private (but available) glibc functions was made impossible in the changeover.
There are a few models that will work in this case. First, the older version of glibc can be included with Oracle, and set LD_LIBRARY_PATH or LD_PRELOAD to load those libraries first. Then there is no problem.
Talk to your vendor. Ultimately, if you want to pay to use their software, they have a responsibility to ensure you can use it with some ease.
Young band makes it on the local scene.
They cut a CD, because it is fairly easy, and nowadays everyone is doing it.
They go nowhere.
Record company comes along, and offers to make them rock stars, if they will sign over all copyright for 10 years and guarantee 7 new CDs, with an opt out for the recording company if the band crashes and burns. This is the dilemma. The band either joins the market forces and potentially becomes rich, or tries to make it on their own, and dies poor.
Access to recording studios has been costly, but never a limiting factor. Recording studios PROMOTE artists with ways and means beyond that of any garage band.
And we all KNOW good marketing and sales beats a good product every time. Look at Iomega and Syquest. CP/M and DOS. Heck, look at ANY Microsoft product.
Napster is nice for making the bands closer to their fans, but AIRTIME and PROMOTION with lots of CASH will continue to make bands.
Besides, how efficient is it, really, to burn food and take 30min. to get to work on a bicycle, when you can get there in half the time with one of these? To do a given amount of work in a given context, animal power is not always more enviro-friendly than machine power. If you want to bitch and moan, counter that having entertaining devices like the Wheelman disincentives forcibly packing ourselves into subway cars.
Animal power is 100% renewable, since animals consume foods that are 100% renewable. With respect to CO2/O2 balance, these work on a SHORT timescale compared to fossil fuels.
Animal power is also dramatically more efficient compared to fossil fuels, since our little cellular engines are quite efficient.
However, clean energy is generally a mask for electric engines that require fossil fuel burning at a power plant instead of on the road. They help places where loads of cars congregate, but they don't help the global picture.
If a substantial portion of people use SMALLER gas powered transportation, we save a LOT of total gas and energy usage, of the sort whose CO2/O2 cycle is thousands of years long. And perhaps it would be more popular (result in more savings) than trying to get people to use their own power.
By contrast, PCs are still incredibly broken AT architecture machines.
What matters ultimately is how they run, not how well they fulfill YOUR idea of what a server architecture should be. Apple has a real problem with their CPUs that other server RISC based CPUs do not have - they are about two times too slow to keep up with the AMDs and Intels of the world.
However the PC architecture may be, linux runs quite well already. So does FreeBSD. People will NOT switch to OS X when it comes out, but some Mac people will change to x86 Mac OS X. Because ultimately, the cheaper computer beats the better architecture every time. A lesson learned the hard way by IBM in the 80s, Apple in the 90s, and now mainframe Unices are learning it. If a computer can get the job done for $500, why spend $2000 to have 'a nice architecture' that provides no real functional advantage to 99% of the users ?
Why should we expect companies to give ??
Companies are beholden to the bottom-line, the dollar. The onus is not on the phamaceutical companies on this one. We should expect resistance from them. Deaths of people in third world countries do not hurt quarterly reports - but free drug giveaways do.
What is really needed is a non-profit organization being given licensing over critical patents so that the drugs may be made for distribution in places where drug company prices are absurd. $10000/yr for AIDS medications seems high but affordable in San Francisco - but patently ridiculous in some African nations.
This charge MUST be backed strongly by the president of the US. He may claim it doesn't affect American interests. He may claim the corporations have a right to their patents. Or he may do the right thing. Time will tell.
The first step will be a re-affirmation of the trade watch list - not banning companies that provide generics that work around US patents to AIDS victims.
The copyright laws in the US are designed under the US Constitutional sentence
"To promote the progress of science and useful arts, by securing for limited times to authors and inventors the exclusive right to their respective writings and discoveries;"
This means that authors are guaranteed certain rights about performance of their artwork, copyright of their writings, protection from conpetition for inventions... the purpose is to encourage authorship and invention through MONEY.
It has little to do with whether something MAY be copied free. Would we even have ssh right now if the RSA patent had not existed ? Would we even have mp3 encoding without patents ? (Personally I'd drop the mp3 patent, but it clearly had a large impact on mp3 spread).
Those are the arguments that need to be made. I think you can point at bad patents, and problems with duration of both copyright and patent, and use of copyright to justify EULAs... but I find it hard to believe one would argue that all software and all digital artwork should be public domain.
That is anarchy.
The RIAA has had their day. It is time to step down. May free art prosper.
The RIAA never really had control by having control of production. What they really have is control over marketing of young artists. Almost no artists that are successful today could have made it without record contracts fronting them a lot of money for an album, marketing the album, and providing a sales and distribution network.
And you know what - if you say screw the RIAA, all that is left are artists and songs. If you feel that marketing, sales, and distribution do not make bands popular, I have two words for you: Milli Vanilli.
You obviously know how to get good advertising.
I understand that you are a complete guru/Linux god, and are incapable of fucking up configuration of your box. If you want to host your own DNS and mail, then use a service that allows you to do that.
To put it mildly - it is pretty difficult to screw up and leave relaying open if you are not using sendmail as your SMTP. In qmail, for example, no relaying is the default.
As for DNS, it doesn't cause an issue with relaying/spam. You only list your box, you test it forwards and backwards, and you generally will know very quickly if it doesn't work properly.
Many professional sysadmins accidently leave open smtp relays on the internet
It is uncertain they should be called professionals then, Bare minimum, the longest O'Reilly book deals with sendmail conf at length.
How do they manage to keep track of the IPs used by major ISP's mail servers? Lucky guesses? Laborous investigation?
Step 1. Reverse DNS lookup on domain name.
Step 2. Check domain name entry for IP address
for MX value
Step 3. Compare IP address and MX address. If they are not equal, bounce mail.
Your mail server should be relaying through the SMTP server that PacBell assigned to you. Period.
Why is that ? Home DSL users cannot be trusted to configure an SMTP server ? As it is, when you get a static IP address, it is MXd to pacbell anyway. You cannot relay your email server through Pacbell. You need to host your own domain name to receive email(because of the MX problem). If you want to have reverse DNS work, you also need to pay another $100 to Pacbell, the fee they charge for adding 5 lines to their named configuration files - blind highway robbery.
A much easier solution, if you can configure a box, is to host your own DNS and SMTP. You save $100.
If Pacbell has a problem with that tough. The DSL line is a monopoly, but there are about 10 providers I could use other than them.
If my ISP is concerned with spam I invite them to check me for relaying, and port scan me too. Go for it.
To those thinking Bush may change something in the handling of the trial, consider
1) The states are co-prosecutors. They do not, and many have stated they will not, withdraw.
2) The DOJ won a resounding victory in the trial. Backing off would make you look like an idiot.
3) The number one senator in favor of the monopoly ruling is Orrin Hatch, the Republican Senator from Utah who is also chairman of the Judiciary Committee. Recall Caldera is in Utah, and they've already had their case with DRDOS against Microsoft. If there is one Republican Senator that can exert an enormous effect on judicial appointments, it is Orrin Hatch. Bush cannto piss him off.
It is the mission of the Attorney General, as prosecutor, to seek the STRONGEST ruling possible, and then the settlement/punishment that is in the best interests of the state (or country). The Microsoft case will be prosecuted to the fullest extent in the appeals case.
If the appeal is less strong than the initial ruling, which seems likely, then expect the Attorney General, maybe Ashcroft, to take some pro-Microsoft action. But he must be careful to do that in a way that keeps the states on board.
The mouse buttons can be VERY easily emulated via the yaboot boot manager. Just pass it the boot varible for the adb mouse button keys, and wa-la, you're there. Besides, you've got your hands on the keyboard, so emulating the 2 mouse buttons is NOT a problem. LinuxPPC/Yellowdog/DebianPPC _all_ have support for USB mouses and keyboards, so just plug in a friggin' mouse.
It would be REALLY nice to be able to take advantage of the Apple hardware without feeling like a second hand mouse user. Some people use their mice a lot. Some people use their keyboard a lot. You, apparently, are one of the latter. Good for you.
But the ergonomic benefits of actually having an intrinsic 3 button mouse should not be lost. I have two buttons on my HP Omnibook, and I REALLY REALLY wish it had three. In fact, so much so that I was willing to pay a little more for three buttons. Unfortunately, I wasn't willing to pony up for a Thinkpad, which is the only real option given my other constraints.
So take it easy on someone who actually appreciates having three buttons on the dern laptop. It could be a benefit, despite the fact that you live happily without it.
It's really sad to see the poster of an article so poorly informed, and such is often the case with Mr.Taco's posts.
An ad hominem attack is poor taste. Say something meaningful. Don't troll.
Microsoft licenses their software on a per-machine basis.
This is a legal sham. You PURCHASE the materials. At that point they are protected by copyright. They CAN be installed without accepting the EULA. A skilled hacker could hack the assembler, at least in theory.
You only need to accept licensing terms to do things not normally allowed by copyright, such as distributing copies. Microsoft would have you believe it is legally tenable to remove MORE of your rights as an owner of copyrighted material by the act of USING their copyrighted materials.
This is a set of rights INVENTED by the Microsoft, and perpetuated to financial gain by the software industry. It has never been legally thought out or challenged in court. And it really should.
Look, our system of copyright and patents existts to encourage limited time monopolies on novel inventions and writings. EULAs do not fit in this framework.
They fit in the framework of "use our product and bend over."
"First they ignore you.
Then they laugh at you.
Then the fight you.
Then you win."
-Mahatma Ghandi
How hard would it be for a manufacturer to start selling NICs with a custom MAC address (if this doesn't exist already)?
Most do this already, although the instructions for setting MAC addresses are not commonly available. Why would someone like 3COM want to manufacture all their cards differently ?
They want to make all their cards identical, and use the easiest possible technique to set the MAC address.
But anyway, this strikes at a very critical issue. When I buy as licence of Whistler, do I have an intrinsic right to use that software on any machine I feel like, whenever I feel like ? Provided of course the software runs on only one machine at a time. Or does Microsoft have the right to dictate how and when I can use their software ??
Software is protected by copyright. This is akin to saying I can use the copyright in any one place.
Suppose we apply this analogy to books. Is it possible I would have the right to read a book at home, but not anywhere else ? And that the book would magically become pixie dust if I tried to read it at work instead ?? It is really quite silly. I have the right to read that book wherever and whenever I like. I can even make copies as long as I keep them to myself. I can even give that book to my friend, as long as I give him all my copies too. That is copyright law applied to books. But for software, somehow the rights are completely different.
Microsoft is perverting copyright protection into infinite time patent protection through schemes like this one. Copyright was never designed to offer such protection, and the patent system actually expires after some time.
It is really quite silly. Microsoft cannot devise a copyright protection system that also protects consumer rights. They want to protect machine exclusive licensing (which I believe to be illegal if ever challenged). Then they will go to a strict licensing model. Then they will tell you to bend over.
I don't know enough about the law to really comment on this, but if the federal lawsuit is killed by either the court of appeals or the supreme court (which after their election ruling would be unsurprising; unfortunately the court seems to have become totally politicized), the states' chances would probably be weakened.
This is a different issue. If the courts destroy the case it is no longer an issue of Bush affecting it.
But still, the states will press onward. Bush's best bet if he wants the case dropped is to do a poor job in court (it will be tough to do worse than Microsoft though based on the first trial), and try to lose or convince the states to back off.
In all likelihood he will let the case run its course, since he doesn't have a strong hand in guiding it anyway.
No, the point was deeper than that.
First, Bush would risk looking like a complete idiot. The justice department scored a resounding victory. How does it look if he stops the prosecution when he is three touchdowns ahead at halftime ?? He looks the fool.
Second, there is the Orrin Hatch factor. Don't forget the judiciary committee's power in judicial appointments. Jesse Helms has blocked appointments of black justices on racial grounds over and over again, not to be overridden during Clinton's tenure. It is a farce. Bush cannot risk losing a Republican on the Judicary Committee, especially the chair. Come to think of it, Bush cannot risk losing any Senate Republicans for two years (until the Democrats take over).
Third, even if Ashcroft (another racist, BTW) offers a deal after the appeal, the states do not have to agree. The federal government has only the power of suggestion to the state's attorney generals, and many of them are pursuing the case as is their duty - to win, and to win as big as possible, and then to consider whatever punishment/settlement is in the public's (and in this case the consumer's) best interests.
For all these reasons the case will ride unperturbed through the appeal. At that point expect Ashcroft to try to soften the settlement. In doing so he HAS to keep the states on board, so he cannot go too far.
Bush as President means almost nothing to the Microsoft case.
Render a MS-Breakup irrelevant? Is there anybody out there who thinks that the prosecution of MS is going to continue under Bush?
It is really simple.
Orrin Hatch (R) is head of the Senate Judiciary committee. Caldera is from his home state of Utah, and they just won a lawsuit about DrDOS, and view themselves as competitors in software with Microsoft. Orrin Hatch is one of the most outspoken persons in the nation against Microsoft as a monopolist.
Now, what are the odds that a Senate Judiciary Committee can block ANY appointment Bush tries to make if it is 50% Democrat and the Chairman takes an anti-Bush stance because of his actions against Microsoft ??
Also, the other party in the antitrust action is the states, and they will continue anyway. Really, Bush has nothing to do with it. Microsoft is going to have to fight their own fight in court.
This is classic monopolist action.
They had a low quarterly revenue. This is quite easy to rationalize, and it is not about market saturation. It is about the economy slowing down, and people no longer buying new computers. Microsoft lives and dies off new computer sales. They OEM their Office quite inexpensively, and OEM their OS. People talk about poor Windows 2000/ME acceptance, but Windows 95, 98, and NT ALWAYS generated primary revenue off OEM sales. If OEM sales drop 20%, Microsoft loses money. But make no mistake about it - they still have an insane profit margin.
To get around their revenue loss, they squeeze the people who have the most invested in Windows - IT shops. These people have integrated networks of Windows, and Microsoft has them by the balls. No more volume licensing is a sure fire short term strategy. Over the long term it could hurt them.
But marginal markets are already thinking about linux as full shops. And it is happening more and more. Some use them only for servers, some use them for desktops as well.
But this move by Microsoft is about squeezing people who have no choice for more money. That is the way a monopolist uses the monopoly to their advantage. That is the Microsoft way.
1. Why would you think that the linux softupdates implementation would be any better? If there is a significant difference, it'd be in favor of FreeBSD, as McKusick is a FreeBSD developer and all.
The phase tree algorithm will not be superior initially. But it raises the possibilities that the file system could be made algorithmically much faster by utilizing the phase tree for file searches. FFS is already log structured, so this could be a wash. The phase trees could be faster. The phase trees could always have log structuring added later. Phillips is a long long time file system/database programmer, who has decades of experience with atomic soft update algorithms.
2. Sure, you could take major parts of the FreeBSD (or any BSD) kernel and start from there, but at what point would it just be easier to take BSD and add what Linux does better?
At this point linux kicks the crap out of any of the BSD ports for SMP machines. Especially multi-CPU multi-NIC machines. And the fine grained locking is quite arduous to add and debug. Since I use an SMP machine at work regularly, this matter a lot to me. For example, Linus specs his four CPU machines at 370% of the kernels with only one CPU compiled in.
To a single user machine, I don't think there are such strong arguments for linux over BSD now, especially as FFS + soft updates is a clear advantage. Where will the future lead - I don't know. But I am pretty sure I will be using linux, and possibly also FreeBSD. Both are great. Discussing the relative merits of one shouldn't automatically prompt one to say all development should be dropped for the other.
May they both live long and prosper.
OK what else really needs to be in the kernel, what needs to be fixed.
/. postings on Tux2. The linux versoin ought to be superior to FFS + soft updates, although the FreeBSD version works now. Journaled file systems should be added too, although I think soft updates, or phase trees, are better conceptually.
There are two things done by FreeBSD that are really much better than linux.
Thing 1. A filesystem with soft updates. This creates a file system without journaling overhead that has a journalled file system's protection against power loss. It uses atomic updates of groups of files - see recent
Thing 2. Scheduling. Linux scheduling is getting worse with each new generation kernel. That is supposed to be addressed in the next series. There was even the suggestion to start with the FreeBSD scheduling algorithm, and try to improve from there. Nothing like open source in action.
Where can I get an off the shelf G4 PPC linux box with a real mouse ?
No backward compatibility break between rpm-3 and rpm-4 for debs. :-)
/tmp clean on every boot. And /var/run.
.spec, rpm -bb .spec), you get to manually enter stupid info that could be auto-detected 99% of the time. Lovely.
/usr/X11R6/bin/Xfree86 instead of /usr/X11R6/bin/X, and the new config file is /etc/X11/XF86Config-4. The new install doesn't conflict with the old one that way (since 4.0 currently doesn't work for a lot of people).
.rpm-.deb issue. This is a distribution issue. If you're claiming that .debs inherently get more testing than .rpms, than I claim that you're full of it. To be fair, if you're slamming RH 7.0, you're on target.
Not an issue for at least RH and, presumably, other RPM-based distros (though I don't know, so I can't swear to it). RH just runs out and makes an rpm-4 compatible version of rpm that runs on old distros. Easy. Fun. Besides, presumably all the reinstalls you've had to do when the unstable Debian tree breaks your system add up to at least some form of "upgrade", hence lack of backward compatibility?
If you don't want to have to fuss with it, simply use the stable tree - that is what it is there for.
If you don't mind the occasional manual futzing with your box, the Debian unstable tree is for you.
And if you want a real pain in the ass, use a RedHat *.0 release. Totally unstable compared to Debian unstable. And RedHat calls it a production release.
Don't get me wrong - I administer RH and Debian boxes. I find I spend much less time screwing with the Debian boxes, and the initscripts and cron jobs (slocate, man) do more intelligent things. Like, for example, wiping
Basically, I tried Debian as an experiment. It was slink, and a royal pain to install (potato is supposed to be much easier). Then it was a pain to get everything the way I liked it, as I didn't know much of apt and dpkg. Then, I used the box. And over a few months, I realized the Debian box was running more smoothly, and upgrading more easily, than the Redhat box. And this is mainly due to clean packaging and apt and debconf.
From Debconf's own description: "It's a way to get rid of all those annoying questions Debian packages often ask when they are installed; or a way to present them to the user via a varierty of UI's.". Gee. That's just swell. Instead of auto-detecting things like most RPMs do (following a Red Hat convention) and then letting you customize things if you want something special (rpm -i blah.src.rpm, edit
Really, it does things like
1) install XFree 4.0
2) Prompt you whether you want to make XFree 4.0 the default
3) Remind you that Debian XFree 4.0 uses the binary
4) Ask if you want to autoconfigure XF86COnfig-4 or use the existing version.
So you can see, you have multiple options, all of which any reasonable user might choose. With redhat you only get a broken system until you can hack up a new config file.
And for another example, for gdm, the debconf entry asks if you want to keep or overwrite the gdm.conf file.
More testing.
Again, not a
Really, this is critical. Debian packages get a little testing by the maintainers, and sometimes others who download them from a web page. Then, they get placed in unstable. Thousands upon thousands download unstable (since apt works so well), and an open bug reporting system checks all the bugs. The package is changed again and again until it is assumed everyone in stable can use the package without issue. Then it is moved to stable.
Redhat is trying something similar with rawhide, but it gets a few orders of magnitude less testers.
Actually this has a ton to do with reliability. The entire point of having a distribution is that they will ensure that packaging is reliable and consistent. When third party vendors start offering Redhat RPMs, or Mandrake RPMs, or you start using contrib RPMs, package dependencies get less consistent, and the machine is just not as clean. I get everything I need at debian.org in woody non-us and non-free, all packaged by debian maintainers.
.deb/.rpm battle ended a long, long time ago. RPM won. How many people do you see coming out with .debs of their software? Sure, they *exist*...but at a 1:20 ratio to RPM producers.
Really? Mmm... Do you also purchase only Microsoft software, avoiding those nasty "other" vendors that prevent you from having a "reliable and consistent" experience? Because you could replace "Debian" with "Microsoft" there and have a sentence that could come from any number of idiot NT admins that I've run into.
Point well taken, but the environment is completely different. Everything I need under linux is free (as in speech), and can be packaged by a single party. And the single party is not a vendor with a monetary incentive to lock me in, but a collaboration of hundreds of volunteers.
There is nothing preventing RedHat from packaging everything that is packaged as RPMs for RedHat from third party vendors. The packaging basically consists of taking the installation files, and wrapping them together, setting sane dependencies, and pre/post install/uninstall scripts where appropriate.
Put another way, in a Microsoft world one of the single largest sources of OS trouble is third party software clobbering system dlls. If Microsoft could package that software properly, Windows would work much better. Of course, in a Windows world the third party software is actually competition, so this would never work.
In free software the third party software is a welcome addition to strengthen the distro. See the difference yet ??
The
This may surprise you, but Debian as a distribution is still growing rapidly. And the net number of people converting from RPM based systems to deb based systems is substantial.
But that is beside the point. Debian will continue. Its volunteer packaging force is growing rapidly, as is its user and testing base. There is no monetary bottom line affecting the distribution - only the joy of having free software that works the way it should. As long as the packagers keep packaging, the distro will continue. As long as the userbase keeps expanding, it will grow stronger and stronger.