Wednesday, September 19, 2007

Linux.com Repost Flood

The following is a repost of a series of posts made on Linux.com by myself. Several of these comments were made in response to a spammer, or spammers, and the originating threads were terminated. In the interest of preservation I am putting my comments down here. The substance of some of the posts is documented where I have directly quoted the spammer(s).

*************************

well. First thing, nobody was vulgar or obscene until you uttered the first obscenity. So, our gut reaction is going to be that we don't want you to begin with. The second portion of your statement that appears to be designed to inflame opposition is to indicate that Linux is a niche OS. Well, lets talk realistic market share for a minute. Both Xandros and Linspire have each surpassed the total number of retail boxes sold as Apple has. Linspire alone sold more Personal Computers through WalMart than Apple did in all of it's retail channels for over 3 years in a row. Dell, Levono, and now HP are all now offering Linux as a Desktop option, for desktop computers. Linux and Apache just don't rule the server market, the domination is total with the nearest competitor almost 40-50% behind. Okay. We don't have a reliable metric to prove that Linux has over 100million desktop users. We know that the figures cited by Distrowatch are meaningless. We, do, however, have some insight into the server sales of IBM and Sun Microsystems, as well as figures from AMD, and Intel. We know that Linux is the majority option in the server market. We also know that Solaris and other BSD's are not the majority options, and haven't been for several years.


So, Linux isn't in a niche. OpenBSD, FreeBSD, NetBSD, Novell Netware, Solaris, and Apple's own Server business, those are niche products.

The final portion that seems designed to inflame is the line that we did not try to convince you to stay with Linux. Here's a clue for you... We Can't. We cannot make you decide anything. The source code is there for you to look at. The mailing lists are there for you to look at. The developers are there for you to talk to. We cannot convince anybody to chose Linux, or any Open Source product. We, however, can give you the choice to use our product. We also then must recognize your choice to not use our product. I, for one, hate Gnome. I think that it's a dead-end project and I'm fairly convinced that the Gnome-Dev team has about as much business designing a desktop interface as I do. However, I will not stop you from using Gnome. I will try to give reasons why you shouldn't, and I will give you the opinion that XFCE is "Gnome Done Right," but that's the extent. If somebody were to tell you that you could not use Gnome... then I would have to take issue with that. It is your choice.

From my point of view, having read the comments stated, it seems to me that you have already decided that Linux is not for you. That's fine. You have expressed your desire to go to another platform. Fine, great. Other platforms offer competition, and if you want to use them, that's your choice.

What you need to ask yourself is this: What keeps you using Linux? Why would you continue to use the product if you don't like the way it is being managed? What do Solaris and the other BSD's lack that have kept you from moving to them to begin with? Why would you indicate that moving to another platform is a threat that we should somehow take seriously, and that if we don't cater to your whims that such a lack of action would be a bad thing? Linux has hundreds of millions of users, and the product is growing on a daily basis. Losing one or two, or even several hundred isn't going to hurt Linux.

The fact is, we don't have any reason to convince you. We don't even know who "you" are. I find myself echoing the sentiments of other posters. Don't let the door hit yourself on the way out.


****************


Linux users' arrogance can be so high, it actually harms the OS

Deary. That is true of all operating systems, not just Linux. I can easily point to rants made by Theo de Raadt where he defines pure arrogance. I can link to posts on Ubuntu forums where the response consisted of "RTFM" or absolutely no response at all. I can point to newsgroup postings and forums and mailing lists for Apple Mac developers that will make your skin crawl with how superior they think "their OS" is. I can point directly at Microsoft's own top executives and how they brush off the continuous assaults made upon them by Open Source supporters, computer security experts, and even national governments. I can easily bring up hundreds of forum postings about common encounters with Best Buy and Circuit City employees who snob down everything that isn't Windows, or the most recent PC World disaster where the local store refuses to fix a hardware problem after being ordered to do so by Corporate.

The fact is this: Users arrogance alone does not directly help or hinder the overall OS. Now, if you want to think that it does, and you want to list a singular source that I quite frankly have never heard of, hey, fine. That's your social circle, and that's your business.

Speaking for myself, I realize that I can't control everybody's opinions. I can't control everybody's actions. Some people are going to the hole in the south end of a north bound donkey. What I can do is this: Ignore them as best I can, and let my own words and my own actions carry their weight.


***************


hang out on Ubuntu's forums or IRC channel for 5 minutes. I think it'sjust as bad, if not possibly worse. Every group has their amount of... "bad" people. The BSD's I think are more noticeable since they are much smaller in terms of users and developers, which means that covering the BSD side means having to cover the Drama Queens and Kings. It also, I think, doesn't help that the leader(s) of OpenBSD seem intent on being screaming raving lunatics every chance (they) get. As I said in another post, I couldn't tell you who leads FreeBSD or NetBSD, and I understand they have a much larger share of the BSD market than OpenBSD. I can tell you a couple of people involved with Solaris, but other than Sun's CEO, not anybody actually important to the project.

nyways, back to the (semi)point. Covering Linux and Windows? There's enough people trying to make good without being jerks that news pages are not flooded with "he said"/"she said" shenanigans. In several ways, it is about perception.


******************


Torvalds already says he doesn't care about user freedoms, only efficient software.

Hmm... I'm really not sure where to start with this. My first reaction to is point to the numerous comments Linus has made during the GPL3 development process. As I read it, Linus seemed to be concerned about the freedom of the software itself. Specifically in an interview with Forbes Linux stated that he didn't like the GPL3 because it placed limits on what could be done with the software.
For example, the GPLv2 in no way limits your use of the software. If you're a mad scientist, you can use GPLv2'd software for your evil plans to take over the world ("Sharks with lasers on their heads!!"), and the GPLv2 just says that you have to give source code back. That certainly sounds like Linus "cares" about user freedoms.

Does Torvalds even want to end Microsoft's monopoly on the desktop,

I'm not really sure where Microsoft enters into this discussion. Linus said this in an interview with LinuxWorld :
I don't actually see it as a battle. I do my thing because I think it's interesting and worth doing, and I'm not in it because of any anti-MS issues. I've used a few MS products over the years, but I've never had a strong antipathy against them. Microsoft simply isn't interesting to me. The thing is, Linus just doesn't care about Microsoft, as do many of other open source developers. They don't aim for Microsoft because Microsoft isn't a benchmark. Lets be honest, anybody involved in Open Source Communities probably has the opinion that Microsoft writes sloppy and buggy code. Focusing on Microsoft as a target? Well, quite frankly, that would be like aiming for an open sewer pipe if you wanted to go swimming. Speaking for myself, I'm glad that Linus doesn't find Microsoft interesting.

This isn't to say that many of us don't have an interest in breaking Microsoft's monopoly on the desktop, but that battle is actually going to fall to KDE, XFCE, and other quality desktop environments in providing user friendly desktops. That battle is going to fall to the X.org development teams to provide functional dual desktop support and better graphics support. That battle is going to fall to Compiz and Red Hat's AIGLX developers, and it's going to fall on the Samba team, and it's going to fall on individual distributions.

Saying that Linus alone is responsible for the battle against Microsoft is like trying to pin Economy or Gas prices on the sitting President.Reality doesn't work like that. Congress is the target.

The Kernel team has their job, yes, but they are only a small portion of the overall scheme. That, however, doesn't address the real issue of Microsoft's Monopoly. Keep in mind that many business's have a saying, nobody got fired for buying Microsoft. Keep in mind that you, personally, vote with your wallet. I don't expect Linus to speak up for me. I speak up for myself, and the fact is, more Linux users are speaking up. More people are beginning to ask Independent Software Vendors just how independent they are. That is what is changing the landscape, not the work of just one person.

or is Linus content to just maintain the status quo with GNU/Linux withering on the vine?

Um. I'm sorry, but I have absolutely no idea what in the world you are talking about here. Last time I checked Linux usage was up in the Cell Phone market, was up in the embedded device market, and was up in the retail PC market. I've seen hundreds of headlines, stories, and editorials about Ubuntu Linux going on Dell system; about Novell Suse going on Levono systems; Levono looking for an alternative Linux supplier; and HP entering the Linux desktop market. There has been the RadeonHD driver, an official Open-Sourced driver from AMD, and the code dump from Intel for their graphics drivers. I don't know where you get the idea that Linux is withering, the only market that it appears to have decreased in is the web-server market. Even that decrease is not a real decrease as the number of servers not using Linux+Apache, but moving to other services like lighttpd, and the google web hosting service.

For many of us, this is a vote of no confidence in Linus Torvalds' leadership. This is a challenge for Linus to get with the program or get out of the way.

Um. Okay. Get with what program? Do you really want Linus to start targeting Microsoft? I'm also failing to see how this is a vote of No Confidence. Last time I checked, Linux was getting results. OpenBSD, FreeBSD, Solaris, Novell Netware... and so on... I see those products loosing results. Theo De Raadt's latest outrage against the Atheros driver only served to turn several users off from even looking at BSD. It's one thing to be in the right. It's another thing to rant and spew after the problem has been addressed promptly and politely. Now, I couldn't tell you who on earth runs NetBSD or FreeBSD. I could tell you that Jonathon Swartz helps with Solaris over at Sun, but even Sun has found profit in selling Linux servers. I can, however, recognize Alan Cox, Linus T, Andrew Morton, and Ingo Molnar. The only reason I know a guy who used to do Linux driver development by the name of Con K. is because he was a media attention seeker who, in my view, threw a baby's fit when he didn't get his way.

Now, if you have somebody else in mind who would be good enough to lead and manage the Linux Kernel... by all means, speak up. But keep in mind that such a person would have to deal with the existing kernel driver team. I know that many of the Kernel Lieutenants have expressed that they do not want Linus's job. Nobody really wants to try to manage the kernel project. Reminds me of a line from the Sean Connery film, First Knight.
Once in a lifetime, you meet a man so fearless. No man can touch him. While you're waiting for him, you can practice on me. - Lancelot. Sure. Linus may not be what you want, but he's what we have. And like it or not, there is nobody else waiting in the wings.


***********


Audio, in general, is an issue with Linux. Yes, lots of sound cards are supported. But, exactly what is the audio standard? I can tell you that OpenGL is the standard for rendering in 3D, but I'm not aware of any consensus on the default Audio Input. As I see it Linux has two primary audio systems, ALSA and OSS. I also have seen libraries relating to SDL and OpenAL, and I'm aware of Gstreamer, Xine, and I'm fairly certain there is another engine out there.

If you don't get the point, trying to pin the performance of a single application such as Audio playback on a single point of the kernel isn't something that can be done. Yes, the scheduler is going to have an impact on it, but quite frankly, if you are rapidly opening programs? That's more likely to cause a run-in with your RAM access and your hard-drive access.

A quick experiment you can try is to have Microsoft Windows boot up and launch 10 programs at once. Have the 7th program be an audio sound.

Since I know you won't do the experiment as you are no doubt a loyal Linux user without Windows installed on any computers... so I'll go ahead and spoil the story for you. Windows Slows down too.

Now, I know for a fact that if I run Synaptic and have it update my system, I don't get any audio playback loss on my systems. Does it mean that I have more knowledge than you and that I know how to work my systems? Perhaps it does... but Perhaps it doesn't. Each package management system runs differently. Now, I know that OpenSuse's package manager is a system resource hog. I've learned the hardway that when it runs, turn everything off.

Now... if I, somebody who only has OpenSuse on one machine and Mepis on all the others can figure that out... I'm really surprised that it isn't common knowledge.


**************



The time and effort that people put into developing open source software is amazing, but people still need to pay their bills and feed their families. If a code change is proposed that would benefit performance on enterprise servers at the expense of performance on desktop computers, the developers must favor their only source of income.

Performance is performance is performance. Several years ago Intel said 64bit was meaningless on the desktop and continued right on with making 32bit Pentium4 chips while AMD was pushing Athlon64 out the door. Years before that hard-drive vendors said RAID was meaningless on the desktop and finding a hardware RAID card could run into thousands of dollars... now, both Radeon Xpress and Nvidia Nforce are doing RAID on chip in motherboards under $60(US).

The idea that changes made to increase performance in the server market is going to decrease performance on desktop computers is... ludicrous. Performance is performance is performance. Okay, granted, home computers are only now getting the ability to handle 4 threads at once in the Athlon64 Quad-Core Barcelona. Big deal. You could get Servers with 32, 64, and 128 processors. However, all of the work making the applications SMP aware for those massive systems has a direct impact on improving performance on the small systems. Users can take advantage, today, of AMD Barcelona Quad-Core, instead of waiting for software to be made SMP aware.

There is a trickle down effect. Technology developed for theserver market will eventually make it to the home market. Maybe not today, but it will get here. The work done by Evans and Sutherland on their massive R300 Based GPU boxes with 64 GPU's is coming right back and is being used in CrossFire management and setup.

I think the idea that Server performance was hurting Desktop performance was started by Con K. What I don't know is if Con K. was out to distribute F.U.D. because he didn't get his way or if he actually thought he had a legitimate point. I personally agree with Linus. 3D gaming isn't the only performance metric out there. I care a lot about hard-drive efficiency, memory usage, and processor efficiency. I'd rather have a kernel that is built to handle any situation I can throw at it in a decent manner, than have a kernel that can handle one situation really well, while having a hard-time at others. This isn't to say I don't care about 3D performance. But, I know enough that 3D performance comes more from the graphics drivers than the kernel itself. If OpenGL is setup correctly in the drivers, the kernel itself shouldn't have that much impact on the final frame rate and response.



************

Fork the Linux kernel to get REISER4.

hmm? Okay, I must admit I'm a little confused as to how forking the kernel involves getting a hard-drive support format implemented. Never mind a format that is effectively dead as is. REISER4 offered no significant performance or feature improvement over EXT3 (braces for the impact of synthetic benchmarks), and EXT4 has been implemented since 2.16.19. Adding REISER4 support at this point and time is um... Well, I don't want to say meaningless, but it would be a useless manuever. As is, REISER support overall has been being dropped by several shipping distributions.

Fork the Linux kernel to get GRAPHICS that work (not sabotaged crap).

Okay Mate. I have not idea what in the world you are on about here. Whose graphics are sabotaged, and how in the world does the kernel have anything to do with that? I'm afraid I'm just going to have to pass on trying to answer this without further data.

REISER4 is a great filesystem.

Um. No. It wasn't. REISER4 was an improvement to REISER3, so it wasn'tbad. But it wasn't some magical miracle pill that would fix all the problems associated with hard-drive access and response. Note the key word,
was.

GRAPHIC drivers need to work.

Again, what in the world are you on about? I have ATi cards from Radeon PCI AIW to Radeon x1900, and Nvidia cards from TNT2 to Geforce 7900. I have laptops with Intel GPU's and Radeon Xpress GPU's. All of them work great... if I use the appropriate drivers. Thing is, Graphic drivers don't fall upon the kernel developers. The drivers fall upon...

THE X WINDOWS DEVELOPMENT TEAMS : THEY ARE AT X.ORG

Yes, it's the kernels responsibility to handle the driver requests, but it's not the kernel teams responsibility to build each and every driver. It is up to the vendors who make the hardware, and up to those involved with X.org. If you honestly believe that Linus, or the kernel development team for that matter, are alone in responsibility for graphics support, you simply do not understand how a Linux distribution is built. If ATi does their job properly with building the Catalyst Drivers, I should be able to move from one driver to another without changing the kernel. I also look to the people who create my distribution to provide an appropriate ATi driver install. I don't look to ATi to provide an installer for my OS.

There are many LinuxKernel SABOTEURS, that work to prevent the Linux ever truly becoming a threat to Microsoft.

Okay, lets presume that this statement is somehow true. Prove it. (btw, fixed your spelling mistake)

The fact is, Linux as an OS is a threat to Microsoft right now. It is more stable. It is faster on the same hardware. It has far better out of the box compatibility. It actually has a real working 3D desktop. Several distributions are scoring OEM wins and getting Linux as a factory pre-installed option. The LiveCD OS started by Knoppix and blown open by Mepis has ccreated countless derivatives, one such Mepis Knockoff being found on Dell computers now, Mark Shuttleworths Ubuntu Linux.

Where you get the idea that the kernel alone keeps Linux from being a threat to Microsoft is beyond me. The kernel team does their job, and from a technological aspect, they passed anything Microsoft could do over a decade ago.

The real problem is I think somebody has been lying to you. Now, I don't know who. I don't know how you got the opinions you hold. What I do know is what I see, and what I see are a bunch of shouted terms and a couple of links to some sites, two of which are owned by the same registrant out of Vancouver Washington Sorry, but I can smell the manure from here on that.

Now, if you want to hold the opinion that REISER4 is worth being in the kernel, here's how you go about getting it in.
Contact your distribution vendor and ask them why REISER4 isn't offered as an option at install.

Did you get that? Here, I'll say it again

Contact YOUR distribution vendor.

Now, since i don't think you understand, I am going to spellthis out real clearly for you. In baby terms.
REISER4 can be added at any time to any distribution, by anybody who feels like it. There are hundreds of packages, drivers, and other tools that are shipped daily that are part of a reconfigured kernel that are not in the main kernel. That's one of the differences that sets distributions apart, what is added to the kernel. REISER4 does not have to be part of the Linux kernel to be used on your machine. Addition to the Linux kernel as a shipping part of the main tree is an indication that the technology has become widespread or is worth using. The fact that several distributions are stripping REISER support out, period, full stop, indicates that there is a general trend away from the REISER format overall. Now, if the opposite was happening, if REISER4 support was being added in or being made the default option on distributions, then yes, it might make sense to pursue kernel inclusion.

For now? it's just best to forget REISER and start working on the problems with the dozens of other file systems supported by Linux, or by working on getting Sun's ZFS ported to the Linux platform.


***********

Wow, you really are an idiot. If you like Reiser4 so much, just compile your own kernel or send an email to your distribution. That website's benchmark also show that Ext4's performance is very similar to Reiser4's, and Ext4 has the added benefit of providing backwards compatability with Ext3/Ext2.

Thank you for repeating what I already said, but I think it is obvious that the person who is responding has no interest in doing their own research into the matter. The thing is, I have had a much better coder than myself assist in adding ReiserFS into a Mepis 6.0 Final build, and into a Fedora Core 6 build. We tested on a couple of different hard-drives in single drive mode: Hitachi Desktar SATA 300 @ 80gig, Samsung SATA 150 @ 250gig, Western Digital 150 SATA @ 120gig, Western Digital 100 ATA @ 120 gig, Maxtor 133 ATA @ 120 gig, Hitachi Desktar SATA 150 @ 160gig. We also tested on a couple of different RAID 0 formats, using a couple of Promise controller on 2 of the Maxtor 133 ATA @ 120gig, using Nforce3 RAID SATA with the Hitchi SATA 150's @ 160gig, using Radeon Xpress RAID SATA on the Hitachi 300's @ 80gig, and getting fancy by sticking the SATA 300 cards on SATA 150 ports, and by running SATA 150 drives on SATA 300 ports.

Know what we found? No difference from ext3. We went from an AMD K6-2 @500mhz and a Pentium3 @ 500mhz, up through Socket 754 Athlon64 @ 3200+, Socket 939 Athlon64 X2 @ 4200+, Socket AM2 Athlon64 @ 4800+, and had a Socket A 2500+ Barton in there somewhere. We had a couple of systems from Intel, the Radeon Xpress 200m holding an Intel P4 630, and at the time brand spanking new Core2 system using an Intel board.

In our testing we couldn't find any repeatable real life situation where the REISER4 release was faster than EXT3. In most common hard-drive scenarios, there was no difference that could be detected.

So, why didn't we publish our results like good little academic researchers? Note the OSs we were using. Mepis 6.0 and Fedora Core 6. yeah, we were wrapping up our research just as Hans Reiser was being hauled off on murder charges. We made the choice to not publish our results because A: REISERFS was going to be as good as dead with Reiser gone, and B: we felt that publishing our results, even using our casual blogs, would be seen as a hack job against the guy. We've seen it before, the claims of people with an Axe to grind. So we junked it.

Looking back now, with the people who simply refuse to give up on REISERFS entry into the mainline kernel, maybe we should have published our results. At this time though, the details would be approaching over a calendar year out of date, and with improvements to the baseline kernel, and other sections of the GNU system, our original results would be meaningless. I'm not really interested then in repeating all of the tests, even on the same hardware, unless somebody pays for it this go-around. Maybe Namesys improved REISERFS so it is actually going to surpass EXT3. Thing is, EXT3 was "good" enough, and EXT4 is better. EXT4 is backwards and forwards compatible. REISERFS isn't. I could continue on into the technical details of why REISERFS is bad. That wasn't good enough argument a year ago, and I doubt it's good enough argument now.
Post a Comment