The Fedora project was urged to be more community oriented and the editor showed an example of how community involvement helps to reduce Fedora boot time. Also, some intro on Sun’s Solaris 10 over at LWN.
LWN on Fedora and Solaris 10
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Hmm I also looked at posting the Solaris article on exp zone… Decided against it, though.
The thing that wonders me is that they say “a less friendly oss license [than the GPL]”. However, let’s not get into that
I’m far more interested and excited about the open-sourcing of Solaris than most people, so it seems. It will be good for both Linux and Solaris– Linux will finally have a true competitor, and Solaris will gain more attention and users, since it will become easier to obtain Solaris. In the end, both will profit from it, I’m sure of it.
I don’t understand why so many people get all pissy against Linux/Solaris about this. Both will profit from this (depending on Solaris’ license, of course).
At the beginning, the author promises to talk about at least some of the new features of Solaris 10. But after some 4 or 5 sentences, the description of new features ended, and a tirade of OSS-and-Solaris topics are flooding.
This guy is sick, and it’s a bad sick.
This is the type of article that gets Linux users in trouble, read a couple of press releases, make a couple of assumptions and write an article about an OS in a less than friendly light.
There is nothing in the article that talks about anything that cannot be read in any other article or press release published. So my guess is this guy has never used Solaris 10. If he had, I somehow think his article would be a lot different.
If he’s quiet than you know somethings getting done, he’s been squeeling since FC1 beta =)
He’s right though, this is something that needed to get done and I hope the relationship between fedora community and redhat is done a bit differently than most open projects. I’d like to see RedHat maintain the ability to be firm on Fedora’s direction yet still listen to the community. Meaning don’t let any peice of code in cause its cool or saves RH dev’s the work of writing it correctly, have very, very strict guidelines like linus does: “that patch is garbage what were you thinking! re-write it and come back” I like that demand quality attitude.
“The thing that wonders me is that they say “a less friendly oss license [than the GPL]”. However, let’s not get into that
”
SUN has said numerous times they will release the Solaris as open source with an Open Source Initiative compliant license, such as the GPL< BSD, Mozilla license and others.
Also why have they not designed fedora to be used for i586 or higher ?? Who use a 386 or 486 for that matter nowadays for a desktop ? Move up with the times Fedora and compile your packages for at least i586 support or i686.
Slow on what system, yours or someone else? Fedora with Gnome will slower on old system like AMD K6 than Fedora using light desktop like XFCE.
“Also why have they not designed fedora to be used for i586 or higher ??”
read the release notes before posting would you?
http://fedora.redhat.com/docs/release-notes/fc3/x86/
Fedora Core 3 is optimized for Pentium 4 CPUs, but also supports earlier CPUs (such as Pentium, Pentium Pro, Pentium II, Pentium III, and including AMD and VIA variants). This approach has been taken because Pentium-class optimizations actually result in reduced performance for non-Pentium-class processors, and Pentium 4 scheduling is sufficiently different (while making up the bulk of today’s processors) to warrant this change.
Yeah, fedora is slow. I experimented with the benchmark script mentioned in the article, with some tweaking I got the boot time on my main system down to 47 seconds (not counting loading gnome). There is still room to reduce that by another second or 2. The system is much more responsive now than on default load though.
I loaded FC3 on the 450MHz test box next to me that was running XP, FC3 is noticeably much slower to start and shut down on that box, it is also much less responsive to use. I am sure with some tweaking I could make it usable but I don’t want to spent the time tweaking it. You click, then wait, type then wait..
I don’t have the patience.
Being OSI complient doesn’t have to mean that much; within the OSI framework you can still make the license quite painful for developers, QPL being a good example with its demand for changes being patches. The fact that Troll Tech, who even had the backing of KDE, still felt compelled to change to the GPL eventually should tell Sun something about how receptive their target audience will be with a new non-GPL compatible cumbersome license.
Same goes for Mozilla, they also tried going with their own license, but in the end relicensed to be more edible to the wider communtiy. And the MPL was actually technically a nice license.
The lackluster interest around Darwin is maybe the best example of what will happen if Solaris comes out under a non-GPL compatible license.
Sun should have some balls and put Solaris out under the GPL or even better the LGPL. Instead of being worried about people porting Solaris code to Linux, they should be more positive minded and think that having a GPL compatible license will enable porting of Linux code to Solaris much easier. Like drivers, which is truly is Solaris’s Achilles heel.
The LWN piece has, in common with far too many similar pieces published in the Linux media, a broad disregard for facts and support.
It isn’t enough to make a broad assertion without substantiating it with facts and evidence. This anonymous piece begins by claiming that “The Fedora Project, after more than one year, has not become a “community” project by any means. It is centrally controlled, and many crucial decisions seem to come from some sort of smoke-filled room in Raleigh.”
What’s missing here? For starters, a definition of “community”. Since the author is levying a criticism at Fedora, he (or she) has the obligation to explain why Fedora’s alleged behavior is bad and how he defines “community”. The author is obviously relying on his expectations that his readers will share his view of “community” and Fedora’s alleged behavior. But, building your case by relying on the unchallenged assumptions and biases of your readers is polemical rather than convincing. In other words, throwing a bone to the already converted.
As for the smoke-filled room, does anyone really doubt that Red Hat facilities are smokefree? The author provides no evidence to support this blantant attempt to portray Fedora and Red Hat managers in a sterotypically negative light.
The author makes no attempt to explain why Fedora should enlarge community involvement. What constitutes sufficient involvement? The author mentions the Debian project. But, why would RedHat relinquish control of an effort it created for its own purposes and which it continues to support and finance to a scattered network of unseen developers who may or may not share RedHat’s objectives for Fedora?
“Community” control has not made Debian markedly superior in quality or innovation to Fedora or any other mainstream Linux distribution. Given the levelling impact of open source and the GPL, no single distribution is ever going to take the lead for any longer than it takes for the competition to incorporate their code.
This is the type of article that gets Linux users in trouble, read a couple of press releases, make a couple of assumptions and write an article about an OS in a less than friendly light.
It’s a pretty decent summing up of the difficulties that Sun and Solaris face, and it’s typical that people who accuse people of Linux zealotry cannot read an article and see what’s in front of them.
Sun are simply going to need an effective license when open sourcing Solaris, otherwise it will not polarise developers. That much is clear. Look at Fedora – it uses genuine, GPL’d and open source software but the processes that underpin it have just not been worked out. Sun has a much more difficult task, and they’ve talked about open sourcing Solaris but not a jot about how it will be done.
He’s also correct about companies who’ve invested time and money in Linux – they’re just not going to move back. There’s no reason to, and many are being misguided that throwing technical features like DTrace at it is going to make a difference. It isn’t, however good it happens to be.
The fact that Sun are also giving away Solaris and many x86 servers as well begs the question of whether their strategy is sustainable and how on Earth they’re going to make money. If they’re giving away Solaris, why bother with Solaris? I’ve heard about all this value added stuff people seem to think Sun will supposedly sell, but its going to have to be bloody expensive to keep the revenues they’ve got now and to stop them sliding. People will pay decent money for a server, an operating system and support but they’re not going to stomach being given servers and Solaris and then being hit for all sorts of services on the top. Rather than re-modelling their current revenue stream, Sun have effectively ditched it and created an imaginary new one. That’s a very dangerous response for any company to be making to new threats and competition.
Maybe Sun thinks that if they give away Solaris and cheap servers for long enough the Linux problem will go away and they can go back to the way things used to be. That’s a bad idea.
So my guess is this guy has never used Solaris 10. If he had, I somehow think his article would be a lot different.
Why do people from Sun, and around Sun, think that its all about Solaris 10 and using Solaris 10? The author is right in that open sourcing Solaris and regaining any customers lost to Linux is not going to be about the technical features of Solaris 10. Sun’s situation and the issue of opening Solaris has absolutely nothing to do with how good its technical features happen to be. Sun and some people around here will wish that was the case, but it isn’t.
It’s time for Sun to wake up, smell the flowers if necessary and look and think about what they’re going to do with Solaris and what people are going to pay sensible money for. Solaris is still an asset, so I don’t see the worth in them giving it away (in a support sense as well) across the board. That may mean accepting a drop in revenues to achieve long-term stability and survival, but Sun are not going to realise that let alone do it.
That’s just not Linux zealotry, that’s the reality of their predicament.
SUN wants to sell support for Solaris, they want to adopt a similar business model to Redhat’s. That’s how they will still make money from Solaris.
SUN wants to sell support for Solaris, they want to adopt a similar business model to Redhat’s. That’s how they will still make money from Solaris.
I’m talking about the fine detail as to how they will do that, not the stupid general statements we’ve seen. “Oh, Sun will just make money from value added services…”. Doesn’t work like that in practice I’m afraid.
Given that Fedora uses fully GPL’d and open source software and they still don’t have their community processes worked out, and Sun also have license issues to solve and a whole community to create around it?
That’s how they will still make money from Solaris.
Sun aren’t going to make money from Solaris – at least not directly. As I’ve already described, the revenue stream people seem to think Sun will have on top of Solaris is imaginary. It currently doesn’t exist. They’re assuming the money they make from Java and other areas of the business will somehow be boosted to cover the expense of giving things like Solaris away. There is no reason at all to believe that will happen because that’s what they’re doing anyway.
You should take into consideration that Sun is still trying to sustain an expensive dieing chip architecture, in light of less expensive, just as well performing, and more sensible alternatives.
Redhat does not need to worry about maintaining its own hardware business. And when you think about it, they don’t really maintain alot of software.
The slowness I am talking about has to do with the GUI and redraw speed and overall snappiness of the GUI. I had Fedora Core 3 running on a AMD 3200 XP PC with a Geforce 5200 FX card with 256 megs of video ram and Nvidia’s latest drivers installed. The redraw rates of the gui and over all feel felt sluggish when compared to other distros running KDE or Windows XP. Now maybe it might have something to do with Fedora Core 3 itself or Gnome or both but it was slower then I expected and have experince under other distros running KDE 3.2 or 3.3.
hehehe…This because they don’t have a hardware biz and you are correct about the software side.
I can’t even get the FC3 installer to load on my PC when booting from the DVD image, so I couldn’t say how much faster it is 😐
Okay, enough is enough. David, you wrote:
He’s also correct about companies who’ve invested time and money in Linux – they’re just not going to move back. There’s no reason to, and many are being misguided that throwing technical features like DTrace at it is going to make a difference. It isn’t, however good it happens to be.
This reflects a serious misunderstanding about information technology. Do you honestly think that Linux is so entrenched that no technology can displace it? If so, let me be the first to break the news to you: users of information technology are in a business. Their objective is to be profitable, not to deploy trendy technology. If the trendy technology happens to best help the business be profitable, then fine — but don’t be fooled into thinking that running the trendy technology has somehow become the objective.
To wit: I have spoken to many, many companies over the past year — usually to very high ranking technologists. Many of these shops have adopted Linux in significant quantity, and some have had the objective of adopting it exclusively — so these weren’t Solaris-friendly audiences. But once they saw a demo of the technologies we have in Solaris 10, without fail their plans surrounding Solaris changed. Universally, they decided to aggressively explore Solaris 10 by running Solaris Express. Often, they put further Linux migration plans on hold until Solaris 10 had been fully evaluated. And in more than one case, Linux migrations have been stopped dead in their tracks and replaced with Solaris x86 and/or Solaris Opteron migrations.
In a few cases, customers have been so enthusiastic about Solaris 10 that they have been willing to go on the record about how the technologies in Solaris 10 have changed their future plans. These are not small or trivial companies — and they include among their number a customer that I hold in absolutely highest regard: FedEx. So, David, this brings me to the existence proof that your assertion is incorrect:
http://www.computerworld.com/hardwaretopics/hardware/server/story/0…
I know you won’t bother to click on the link, so here’s the key quote:
Prior to the development of Solaris 10, interest in Sun’s systems “was somewhat dwindling” at FedEx, said Don Fike, the Memphis-based company’s technical director. But now that FedEx’s IT staffers have evaluated the new software, “interest is extremely high—a major shift here,” he said.
I think that pretty much speaks for itself. I’ll only add that that the “major shift” isn’t only at FedEx…
Interesting how people have the screaming match of BSD License vs GPL, whose to say that SUN might simply go for the same license Apple uses for their Darwin code (APSL)? IIRC, that is OSI compliant, and allows a good mixture of multiple licenses in the one code base.
Lets stop this “fuss’in and a feuding” and start looking at the positive aspects of Solaris and it being opensourced. If Solaris IS opensourced, then there will be another competitor out there for Microsoft to take fire against. An truely enterprise class operating system running on commodity hardware with a large company standing behind it.
He’s also correct about companies who’ve invested time and money in Linux – they’re just not going to move back. There’s no reason to, and many are being misguided that throwing technical features like DTrace at it is going to make a difference. It isn’t, however good it happens to be.
Bullcrap. That is the same thing said back when Linux momentum was starting; the nay sayers said, “oh, people will never move, they’ve already invested so much money with Windows, why would they throw it out?”.
Get with the programme, companies will move if they see a definate benefit; Solaris can run all the opensource applications that can run on Linux, Solaris has all the commercial server software that Linux has.
The issue will be cost of support, reliability and other measurable results; not the “feel good factor” that geeks feel knowing they run an open source system.
Adam (IP: —.clv.wideopenwest.com) – Posted on 2004-11-29 01:10:04
You should take into consideration that Sun is still trying to sustain an expensive dieing chip architecture, in light of less expensive, just as well performing, and more sensible alternatives.
Oh, isn’t it interesting that for a dying architecture, the number of systems sold by SUN (SPARC) have increased each year; yes, thats right, the revenue may have decreased, their prices have decreased but the number of systems have increased.
Oh, and if you take R&D out of the picture, SUN made a nice $300million profit last quarter. Look at the facts, not the emotion. If SUN was dying, it would have died ALONG time ago.
Maybe it’s just me…fedora core 3 on my laptop has a boot time under 30 seconds and I’m very happy with the performance. Of course, I’ve disabled all the “extra” services (hardware detection – it’s a laptop; ifup on anything – again…laptop; no graphical boot loader; lilo with noprompt; starts in runlevel 3; etc).
I am interested to see where Solaris goes in the near-future though. I’d love to see what the community does with it (or doesn’t do with it).
You’d need to eliminate the Nvidia driver as a factor before placing all the blame on Fedora or Gnome.
I’ve used the current Nvidia driver with a variety of distributions and XP, all on the same hardware. It’s my impression that redraws and such were faster on XP than on any Linux distribution. None of the Linux setups were annoyingly slow, but XP was definitely snappier.
It’s a pretty decent summing up of the difficulties that Sun and Solaris face, and it’s typical that people who accuse people of Linux zealotry cannot read an article and see what’s in front of them.
That article is utterly useless. The very few words describing the merits of the OS itself (about 10% of the whole “article”) are partially wrong, too! There is no freaking way an intelligent person is going to defend that junk.
As for the whole licensing issue and how company X makes/will make money, I will not discuss that now. But the article is still crop of crappola.
If I were Sun i would just open up a safe version of the operating system. By safe i mean releasing perhaps a little older version, or atleast a version that does not include the latest and greatest solaris features and features that may have lisence repurcutions. Open it up and market it as the way Red Hat markets Fedora with RHEL being comparable to Solaris 10. Much like how the Netscape – Mozilla partnership started. Allowing a personal Solaris version with desktop drivers so common users/programmers could use it and if they want improve on it. Soon who knows it may even become as good as their Solaris 10. But thats just me.
Don’t blame Nvidia. I’ve used FC3 with the stock GPL driver and Nvidia’s driver and it’s slow either way !
I’m trying to figure out how Sun could maintain any credibility by following your suggestion of releasing an older version of Solaris for the open source base; and I simply can’t see it.
What would be the point in anyone participating or innovating?
As we have said previously, what is being opened will be the current source tree.
The new stuff that you think we might have licensing problems with is actually the stuff that I would have thought obvious that there is clear title (e.g. Dtrace, Fire-engine, ZFS, Zones, …).
The idea is that we truly wanty to move to an open sourced model, difficult as this appears to be for many to accept.
Alan.
I’m not blaming Nvidia. I simply suspect that their driver performs differently with different distributions and that the Windows driver performs best of all.
Everytime a new Fedora release comes along, it seems to be instantly bashed by the “mine is better than your’s” crowd.” On my hardware (Athlon XP 2.7, 1-gig ram, Nvidia 5900XT, 17-inch flat panel) it is at least as snappy as any other Linux flavor I’ve used, and considerably more polished than most. By default, it does run a number of services unnecessary for desktop use, And, logically, any software running on older and slower hardware will take a performance hit.
I had Fedora Core 3 running on a AMD 3200 XP PC with a Geforce 5200 FX card with 256 megs of video ram and Nvidia’s latest drivers installed. The redraw rates of the gui and over all feel felt sluggish when compared to other distros running KDE or Windows XP. Now maybe it might have something to do with Fedora Core 3 itself or Gnome or both but it was slower then I expected and have experince under other distros running KDE 3.2 or 3.3.
I have a system with AMD Athlon XP 2500+ Barton, Geforce FX 5600XT with 128 Meg of VRAM with the latest driver, 512 MB of total RAM. Obviously you didn’t talk about the amount of physical memory RAM you have on your system. I haven’t faced the problem you mentionned under Gnome even though my system is inferior to yours. Try to ask Ubuntu users if they face the same issue.
Interesting that you mentionned KDE 3.2 or 3.3 as Fedora Core 3 has it included so you should compare both on the same desktop environment.
”
The idea is that we truly wanty to move to an open sourced model, difficult as this appears to be for many to accept. ”
you know why thats difficult to accept?
Because the Jonathan guy LIES about redhat being proprietary and not LSB compliant and other such utterly stupid statements. you guys dont even consider Novell as a competition while repeatedly attacking Redhat.
If you guys had handled the situation anything close to gnome involvement it would have been great
Sun – technology is good
sun marketing and CEO people are liers which is basically the failure of it
hehehe…This because they don’t have a hardware biz and you are correct about the software side.
They want to effectively give away the x86 hardware as well and make their money from software and services apparently, so their software and hardware strategies are very linked. It looks like an effort to stem the tide of the x86/Linux combination. I hope they have a lot of patience and deep pockets.
But once they saw a demo of the technologies we have in Solaris 10, without fail their plans surrounding Solaris changed.
No they didn’t. Wherever it is that you live it certainly isn’t Earth, but being from Sun I guessed that already. You can say it a thousand times and put it into as much bold text as you like to try to convince people – technology by itself is not a reason for anyone to move if the existing technology is relatively cheap and is doing what people want it to do to start with.
The reason why Linux has succeeded where even Windows has failed as a commodity OS on x86 is that the distributed development model produces not just excellent (and good enough) server technology, but it is just an excellent way for everyone to keep support and development costs down and not put massive effort into keeping their own proprietary systems going. The technology is undoubtedly good enough overall, but you can’t point to single features like DTrace and say “That’s a reason to ditch everything.” The real world just doesn’t work like that.
The link you gave reads exactly like one of those Microsoft articles where they give away software and consultancy services to make their products look good, and that’s exactly what this is. None of what they’ve said makes any sense for any business to move to unless they’ve been handed stuff on a plate to stop any migration they may be making – which is exactly what has happened I’ll bet.
The acid test is has Linux etc. been cheaper for a company and does it work at least as good as, if not better, than what we could have bought? The answer is yes and yes, and that’s all that matters.
And in more than one case, Linux migrations have been stopped dead in their tracks and replaced with Solaris x86 and/or Solaris Opteron migrations.
Wish, wish, wish. Even if that is the case, I hope you have patience and deep pockets to keep it going because that’s all that’s propping it up. You’ve just proved me right that this is what Sun is trying. I’ve offered ample reasoning as to why this is just not viable in the long-term.
“Oh, oh, let’s give everything away, try and stem the tide and we’ll close our eyes and just hope that this big bad Linux thing goes away.” I know how the old-guard at Sun thinks, and they’re not unlike Marks and Spencer in the UK at the moment. Linux together with cheap x86 hardware has a workable business model right into the long-term because no company needs to offer any customer anything special.
That is the same thing said back when Linux momentum was starting; the nay sayers said, “oh, people will never move, they’ve already invested so much money with Windows, why would they throw it out?”.
Because it was a cheap OS environment, and cheap to maintain for companies like Red Hat, surrounded by cheap hardware that people realised they could use instead of the expensive stuff they were buying. That advantage in the market has been and gone, and it isn’t with Solaris.
And yet this moron keeps regurgitating the same reply again.
Solaris + x86 hardware is cheaper than Red Hat + x86 server. Don’t want to pay for support, easy, don’t pay for it, just rely on the free security updates SUN will make available for their free version.
Stop lying; acccept the facts; Solaris is back, its on the x86 server and its dirt cheap.
Stop trying to dig up red herrings to justify your dogma; put you ego aside for a moment and realise the positive aspects of it rather than whinging and whinging.
[T]echnology by itself is not a reason for anyone to move if the existing technology is relatively cheap and is doing what people want it to do to start with.
[Y]ou can’t point to single features like DTrace and say “That’s a reason to ditch everything.” The real world just doesn’t work like that.
Bryan pointed out that you fundamentally misunderstand how people use technology in their business, and provided a set of examples that you should consider. Yet your reply shows, more than ever, that you’re still completely missing the point.
Technology, in a vacuum, isn’t necessarily a resason for a company to abandon its deployment and switch. However, you’ve failed to realized that these features don’t exist in a vacuum. Are you going to argue that a company would opt to choose a set of technologies which make it harder for them to get their work accomplished, and requires them to use more resources to do so? You’re arguing against innovation, and against a customer’s good sense for picking a better deal. If I the customer find that using a new technology I can diagnose incredibly complex problems in a matter of minutes instead of weeks, I’ve saved money if I use that technology. I can spend my time doing other things that are directly related to my core business functions. If I find that I can use one server to do things that I previously needed 5 to do, why would I go out and continue to buy 5 servers? It’s a waste of resources that can be better applied elsewhere. You fundamentally misunderstand how the features in Solaris 10 can help customers save time and resources in ways that Linux cannot.
Customers don’t want DTrace because its trendy. They want DTrace because it allows them to save money. Are you suggesting that customers should continue to waste time/money/resources when they can be conserved by employing a new technology?
Linux together with cheap x86 hardware has a workable business model right into the long-term because no company needs to offer any customer anything special.
You’re suggesting that no company should bother to innovate because Linux has done all of the innovation. This is as rediculous as it is false. Are you Amish? What have you got against technological progress and innovation? Customers will always be willing to consider new solutions that let them be more productive and do more with less.
The system has 1 gig of ram so no it’s not an issue with not having enough ram.
P.S. KDE on Fedora runs faster now that I tried it and I know for sure it’s a Gnome issue know. The thing is thought that KDE on Fedora is not as well intergrated as with Gnome. It’s obvious that Gnome has gotten more of the attention when it comes to polishing and making it intergrated with apps.
I have experience of Ubuntu, Mandrake 10, Fedora Core 2 and Suse 9 on two aged machines: a Athlon 1700+ nforce with 256 ram and a very elderly P3-500 with 192ram.
FWIW I found FC2 to be slower than Mandrake or SuSe – in fact it was frustratingly slow on the older machine where Mandrake was quite usable. I recognise that this may be because I did not optimise the Fedora install as suggested here.
However, Ubuntu is a whole different ball game. It flies on the P3 – in fact general desktop apps (although obviously not processor heavy stuff like games) appear faster under Gnome on the P3 than they do under IceWM under Mandrake 10 on the Athlon!!!!!!!!!!! If you haven’t tried it, do so. It’s opened my eyes the debian way for the first time…..