I dunno, I don't see too many solaris based web servers on Netcraft's list of longest uptimes. They're all running FreeBSD
As I pointed out in another post, the FreeBSD boxes are probably just a bunch of basic machines serving the actual webpages on a round robin basis. The heavyweight work (data stores, user accounts, news feeds or whatever) will be done by machines that NetCraft doesn't see - and they are often big iron like Sun or IBM servers.
How many servers are (or how many admins actually want) >32 processors anyway? Isn't it much easier to cluster your servers together?
I see this come up quite often on places like Slashdot, the old "why buy big iron when you can cluster?" argument. The reason people pay for big iron (be it from Sun, SGI, IBM or whoever) is because writing a multithreaded app that runs on one machine is far simpler than writing a distributed one. If it's well written, then it will outperform a distributed app unless you use a disproportionate amount of clustered hardware.
For many applications, a bunch of separate machines can be a good solution. Think of a server farm for a large website, where each machine handles requests on a round-robin basis. However, the moment they need access to a single resource (such as a database of user accounts), then you need a single heavyweight machine to handle this resource. Otherwise you're introducing a potential bottleneck, have to write a very elaborate distributed system or invest in some serious enterprise software (which probably runs on a large Sun box anyway).
And keep in mind, this is the company that told us hardware should be free, and people should only pay for software.
I think you misunderstand what Sun were trying to say. What they meant was that you buy a lease, which essentially means that the hardware is free, as you're actually paying for support. Of course, if you stop paying for support, then they take the hardware back...
The original SunOS was BSD based, but Solaris was built on top of SVR4. Extra features from the old BSD based SunOS were shoe-horned into Solaris, and various bits put into a compatability directory (ucb).
in all the coverage I've heard on this the only mention made is of the east coast of the US. Just one more invisible catastrophe to hit the third world...
If you look at the diagrams, the vast majority of the energy released by the slide would travel westwards. The coast of Africa looks like it would be far less at risk in this scenario than the Eastern seaboard of the US. It's the Caribbean islands that are overlooked in a lot of the press coverage, as they are most at risk.
I think that sun should come out with a new open version of Solaris that is fully compatible with the current version but integrates some flavor of BSD.
Solaris does integrate some flavour of BSD - SunOS. The predecessor to Solaris was BSD based, as Suns founders were ex-Berkeley people who had worked on Unix. When Sun bought a license for SVR4 from USL (or it may have still been AT&T at the time), they integrated much of the extra bits that their BSD version of Unix had.
Now if they open sourced it, that would be great, but I guess that really depends on the fine detail of their SVR4 license.
While BSD License really is the most free, it would allow IBM to put a lot of effort into it, and then have MS swope in, modify it, and sell with a sorts of closed APIs, etc.
No they wouldn't, because under the BSD license, IBM wouldn't have to publicly distribute their modifications to the NetBSD code.
The real reason IBM go with Linux? They have more expertise with it than any of the BSD's.
The slow pace of recent releases is down to the amount of work going into NetBSD 2.0. This will feature decent SMP support and high performance threading using Scheduler Activations. The improvements in performance compared to the 1.6 branch are extraordinary, and unlike Linux, where reecnt work has been aimed at improved performance on high end hardware, the NetBSD improvements are generally applicable to all classes of machines.
It's actually quite amusing to see NetBSD development labelled as "slow" compared to OpenBSD. The reverse is very much true, as Open takes from Net far more than Net takes in return. For example, the rapid implementation of SMP support in Open was due to Net having done most of the work already.
The NetBSD developers are also not very vocal about what they're working on. In the Linux world, we have things like Kernel Traffic picking over the minutiae of mailing list activity. There's also the massive preponderance of "Linux branded" press coverage, which gives an emphasis to goings on that are fairly peripheral to the Linux kernel (GNOME and KDE for example). The only real way to get a feel for what's going on in the NetBSD world is to study the CVS logs - not something that many people aside from the developers would do.
If you're finding that the GENERIC kernel is too bloated, then the quick fix is to create a custom kernel usung the adjustkernel script. This parses the output of dmesg, and creates a custom kernel config file with only the devices found on your machine enabled.
On my laptop, I was able to pare the kernel down to 1.8Mb. Not such a big deal on a machine with 512Mb of RAM, but it's useful on something like my Vax which only has 24Mb.
Contrary to what I posted earlier there apparentely is already one hell of a tool that creates svg-graphics, namely sodipody.
Even better than Sodipodi is Inkscape. It's a fork of Sodipodi, ported to C++ and more actively developed. It's well worth checking out the screenshots page. Sodipodi is pretty much a one man effort, and that man (Lauris Kaplinski) doesn't get much time to work on it. Inkscape is developed in a more collabarative manner, and has made impressive advancements during its short existence.
It lowers the cost of customer support. Compare
Sir, did you install the Flash plugin? No? Well, you need to install it.
to
Sir, did you install the SVG plugin? what? yes, it's Ess-Vee-Gee... Yes, Sarah-Vostok-Gargoyle... No? well you need to install it.
The articles are talking about mobile phones. The end user is not the one who is going to be installing a Flash or SVG 'plugin' on such devices.
I thought the fundamental design goal of the web was fail-safe information exchange?
No, that was the design goal of the Internet, which the World Wide Web sits on top of. The goal of the Web was to provide easy distribution of scientific data and analysis (Google for Tim Berners-Lee, CERN and the birth of the Web). Flash fundamentally breaks that idea by limiting the viewing of content to the couple of platforms that Macromedia deign worthy of a Flash plugin. Ever tried viewing Flash content from a SparcStation running NetBSD?
Normal "artists" don't ask people to start messing with the clay, the artist make it and the people have a right to watch or even buy it.
Bad analogy. The Web was designed for platform independent dissemination of information. Now, for some people that 'information' may be considered art, but most websites are not supposed to be 'pure art'. For instance, a band's website may be an extension of their sleeve art, image, etc. But it usually contains normal content, like gig listings or release info. That information shouldn't be buried in a Flash movie.
Yeah, browsers were intentionally used to display html only, but things change... These days browsers are used to display multimedia...
A Flash movie may be ideal for the multimedia, but when content is buried in it then it defeats the idea of the Web.
webmasters want to make a few bucks off off advertising
If Flash was only used for adverts, then I wouldn't mind. I'd just never install the Flash plugin. Now I have to use Flashblock on the one machine I own that can run the plugin, otherwise I'm bombarded with *very* distracting adverts (the LinuxToday website springs to mind).
I was thinking more along the lines of one of the modern sequencers
I borrowed my brothers Akai MPC a while ago, and it was a very nice piece of equipment. I couldn't justify the cost of one for myself though, as I'd never use half the features. I'm also hopeless at sampling, so I'd end up using third party libraries.
This is where Apple has traditionally worked, and it's been a strong position for them.
This may just be my experience, but I've found that in the UK studios are more likely to have PC's than a Mac. however, in the States the opposite appears to be true. I haven't seen an Atari ST in a professional studio since 1996 though...
If you do that, then why do you need the kind of timing ST owners blather on about? There are outboard sequencers that would do you just fine.
I don't need anything out of the ordinary when it comes to latency. The main reason I use the ST instead of a hardware sequencer, is the better interaction via mouse, monitor and keyboard. Saying that, I do have an ancient Roland MC-300 sequencer that I use for live stuff (it's a rugged little beastie).
Some of us aren't totally enamoured by software synths and samplers. I'd rather use hardware devices (analogue preferably), and just use the computer as a sequencer and patch editor/librarian.
I still use Steinberg Pro-24 (the forerunner to Cubase) on an Atari ST. It is far more stable than any version of Cubase that I have used on a PC. I don't own a Mac, but I do hear that recent versions of Cubase have been rock solid on that operating system. So if I upgraded that would be they way I'd go. Then again by the time I can afford a decent Mac, Rosegarden (http://www.rosegardenmusic.com/) may have reached a stable release...
As for the poster below who talks about booting from a 720Kb floppy, my ST has a SCSI hard drive, and boots in seconds.
Unless the post-war Iraq Democracy caused Peace In the Middle East.
I wish it would, but like you I don't think it will. In Iraq itself, the new rulers have very little credibility with the ordinary people. During their time out of the country, they argued amongst themselves, and whatever grassroots support they had became disillusioned or disappeared into Saddam's prisons. Add to this the religious and tribal based allegiance that a lot of Iraqis have, and you have a recipe for weak, ineffectual government by coalitions.
No, the Russians do have the most experienced and reliable rocket engineers on earth. That's why NASA are working with them. Following the collapse of the Soviet Union, a number of Russian rocket engineers carried on perfecting the designs for extremely stable and powerfull rocket motors. This work has now been commercialised, and is used in both the Russian and American space industries. I'm sure a quick Google will turn up some suitable references.
I believe that in 20 years, we'll look back and say the same for W.
The difference between Reagan and Bush junior, is that Reagan's anti-Soviet rhetoric was mostly for public consumption. Behind the scenes, there was a great deal of diplomacy going on which ultimately lead to the arms limitation treaties. The Reagan and Bush senior regimes were much more pragmatic than the Rumsfeld / Cheney / Bush junior regime. We'll look back on the Bush junior regime in 20 years time with as much disgust as most people look at it now.
The belligerent attitude of the current regime comes as no surprise to those of us who kept up with what the various neo-con think tanks that influenced the current regime were saying in the mid-1990's. Cheif amongst their suggestions was that Saddam Hussein should be given a whipping for going against the wishes of the last Republican regime. Saddam had been the pet Middle-East strongman of the US throughout the 1980's, but he overstepped the mark by invading Kuwait. Having glossed over his previous gassing of Kurds, the Bush senior regime was thrown into turmoil by the Kuwait invasion. This is why there was a lack of firm comment on the situation from the Whitehouse in the immediate aftermath.
Slighly off topic, but the origin of the word "man" was gender neutral. Some languages still make little disinction between gender (Finnish for instance).
Most programmers I've worked with in the UK have either read "Mythical Man Month" or at the very least heard of it. The same goes for Jon Bentleys "Programming Pearls".
Both books were a little bit of an anti-climax when I first read them, probably because I expected way too much in the way of blinding insights. I found I was like the bloke that Brooks sat next to on a plane journey (described in the second edition) - so much of what the book has to say seems obvious now.
However obvious those insights may seem, big projects still get bogged down with the same old problems. I guess that means managing really big projects is still a bit too much for most of us to cope with.
I had high hopes that Apple would contribute back to the community, but I don't think that has materialized like I had hoped.
Mac OS X uses the Mach kernel with a FreeBSD layer above it. This means that much of Apples work on the Mach kernel is irrelevant to FreeBSD. Mach is a microkernel, which was of course derived from BSD Unix, but it was forked so long ago that few similarities remain.
As far as stability and consistancey goes, only Debian-Stable approaches BSD
The BSD's also benefit from being a complete system, not a kernel with various userland stuff slapped together into 1001 distributions. This means that users running the development versions are using the same userland as the developers, and bugs can be shaken out far quicker.
NetBSD nor OpenBSD have comparably-sized development teams [compared to Linux]
That's not as accurate as you might think. If you look at Linux kernel development, there is a core of really talented developers who work on the guts - the VM system, VFS, threading, etc. Most other people only poke away at the edges, driver writing for instance. If you compare the number of core developers working on Linux and the NetBSD kernel, a similar number of names crop up with regularity on the relevant mailing lists. NetBSD also has a strong following in academia, and gets a lot of input from talented people with plenty of time to tinker. This partly compensates for the lower number of people working professionally on NetBSD (mostly Wasabi employees) compared to Linux. The BSD's also started from an almost complete Unix implementation (4.4BSD Lite), while Linux had to reinvent the wheel a few times.
Slashdot Mirror
I dunno, I don't see too many solaris based web servers on Netcraft's list of longest uptimes. They're all running FreeBSD
As I pointed out in another post, the FreeBSD boxes are probably just a bunch of basic machines serving the actual webpages on a round robin basis. The heavyweight work (data stores, user accounts, news feeds or whatever) will be done by machines that NetCraft doesn't see - and they are often big iron like Sun or IBM servers.
How many servers are (or how many admins actually want) >32 processors anyway? Isn't it much easier to cluster your servers together?
I see this come up quite often on places like Slashdot, the old "why buy big iron when you can cluster?" argument. The reason people pay for big iron (be it from Sun, SGI, IBM or whoever) is because writing a multithreaded app that runs on one machine is far simpler than writing a distributed one. If it's well written, then it will outperform a distributed app unless you use a disproportionate amount of clustered hardware.
For many applications, a bunch of separate machines can be a good solution. Think of a server farm for a large website, where each machine handles requests on a round-robin basis. However, the moment they need access to a single resource (such as a database of user accounts), then you need a single heavyweight machine to handle this resource. Otherwise you're introducing a potential bottleneck, have to write a very elaborate distributed system or invest in some serious enterprise software (which probably runs on a large Sun box anyway).
And keep in mind, this is the company that told us hardware should be free, and people should only pay for software.
I think you misunderstand what Sun were trying to say. What they meant was that you buy a lease, which essentially means that the hardware is free, as you're actually paying for support. Of course, if you stop paying for support, then they take the hardware back ...
BSD was the bases that Solaris used
The original SunOS was BSD based, but Solaris was built on top of SVR4. Extra features from the old BSD based SunOS were shoe-horned into Solaris, and various bits put into a compatability directory (ucb).
These are the models you seek.
in all the coverage I've heard on this the only mention made is of the east coast of the US. Just one more invisible catastrophe to hit the third world...
If you look at the diagrams, the vast majority of the energy released by the slide would travel westwards. The coast of Africa looks like it would be far less at risk in this scenario than the Eastern seaboard of the US. It's the Caribbean islands that are overlooked in a lot of the press coverage, as they are most at risk.
I think that sun should come out with a new open version of Solaris that is fully compatible with the current version but integrates some flavor of BSD.
Solaris does integrate some flavour of BSD - SunOS. The predecessor to Solaris was BSD based, as Suns founders were ex-Berkeley people who had worked on Unix. When Sun bought a license for SVR4 from USL (or it may have still been AT&T at the time), they integrated much of the extra bits that their BSD version of Unix had.
Now if they open sourced it, that would be great, but I guess that really depends on the fine detail of their SVR4 license.
While BSD License really is the most free, it would allow IBM to put a lot of effort into it, and then have MS swope in, modify it, and sell with a sorts of closed APIs, etc.
No they wouldn't, because under the BSD license, IBM wouldn't have to publicly distribute their modifications to the NetBSD code.
The real reason IBM go with Linux? They have more expertise with it than any of the BSD's.
The slow pace of recent releases is down to the amount of work going into NetBSD 2.0. This will feature decent SMP support and high performance threading using Scheduler Activations. The improvements in performance compared to the 1.6 branch are extraordinary, and unlike Linux, where reecnt work has been aimed at improved performance on high end hardware, the NetBSD improvements are generally applicable to all classes of machines.
It's actually quite amusing to see NetBSD development labelled as "slow" compared to OpenBSD. The reverse is very much true, as Open takes from Net far more than Net takes in return. For example, the rapid implementation of SMP support in Open was due to Net having done most of the work already.
The NetBSD developers are also not very vocal about what they're working on. In the Linux world, we have things like Kernel Traffic picking over the minutiae of mailing list activity. There's also the massive preponderance of "Linux branded" press coverage, which gives an emphasis to goings on that are fairly peripheral to the Linux kernel (GNOME and KDE for example). The only real way to get a feel for what's going on in the NetBSD world is to study the CVS logs - not something that many people aside from the developers would do.
If you're finding that the GENERIC kernel is too bloated, then the quick fix is to create a custom kernel usung the adjustkernel script. This parses the output of dmesg, and creates a custom kernel config file with only the devices found on your machine enabled.
On my laptop, I was able to pare the kernel down to 1.8Mb. Not such a big deal on a machine with 512Mb of RAM, but it's useful on something like my Vax which only has 24Mb.
Contrary to what I posted earlier there apparentely is already one hell of a tool that creates svg-graphics, namely sodipody.
Even better than Sodipodi is Inkscape. It's a fork of Sodipodi, ported to C++ and more actively developed. It's well worth checking out the screenshots page. Sodipodi is pretty much a one man effort, and that man (Lauris Kaplinski) doesn't get much time to work on it. Inkscape is developed in a more collabarative manner, and has made impressive advancements during its short existence.
It lowers the cost of customer support. Compare
Sir, did you install the Flash plugin? No? Well, you need to install it.
to
Sir, did you install the SVG plugin? what? yes, it's Ess-Vee-Gee... Yes, Sarah-Vostok-Gargoyle... No? well you need to install it.
The articles are talking about mobile phones. The end user is not the one who is going to be installing a Flash or SVG 'plugin' on such devices.
I thought the fundamental design goal of the web was fail-safe information exchange?
No, that was the design goal of the Internet, which the World Wide Web sits on top of. The goal of the Web was to provide easy distribution of scientific data and analysis (Google for Tim Berners-Lee, CERN and the birth of the Web). Flash fundamentally breaks that idea by limiting the viewing of content to the couple of platforms that Macromedia deign worthy of a Flash plugin. Ever tried viewing Flash content from a SparcStation running NetBSD?
Normal "artists" don't ask people to start messing with the clay, the artist make it and the people have a right to watch or even buy it.
Bad analogy. The Web was designed for platform independent dissemination of information. Now, for some people that 'information' may be considered art, but most websites are not supposed to be 'pure art'. For instance, a band's website may be an extension of their sleeve art, image, etc. But it usually contains normal content, like gig listings or release info. That information shouldn't be buried in a Flash movie.
Yeah, browsers were intentionally used to display html only, but things change ... These days browsers are used to display multimedia ...
A Flash movie may be ideal for the multimedia, but when content is buried in it then it defeats the idea of the Web.
webmasters want to make a few bucks off off advertising
If Flash was only used for adverts, then I wouldn't mind. I'd just never install the Flash plugin. Now I have to use Flashblock on the one machine I own that can run the plugin, otherwise I'm bombarded with *very* distracting adverts (the LinuxToday website springs to mind).
I was thinking more along the lines of one of the modern sequencers
I borrowed my brothers Akai MPC a while ago, and it was a very nice piece of equipment. I couldn't justify the cost of one for myself though, as I'd never use half the features. I'm also hopeless at sampling, so I'd end up using third party libraries.
This is where Apple has traditionally worked, and it's been a strong position for them.
This may just be my experience, but I've found that in the UK studios are more likely to have PC's than a Mac. however, in the States the opposite appears to be true. I haven't seen an Atari ST in a professional studio since 1996 though ...
If you do that, then why do you need the kind of timing ST owners blather on about? There are outboard sequencers that would do you just fine.
I don't need anything out of the ordinary when it comes to latency. The main reason I use the ST instead of a hardware sequencer, is the better interaction via mouse, monitor and keyboard. Saying that, I do have an ancient Roland MC-300 sequencer that I use for live stuff (it's a rugged little beastie).
limited to 100k for a sample
Some of us aren't totally enamoured by software synths and samplers. I'd rather use hardware devices (analogue preferably), and just use the computer as a sequencer and patch editor/librarian.
What's this Atari ST nonsense?
I still use Steinberg Pro-24 (the forerunner to Cubase) on an Atari ST. It is far more stable than any version of Cubase that I have used on a PC. I don't own a Mac, but I do hear that recent versions of Cubase have been rock solid on that operating system. So if I upgraded that would be they way I'd go. Then again by the time I can afford a decent Mac, Rosegarden (http://www.rosegardenmusic.com/) may have reached a stable release ...
As for the poster below who talks about booting from a 720Kb floppy, my ST has a SCSI hard drive, and boots in seconds.
Unless the post-war Iraq Democracy caused Peace In the Middle East.
I wish it would, but like you I don't think it will. In Iraq itself, the new rulers have very little credibility with the ordinary people. During their time out of the country, they argued amongst themselves, and whatever grassroots support they had became disillusioned or disappeared into Saddam's prisons. Add to this the religious and tribal based allegiance that a lot of Iraqis have, and you have a recipe for weak, ineffectual government by coalitions.
No, the Russians do have the most experienced and reliable rocket engineers on earth. That's why NASA are working with them. Following the collapse of the Soviet Union, a number of Russian rocket engineers carried on perfecting the designs for extremely stable and powerfull rocket motors. This work has now been commercialised, and is used in both the Russian and American space industries. I'm sure a quick Google will turn up some suitable references.
I believe that in 20 years, we'll look back and say the same for W.
The difference between Reagan and Bush junior, is that Reagan's anti-Soviet rhetoric was mostly for public consumption. Behind the scenes, there was a great deal of diplomacy going on which ultimately lead to the arms limitation treaties. The Reagan and Bush senior regimes were much more pragmatic than the Rumsfeld / Cheney / Bush junior regime. We'll look back on the Bush junior regime in 20 years time with as much disgust as most people look at it now.
The belligerent attitude of the current regime comes as no surprise to those of us who kept up with what the various neo-con think tanks that influenced the current regime were saying in the mid-1990's. Cheif amongst their suggestions was that Saddam Hussein should be given a whipping for going against the wishes of the last Republican regime. Saddam had been the pet Middle-East strongman of the US throughout the 1980's, but he overstepped the mark by invading Kuwait. Having glossed over his previous gassing of Kurds, the Bush senior regime was thrown into turmoil by the Kuwait invasion. This is why there was a lack of firm comment on the situation from the Whitehouse in the immediate aftermath.
Just out of curiosity...didn't you set up your .sig so that you wouldn't *have* to um...sign every post?
Yup, but I have signatures switched off in my preferences and have since forgotten what my sig is.
Chris
Slighly off topic, but the origin of the word "man" was gender neutral. Some languages still make little disinction between gender (Finnish for instance).
Chris
Most programmers I've worked with in the UK have either read "Mythical Man Month" or at the very least heard of it. The same goes for Jon Bentleys "Programming Pearls".
Both books were a little bit of an anti-climax when I first read them, probably because I expected way too much in the way of blinding insights. I found I was like the bloke that Brooks sat next to on a plane journey (described in the second edition) - so much of what the book has to say seems obvious now.
However obvious those insights may seem, big projects still get bogged down with the same old problems. I guess that means managing really big projects is still a bit too much for most of us to cope with.
Chris
I had high hopes that Apple would contribute back to the community, but I don't think that has materialized like I had hoped.
Mac OS X uses the Mach kernel with a FreeBSD layer above it. This means that much of Apples work on the Mach kernel is irrelevant to FreeBSD. Mach is a microkernel, which was of course derived from BSD Unix, but it was forked so long ago that few similarities remain.
As far as stability and consistancey goes, only Debian-Stable approaches BSD
The BSD's also benefit from being a complete system, not a kernel with various userland stuff slapped together into 1001 distributions. This means that users running the development versions are using the same userland as the developers, and bugs can be shaken out far quicker.
Chris
NetBSD nor OpenBSD have comparably-sized development teams [compared to Linux]
That's not as accurate as you might think. If you look at Linux kernel development, there is a core of really talented developers who work on the guts - the VM system, VFS, threading, etc. Most other people only poke away at the edges, driver writing for instance. If you compare the number of core developers working on Linux and the NetBSD kernel, a similar number of names crop up with regularity on the relevant mailing lists. NetBSD also has a strong following in academia, and gets a lot of input from talented people with plenty of time to tinker. This partly compensates for the lower number of people working professionally on NetBSD (mostly Wasabi employees) compared to Linux. The BSD's also started from an almost complete Unix implementation (4.4BSD Lite), while Linux had to reinvent the wheel a few times.
Chris