I wouldn't trust anything important to Fedora, full stop. That's what RHEL and friends are for.
Bull. Just because RHEL is $$$ doesn't make it any more reliable. If fedora is unreliable then the alternative is another vendor, not RHEL.
You're being deliberately obtuse. I've already mentioned Centos and Whitebox, and that's what I meant by 'and friends'. RHEL is reliable, because RH put lots of work into QA, not because they charge for it (though arguably, this is something of a prerequisite to produce a distro of the quality that they do AND on the schedule that they do it: "good, fast, cheap - pick any two" as the saying goes). Centos and Whitebox inherity ~90% of that reliability by being based on RHEL.
The reason I wouldn't trust Fedora with anything important is not necessarily due to a perceived lack of testing on RH's part, but that they follow a 'roll-forward' policy for updates (i.e. often rolling out newer versions of upstream packages in their errata) rather than the backport approach of RHEL and Debian Stable. As a result, with the latter two distros, you can be fairly confident that the only behaviour that will change if you apply an update is the buggy and/or insecure behaviour that you want to fix. There are no such assurances with a roll-forward policy.
It is, however, fine as a general purpose desktop OS for geeks, just as the free releases of RH were.
No it isn't. Even a desktop OS needs security updates, and even "geeks" don't necessarily want to be on an upgrade treadmill.
Personally, I'm OK with a distro upgrade every year or so on my workstations, because otherwise I find that more and more of the toys I want to play with need more featureful APIs from the libraries they link against. If you upgrade at least every other FC release, you'll get continuous security updates. If you're not so keen on that, but want to use a RH-based distro, use one of the RHEL rebuilds - mostly the same as about RH7.3, but with updates until 2010.
The synaptics site lists a number of independent config tools that would appear to allow per-user preferences (e.g. QSynaptics or ltpconf).
As far as the architecture goes, I believe it's only possible to implement this functionality within the driver, as the pads need to be initialized into synaptics/ALPS mode, otherwise they default to emulating a PS/2 mouse and the extra functionality isn't available at all.
The problem is that several years after the avoidable screwup of claiming "we're walking away from home users and the desktop: go use Windows", I still can't walk into a random computer store, as far as I know, and buy a $20 boxed DVD with any kind of Red Hat Linux on it. This is bad. I can certainly walk into any computer store and buy a copy of Windows.
... for about US$300. The costs of producing a consumer OS aren't in the R&D, they're in the support. Consumers don't seem to want to pair a fair rate for the level of support they generally require.
Don't forget that Redhat's CEO Matthew Szulik Also recommended that desktop users use Windows instead of Linux [silicon.com] around the time that they dropped their desktop distributions in order to focus on enterprise Linux.
That's significantly overstating what he actually said. He was talking about consumer desktop users (as distinct from enterprise desktop users) who have no real support network, don't want to pair a fair price for access to one, and buy all sorts of funny hardware and want to install all sorts of games and applications (usually off of magazine cover CDs or bought from supermarkets) and want them to Just Work. For that type of user, Windows probably is the best choice.
* Under FC2, my laptop battery status worked. I didn't know how much time I had left to charge or left on the battery because the version of the battery status applet didn't estimate in terms of time. But under FC3, the battery applet never shows the correct information (it can't tell when I'm plugged in versus on battery, it can't tell how much the battery is losing power).
Try opening up a shell and doing:
cat </proc/acpi/battery/BAT1/state
a few times. That makes the battery applet sane on my dual-battery-capable Toshiba.
* The trackpad has hidden (and apparently unmodifiable) regions where one can use a scrollwheel-like area, and a couple of areas which function as "buttons" for back and forward in Firefox. Had I set these up myself, these would not be annoying; I could have chosen where they were located on the trackpad, if they were there at all, and possibly when I wanted them there.
This is probably from the newly-introduced synaptic driver; either disable it in xorg.conf and use the [IM]PS/2 driver to go back to FC2-style functionality, or RTFM and tune it the way you like.
With RH9, at least I didn't have to risk a system overhaul every couple of months. Updates came out regularly, and I could upgrade RPMs as needed or required.
Updates for FC1 are still available from Fedora Legacy. Alternatively, consider the Free RHEL rebuilds such as Centos or Whitebox, which promise to have updates available for as long as RHEL does (2010, as it stands at the moment).
I was running RH8 when Fedora was announced, and I was pretty nervous about it. I considered jumping ship to Debian as the next best thing (from an established RH user's point-of-view).
But now that Fedora has established itself, I'm much happier. It moves more quickly than RH's free releases used to (which always used to be a frustration for me when colleagues were raving about the latest toy they built or got installed with Mandrake or Debian unstable or Gentoo) but it's otherwise very similar to RH.
For server uses, I'll generally use RHEL (which moves/slower/ and is supported for much longer than the free releases ever were - good things in Enterprise environments) if ISV support is required (e.g. Oracle, CheckPoint FW-1) or one of the rebuilds (e.g. Centos) if the role was less critical, didn't require ISV support or I planned to replace key components anyway.
An evaluation process, that, if completed successfully, allows (mainly) government IT users to justify their usage of SLES for some roles more easily to auditors, and makes its use possible in others.
Linux needs *easy*, *default*, *out of the box* ldap-based authentication. i should be able to install a distro, select "ldap auth", and then have everything automagically authenticate against it - shell, apache, samba, IMAP, etc etc etc. same on workstations - select "ldap auth", specify the ldap server, and you're done.
RH/Fedora has been doing that at install time for ages - apparently 6.1 or so. How well it works might be another matter - I've never had cause to use it, but it'd be worth a look for anyone who hasn't seen it and discounted it already.
I wouldn't want to give the impression that the Free ATI drivers included in the standard X.org and kernel trees are flawless, but I've also had crashes from the nVidia closed driver, and I've got more faith in the open drivers becoming flawless - eventually - than closed drivers in general.
That, and wasting several hours recently trying to get nVidia's most recent drivers working with the latest FC3 errata kernels on a Toshiba notebook really puts me off using ANY closed drivers in the future (hint: revert to 6111 and patch the kernel AND the nVidia kernel driver - details here).
If you don't want to be bothered with nVidia's closed drivers, the ATI Radeon 9200/9250 should work out-of-the-box with a recent version of x.org and provide 3D acceleration.
Support companies who provide (some) code and docs for Free software!
How far back do they go? Do they set the limit before they start looking? If not, and they're prepared to look back indefinitely, they'll always find a spike!
And yet, they don't always find a spike.
So maybe they're lazy and can't be bothered to go back indefinitely every time. Or maybe they're aware that if they had a 100% hit rate, it'd give the game away and they'd lose their research grants.
Regardless, it's detail that appears to be missing from their description of their experimental method, and it's entire legitimate to poke at it. To fail to do so is to be unscientific.
About mid-way through the FAQ, it details their 'experimental method'. The rednova article points out they saw a spike 4 hours before 9/11 and 24 hours before the asian tsunami. How far back do they go? Do they set the limit before they start looking? If not, and they're prepared to look back indefinitely, they'll always find a spike! In which case, the rate of 'coincident' spikes will of course be significantly higher than random chance!
For reasons connected with my personal beliefs, I would really like this to be a genuine phenomenon.
But, like you, I think it's hooey too.
If they want to convince me, then they need to start making concrete predictions (e.g. "there will be an earthquake at X on hh:mm dd:mn:yy"). They also need to start coming up with falsifiable hypotheses to explain the devices' behaviour and start testing those hypotheses.
I've been very happy with my Samsung ML-1250. I bought it in early 2002 for 99GBP inc. A special offer meant it came with an second 1000 sheet (i.e. half-full) toner cartridge. I'm still on the first cartridge three years later, and it's done nearly 1800 sheets. I'll confess that the coverage is getting a bit uneven now, but that's not unreasonable.:-)
The toner cartridges include the drum, so about the only things that can go wrong are the motor or the fuser (both of which would probably mean doom on any SOHO laser printer anyway). You might expect the cartridges to be expensive, but they can be had for about 35GBP if you shop around (i.e. about the same price as some of the drumless cartridges out there).
The printer itself speaks PCL6, so it works fine with the pxlmono driver in ghostscript. About the only problem I've noticed is that some large PDFs and prints from Autoroute won't print, even though I've fitted an 8MB 72-pin SIMM in the slot underneath.
My company's biggest complaint with GPL is anything developed using GPL libraries must be GPL and released. They just want to make money and contribute back when it's nessisary and important.
In that case, contact the owners of the GPLed library or libraries that you wish to use and open negotiations with the aim of securing the rights you require, quite probably by offering money. If you think the owners want too much, write your own version (note that you can legitimately study the workings of the GPLed version in order to do this).
So for me (and for my dad) Linux IS READY for the desktop. At least Ubuntu is.
Ditto for me with my 65 year old father and Fedora Core 1. The install was done for me by a friend of mine (but I could have just as easily done it myself), and I just setup the printer, added a few applications, migrated his old Windows Netscape Communicator mailbox to Mozilla and copied his old data across. From time to time, I rpm -ivh applications and add icons to the desktop, but other than that, it's pretty much trouble free for both myself and him. The biggest problem I've run into was a dodgy CD-Rom drive that didn't like ripping audio CDs - easily fixed by swapping in the drive from his old machine.
Yeah, well, and I've run Win95 on a 386/16 with 4MB of RAM. So what? It's not like you can do something useful with it.
I'd certainly expect that to be true of that combination. But both myself and my housemate performed useful work with both our machines at the time; graphical web browsing, email, software development. I even ran early versions of StarOffice, KDE and VMware on my 486 machine.
In fact, I don't exactly understand what the hell third world countries need computers for, anyway.
Assuming you're not just being an annoying little troll, how about things like a farmer shopping around to see whether the price they've been offered for their produce is the best they can obtain, or even fair? That's just one use I'm aware of. Given that our lifestyles are very different from theirs, I wouldn't expect to be able to come up with genuinely useful applications for them. But I'm certain that if the hardware is made available at an affordable price, and they're made aware of the sort of things it can do, they'll have enough smart people to come up with applications that are useful to them.
Why the hell would you want an ancient computer that can barely run X11? Normal, graphical Linux doesn't run that well on anything less than 500MHz / 128M of ram.
You're right if you're talking about a modern Linux distro, but I started running Linux in 1995 on a DX4/100 with 8MB and a 2MB S3 video card. My housemate had a 386 DX40 with 5MB and some kind of Trident-based ISA card. Both of us used X11 most of the time.
To someone who's never used a computer before, and has no preconceptions, even apps using the Athena widget set will look fine (assuming Athena can be used or otherwise hacked to use non-Western fonts, or the users understand English).
Re:I'd be happy to pay that without a display
on
The Hundred-Buck PC
·
· Score: 1
Pretty favorably, a desktop LCD consumes maybe half to two thirds the power of CRT for the same size screen?
Good guesstimate; the 17" TFT in front of me uses 19V at 2.6A making consumption just under 50W (though the PSU probably wastes some more on top of that). I usually reckon on a CRT using anywhere from 70-140W.
Slashdot Mirror
Bull. Just because RHEL is $$$ doesn't make it any more reliable. If fedora is unreliable then the alternative is another vendor, not RHEL.
You're being deliberately obtuse. I've already mentioned Centos and Whitebox, and that's what I meant by 'and friends'. RHEL is reliable, because RH put lots of work into QA, not because they charge for it (though arguably, this is something of a prerequisite to produce a distro of the quality that they do AND on the schedule that they do it: "good, fast, cheap - pick any two" as the saying goes). Centos and Whitebox inherity ~90% of that reliability by being based on RHEL.
The reason I wouldn't trust Fedora with anything important is not necessarily due to a perceived lack of testing on RH's part, but that they follow a 'roll-forward' policy for updates (i.e. often rolling out newer versions of upstream packages in their errata) rather than the backport approach of RHEL and Debian Stable. As a result, with the latter two distros, you can be fairly confident that the only behaviour that will change if you apply an update is the buggy and/or insecure behaviour that you want to fix. There are no such assurances with a roll-forward policy.
It is, however, fine as a general purpose desktop OS for geeks, just as the free releases of RH were.
No it isn't. Even a desktop OS needs security updates, and even "geeks" don't necessarily want to be on an upgrade treadmill.
Personally, I'm OK with a distro upgrade every year or so on my workstations, because otherwise I find that more and more of the toys I want to play with need more featureful APIs from the libraries they link against. If you upgrade at least every other FC release, you'll get continuous security updates. If you're not so keen on that, but want to use a RH-based distro, use one of the RHEL rebuilds - mostly the same as about RH7.3, but with updates until 2010.
As far as the architecture goes, I believe it's only possible to implement this functionality within the driver, as the pads need to be initialized into synaptics/ALPS mode, otherwise they default to emulating a PS/2 mouse and the extra functionality isn't available at all.
It is, however, fine as a general purpose desktop OS for geeks, just as the free releases of RH were.
That's significantly overstating what he actually said. He was talking about consumer desktop users (as distinct from enterprise desktop users) who have no real support network, don't want to pair a fair price for access to one, and buy all sorts of funny hardware and want to install all sorts of games and applications (usually off of magazine cover CDs or bought from supermarkets) and want them to Just Work. For that type of user, Windows probably is the best choice.
Try opening up a shell and doing:
cat </proc/acpi/battery/BAT1/state
a few times. That makes the battery applet sane on my dual-battery-capable Toshiba.
* The trackpad has hidden (and apparently unmodifiable) regions where one can use a scrollwheel-like area, and a couple of areas which function as "buttons" for back and forward in Firefox. Had I set these up myself, these would not be annoying; I could have chosen where they were located on the trackpad, if they were there at all, and possibly when I wanted them there.
This is probably from the newly-introduced synaptic driver; either disable it in xorg.conf and use the [IM]PS/2 driver to go back to FC2-style functionality, or RTFM and tune it the way you like.
Updates for FC1 are still available from Fedora Legacy. Alternatively, consider the Free RHEL rebuilds such as Centos or Whitebox, which promise to have updates available for as long as RHEL does (2010, as it stands at the moment).
But now that Fedora has established itself, I'm much happier. It moves more quickly than RH's free releases used to (which always used to be a frustration for me when colleagues were raving about the latest toy they built or got installed with Mandrake or Debian unstable or Gentoo) but it's otherwise very similar to RH.
For server uses, I'll generally use RHEL (which moves /slower/ and is supported for much longer than the free releases ever were - good things in Enterprise environments) if ISV support is required (e.g. Oracle, CheckPoint FW-1) or one of the rebuilds (e.g. Centos) if the role was less critical, didn't require ISV support or I planned to replace key components anyway.
An evaluation process, that, if completed successfully, allows (mainly) government IT users to justify their usage of SLES for some roles more easily to auditors, and makes its use possible in others.
RH/Fedora has been doing that at install time for ages - apparently 6.1 or so. How well it works might be another matter - I've never had cause to use it, but it'd be worth a look for anyone who hasn't seen it and discounted it already.
The appropriate reference to the RHEL manual
That, and wasting several hours recently trying to get nVidia's most recent drivers working with the latest FC3 errata kernels on a Toshiba notebook really puts me off using ANY closed drivers in the future (hint: revert to 6111 and patch the kernel AND the nVidia kernel driver - details here).
If you don't want to be bothered with nVidia's closed drivers, the ATI Radeon 9200/9250 should work out-of-the-box with a recent version of x.org and provide 3D acceleration.
Support companies who provide (some) code and docs for Free software!
a good recipe for a 1.2TB Linux-based RAID for under US$1600.
Show me.
Does it? Have you actually read their material, or just the brief summary linked to by Slashdot?
Yes, I've read as much as I could bear of their material.
and it's entire legitimate to poke at it. To fail to do so is to be unscientific.
Kind of like accusing them of incompetence or corruption without backing evidence, huh?
Perhaps you need to check the definition of 'maybe' in the dictionary. Or do you believe everyone is competent and incorrupt all the time?
And yet, they don't always find a spike.
So maybe they're lazy and can't be bothered to go back indefinitely every time. Or maybe they're aware that if they had a 100% hit rate, it'd give the game away and they'd lose their research grants.
Regardless, it's detail that appears to be missing from their description of their experimental method, and it's entire legitimate to poke at it. To fail to do so is to be unscientific.
About mid-way through the FAQ, it details their 'experimental method'. The rednova article points out they saw a spike 4 hours before 9/11 and 24 hours before the asian tsunami. How far back do they go? Do they set the limit before they start looking? If not, and they're prepared to look back indefinitely, they'll always find a spike! In which case, the rate of 'coincident' spikes will of course be significantly higher than random chance!
But, like you, I think it's hooey too.
If they want to convince me, then they need to start making concrete predictions (e.g. "there will be an earthquake at X on hh:mm dd:mn:yy"). They also need to start coming up with falsifiable hypotheses to explain the devices' behaviour and start testing those hypotheses.
The toner cartridges include the drum, so about the only things that can go wrong are the motor or the fuser (both of which would probably mean doom on any SOHO laser printer anyway). You might expect the cartridges to be expensive, but they can be had for about 35GBP if you shop around (i.e. about the same price as some of the drumless cartridges out there).
The printer itself speaks PCL6, so it works fine with the pxlmono driver in ghostscript. About the only problem I've noticed is that some large PDFs and prints from Autoroute won't print, even though I've fitted an 8MB 72-pin SIMM in the slot underneath.
In that case, contact the owners of the GPLed library or libraries that you wish to use and open negotiations with the aim of securing the rights you require, quite probably by offering money. If you think the owners want too much, write your own version (note that you can legitimately study the workings of the GPLed version in order to do this).
Ditto for me with my 65 year old father and Fedora Core 1. The install was done for me by a friend of mine (but I could have just as easily done it myself), and I just setup the printer, added a few applications, migrated his old Windows Netscape Communicator mailbox to Mozilla and copied his old data across. From time to time, I rpm -ivh applications and add icons to the desktop, but other than that, it's pretty much trouble free for both myself and him. The biggest problem I've run into was a dodgy CD-Rom drive that didn't like ripping audio CDs - easily fixed by swapping in the drive from his old machine.
I'd certainly expect that to be true of that combination. But both myself and my housemate performed useful work with both our machines at the time; graphical web browsing, email, software development. I even ran early versions of StarOffice, KDE and VMware on my 486 machine.
In fact, I don't exactly understand what the hell third world countries need computers for, anyway.
Assuming you're not just being an annoying little troll, how about things like a farmer shopping around to see whether the price they've been offered for their produce is the best they can obtain, or even fair? That's just one use I'm aware of. Given that our lifestyles are very different from theirs, I wouldn't expect to be able to come up with genuinely useful applications for them. But I'm certain that if the hardware is made available at an affordable price, and they're made aware of the sort of things it can do, they'll have enough smart people to come up with applications that are useful to them.
Tools->Options->Load/Save->General->[Standard File Format|Text Document|Always Save As...|Microsoft Word xxx]
You're right if you're talking about a modern Linux distro, but I started running Linux in 1995 on a DX4/100 with 8MB and a 2MB S3 video card. My housemate had a 386 DX40 with 5MB and some kind of Trident-based ISA card. Both of us used X11 most of the time.
To someone who's never used a computer before, and has no preconceptions, even apps using the Athena widget set will look fine (assuming Athena can be used or otherwise hacked to use non-Western fonts, or the users understand English).
Good guesstimate; the 17" TFT in front of me uses 19V at 2.6A making consumption just under 50W (though the PSU probably wastes some more on top of that). I usually reckon on a CRT using anywhere from 70-140W.