My problem is the whole idea of using the government and tax policy to punish "bad" behavior.
What you call "bad" behaviour I define as bad behaviour. Vice versa is also likely to be true. The environment is affected hence owned by all the inhabitants of Earth, and should be protected by all of them. We humans have great power and hence also great responsibility, even in this regard. Such power is, in nowadays societies, protected by so-called democratically elected governments. Hence, this is indeed the task of our government although as of now we're free to discuss useful alternatives on Slashdot.
It really pisses me off to read about childish nerd with his toy car (SUV) which is dangerous for both pedestrians and environment. Then I feel some kind of jealousy about the low tax the childish American pays for his fuel. Then I'm reminded how well that works here in Europe and btw here in the Netherlands cars which are bad for the environment are going to pay substantially more regarding road-tax while the opposite is true for environment-friendly cars. The fuel itself ofcourse, costs still approx the same.
True, its too expensive to replace SATA. Nevermind performance compares. Just run your OS on SSD and keep your data seperate. It'll make your computer much faster, won't cost much, yet makes your setup also more flexible. I wanted to run an OS on an USB stick, but alas it was bad supported by my OS so I ended up with a SSD/PATA. Ironically, those have to reside inside the case whereas a USB stick is plugged on the outside. The latter is in some setups more flexible: have a backup copy or 1:1 copy of USB easily available for the server. If one needs replacement its sticking out and in a key (and pressing a button). Even the secretary can do such simple tasks.
3 IDE-CF adapters cost me 8$ including shipment on ebay last week.
A valid argument, but dirtier in the case and worse airflow. I'm sure they'll be more expensive from a more valid source, and you also need a CF card. My SSD/PATA is really slick and small. Perfect for an embedded computer like e.g. firewall, or even RAID system.
If under windows, make sure you turn off:
* SWAP
* ntfs Access time writes (fs tuning utility, one command from shell, or a reg key)
To lower the number of writes many additional tweaks on OS are recommended. Start in/var. For *BSD there are some sweet howtos available. Search for "USB Flash FreeBSD install howto" or sth similar. Linux is probably similar.
If you want >16GB, you can buy several, then use LVM/dynamic disk/multiple partitions depending on your OS to use that.
I don't want/need 16 GB to run an OS on, thankyouverymuch. I keep data and OS strictly on seperate devices, although I consider exceptions for e.g./usr/ports,/usr/src, and a few more directories because those tend to have more writes than others.
Most people interested in SSD do not need 16 GB either. They just need to put a (custom) OS on the 'disk' and be done with it. It won't do much writes, and will last long. The key term here is 'embedded'. If you thought you were among the target customer base with your RAID-0 (???) game desktop, alas. You're not in the target market segment by a long shot. SSD is not for you (yet).
What fascinates me, personally, is the theoretical construction of using flssh as r/o medium and a (SATA) harddisk to write data to hence r/w medium. If a file is written once and then merely accessed, it'd be written to flash medium. Else, it'd be written to the magnetic disk.
So no need to get overly excited with SSD. They're just an overpriced nicely bundled version of what is already cheaply available, kinda like external harddrives. And they'll keep on being that for a while yet.
Yes, the technology is flash plus some additional features, but you failed to understand an external harddrive (ESATA or USB) has additional advantages over an internal one. Ofcourse, it has disadvantages as well. My point is that there is no such thing as the perfect fruit for the cake.
You also failed to understand SSD has additional features CF does not have. For example, my SSD/PATA has ECC. Yes, it was more expensive, but it can sustain more writes than the average flash medium. Sound, speed and reliability are other arguments when compared with normal drives. All this stuff is nicely summed up on Wikipedia: http://en.wikipedia.org/wiki/Solid_state_drive
PS: There are some FS specifically made for flash drives such as JFF2. My Zaurus ran such.
Well, for now I've decided I want to be less depending on all kind of gadgets but later I might buy a model if it suits my needs (it has to be hackable to my taste). Hence, I've opted for an agenda with pen and paper together with efforts to improve my writing skills. One factor which forces me to write in bad quality is stress. So far, I've established wonders. I also have a laptop, but normally not used for this purpose. I don't have or need a cellphone. YMMV. IMO it all starts with organising skills. If you don't have them, no matter how good your PDA will be, it won't work well. Perhaps I simply did not need a PDA at all.
A genius IS often (if not always) someone who is 'developmentally disabled'. First, at some point on the IQ ladder, there is the definition "insane". IIRC its at 150, which is the same as the definition 'genius'. Then there is autism (especially asperger's), OCD, and other disabilities. Hence, geniuses shouldn't be excluded from this program!
The problem for an outsider is that they will pick "Linux". Now "Linux" is not "Linux" anymore because we want to punish Novell. So we have to make some political ramblings about patent deals and such. Even though the patent claims are FUD. Even though that BS doesn't exist for large parts of the world.
Hence your statement "A common enemy unites, not fractures, a community" is incorrect hoompapa. What is happening is that its fractured by 2 groups while there used to be one. Yes, the friends and foes are identified and yes the 'good' group may seem tight and yes you may argue it is the 'true group standing true to the original ideals' but... fractured, it has.
If you threaten a group of people, they will have something in common. Some will betray and backstab while the others will work together beating that enemy. Novell some of the smaller businesses are the backstabbers. While all this is happening Microsoft wants us to believe they are all ohhh and ahhh to open source, that there is no 'us and them'. I believe that while they realize its inevadable they'll lose ground they do damaga control in all kind of ways. This is an example. Its also a PR stunt to pull some sort of trick from a seemingly innocent yet incredibly dangerous kid.
As for Lenovo: they probably opted for those voucher thingies from Microsoft. Whenever a customer truelly wants to switch to Linux, Microsoft will say: "ohh here have this SUSE druggie." This is damage control, whereas Novell is Microsoft's puppy anyway.
Lockpicking? I don't know. What was used before lockpicking? Guards? Common sense from neighbors et al? What about hunting for mammoths? I don't have a good answer but I suspect there are better ones. Willing to hear yours.
Thanks, I realized that. I kinda said "D'oh" at myself.
I considered. Then again, I don't see people often vaguely talking to themselves.
And Mac OS X runs X11 apps just fine.
Yes, but that is not the issue. The issue at hand related to Cocoa is not MacOSX's incompatibility; its the other Unices (and clones) lack of compatibility. However the issue exists due to the closed nature of Apple & Co.
However, this hardly matters for the discussion of Mac OS X becoming an "official" Linux.
An official UNIX; not Linux. It does matter because UNIX has GUI too. The main one of MacOSX isn't compatible with the rest. I claim the UNIX certification doesn't cover this and hence it doesn't say everything (far from that).
If Mono isn't proprietary enough for you, maybe Qt floats your boat.
QT is GPLed and can be ported easily. The GPLed QT port is an official port and is 100% on par with the proprietary QT port. QT is ported to many platforms. The situation is different than with GNUstep & Cocoa. Those are far from 100% compatible and not ported to many platforms.
Java only became open source when Sun got scared because.NET started to cut into their market share.
Your point? What matters is that it is highly portable, open API, and there have always been testkits to see how compatible your JVM is. Apple doesn't provide such.
Considering that Google does software, for which the idea of patents is just ludicrous, whereas Apple does hardware [...]
What is this MacOSX thing people talk about?
For the last time: Apple (and e.g. IBM) do not simply 'do hardware'. They make (slick) end solutions glueing together hardware (some proprietary), software (some proprietary) and service (in-house or outsourced). This comes with a price tag, ofcourse. I'd call it simply "end-solution". Apple has both patents on their hardware as well as their software.
AMD, they do primarily hardware. VIA, hardware. Intel, hardware as core business too. Microsoft and Google, mostly software but I'd differentiate Google primarily does service (advertisement) and some hardware whereas Microsoft also do service and hardware while their traditional core business was software (and IMO still is but less dependant on it). If Apple did hardware or software you'd be able to run MacOSX on your own AMD64 w/o a hardware vendor lock-in. Or an Apple computer w/o having to buy MacOSX w/it. I mean, compare it with the old SGI (though price diff is apparent).
Many businesses which sell software as core business are pro software patents. Those involving mostly service (Google, Red Hat) and those which are underdogs are usually against software patents.
Besides, one can be pro patents while anti software patents.
Why should the theater decide between "good and bad stealing"?
We don't. We do look at circumstances though. For example: stealing one apple (with minor a) is something different than stealing a Porsche. The penalty for the offense should show that. The profit margins of stealing a Porsche are, for the offended, far higher than the profit margins of stealing an apple (heh if there even are profit margins for second hand apples?!). I rather have a homeless person stealing an apple and eating it, than a Porsche. If I were a salesmen on a market or grocery store I'd give my foods to 'em at the end of the day. I'd 'steal' them right before they'd land in trashbin (cause they cannot be sold anymore). Now, stealing is also something different than copyright infringement, but you've been told that so often I suspect you're simply trollin'.
My point is that a UNIX certification doesn't say much about the portability at a whole IOW not the dreamworld where everyone writes portable UNIX applications (that does not exist) but the practical world where MacOSX applications are often not portable. This reality check questions what the certification is worth. I admit I do not know the full answer, but regarding MacOSX GUI applications (which many run) there is a clear vendor lock-in. Perhaps you would understand it better if you were actually forced to dive in the sea with SGI 10+ years ago.
The compare is also wrong. Mono is opensource, ported to other platforms, and also portable. If you decide to switch away from Linux to Solaris, Mono would ensure you're able to use the same.NET applications on Solaris as you were running on Linux. Although you might not be able to run all.NET applications because the Mono is not a complete.NET implementation yet.
Why, you think, Java is popular? One of the reasons is the one outlined above. The fact (Sun's) Java is open sourced also allows any other OS to port Java to their OS further increasing the practical situation of portability.
Well I got Vista w/my laptop (it was still a cheap deal via my work) but I can assure you I only picked the laptop cause it was cheap hardware-wise not because of Vista. If I could have bought XP or Ubuntu I would have bought that for day a $50 cut. Alas. This makes me wonder whether they know what their customers want or just shove Vista up their throat. I'm not impressed with Vista's performance, btw.
carbon is in the process of being depreciated, and cocoa is quite portable. this makes your statement null.
Your suggestions are not practicle. With portable, we mean we can switch to some different UNIX system to avoid a vendor lock-in. You know? Apple? Vendor lock-in? Okay, now that you have that straight in your head also realize GnuStep will not aid in that at all in evading the vendor lock-in at least not right now. In theory one could use that and translate some system calls and you're a long way much like FreeBSD does Linux emulation. However almost all those GNOME and KDE apps are a mere recompile away and otherwise they're runnable with Linux libraries. With MacOSX software that is not possible either. Else we would be able to run Photoshop OSX port on Linux.
A Unix certification is a bit more than a moniker. It means that the level of software portability between Unix 03 compliant systems is guaranteed to be very high. That may not be important to you but to companies/corporations seeking to reduce costs and development times and to achieve the maximum level of reliability and portability in their business critical software a Unix 03 certification has meaning.
30 USD for for a 4 GB USB is indeed cheap, but 50 USD is probably more expensive than they are. Costs me about 45 EUR here.
1) I want to run my OS _fast_. 2) I want to _fix_ my OS from different computer easily when its broken. 3) I don't _need_ 20 GB. I need approx 4-8 GB for my OS.
USB is a good solution. A disadvantage is indeed that it sticks out.
10 years? Amazing. Where did you buy your crystal ball? Especially its detailed compare with WINE was in one word outstanding! My crystal ball says it'll take 10 years and 2 months, and I think mine is lying.:(
Good point however, to defeat a keylogger, S/Key & OPIE work great. If you suspect a keylogger you obviously don't trust the whole computer hence the monitor argument is kinda moot. In my opinion, at least.
I was more thinking about a situation where you actually have to copy the data over to a local storage device without watching the data itself while protecting the source. S/Key and OPIE only do the latter. A bugged monitor or bugged keyboard is not a viable attack against such situation. A bugged USB port or bugged I/O controller could be mitigated by having the data encrypted (but the key may not be send over the SSH connection in any way whatsoever) for example with GPG. Issueing a few remote commands on an otherwise secure server is also not a problem as long as you trust your SSH client (e.g. USB stick) which is another weak link in the whole story. Now, once you get your USB stick to a secure place with a secure computer you'll be able to read your data e.g. by entering your GPG password.
It's a little more difficult to keep changing passphrases for bulk encrypted data stored locally
Good disk encryption solutions such as LUKS and GELI support key management. However, they don't support S/Key or OPIE.
You know, one thing that struck me in the scheduler discussion yesterday was that no one said 'WTF? Why are you replacing a (working) core component of the kernel with a more-or-less untested one in a minor release?' With that kind of commitment to stability, I'm glad I don't run Linux anywhere important.
You ehm, can like, still run a version which is well tested and came with your distribution? One which, you know, contains only backported security and reliability fixes? The one ehm... the one your Linux distribution provides you? You know, much like *BSD does? Bravo, you figured it out now. Thanks for flaming. Or, if that doesn't suit you, you use the 2.6.x.[1+]-x revisions e.g. 2.6.20.2 instead of 2.6.20. This process is well documented and pointed out esp when they started to change to this way (Torvalds et all did).
Working component is very relative since it doesn't work well in some situations at all.
The FreeBSD folks are also writing new schedulers for FreeBSD 7 (as you know, since you actually talked a bit about this in the other thread), and finally are including a GEOM based journaling implementation. Neat, but again the 7.0 will also follow revisions quickly (esp if not well tested) just like 2.6.24.1 is more stable than 2.6.24. The Linux kernel follows the 'release early, release often' path and the development goes much faster than other kernels (or cores/bases). Plus, I happen to know some folks simply prefer to stick with FreeBSD 4 or 5 on production servers. Once 7 is out, the same will be true for 6 until 7 has matured a lot. You see this behaviour elsewhere too; e.g. Windows world. New != best to run. Same is true for Linux kernel. Not news, sorry.
Look these are the options. It doesn't get much simpler than how I outline it here. Take it or leave it, or throw some code/cash at developers.
I learned that actually creating the NTFS partition isn't possible.
Create the NTFS partition on Windows then. You run Windows as well, right?
Evidently, all it can do is read and, possibly, write to an existing NTFS file system.
NTFS-3G is known as write safe. For a long time.
That's no good.
Too bad for you, then. Contribute to the NTFS-3G developers by code and/or cash and/or bounties. If you can afford a Mac...
Then there is the fact that third party NTFS drivers can cause corruption.
See above; FUD. The userspace NTFS driver via FUSE (NTFS-3G) is known to be write-safe.
After FAT, ext2 comes closest. Clearly it is well supported in Linux. The OS X driver does all of those things, but it causes kernel panics (definitely not acceptable). I'm not sure, but I think the Ext2-IFS driver for Windows supports creation; it definitely supports reading and writing.
Ext2 drivers (the 2 most known Windows ones) are also known as write safe. For a long time as well. Ext3 is simply Ext2 + journaling. Journaling is not supported however the drivers do respect journaling logs. This means that if there are entries in your journal (ie. due to unclean unmount) the Windows driver will not support journaling however your FS will remain consistent. No big deal I'd say.
I fail to see how creating Ext2/Ext3 on Windows or MacOSX matters. Surely you do have a Linux machine?
If your argument is that you're scared for data loss then only use native drivers. Hence, an option could be to use NAS / networking FS such as CIFS and NFS). Also, simply create offline backups as well. It seems people who are too scared to use a 3rd party driver simply don't have a backup of their data. Trust me, if the driver would be unstable, it would have been known and the website of the driver would state such in red colours, as would Google. This is not the case. There might be rare bugs, and you may stumble upon them. Thats true for any FS or driver. So, make an offline (and offsite) backup.
FAT32 is no solution. Yes it is ancient and well supported, but it doesn't have journaling or softupdates. Which means you don't want to run it on a large partition or harddisk. Would you enjoy doing that on a 500 GB HDD? You argue you don't like data loss and such, wish to be pedantic about being able to create FS on any of your 3 OSes, yet you don't mind long (and not always consistent) fsck/scandisk? Strange.
Btw, keep an eye on ZFS. It has byte-swapping, and runs on Solaris, FreeBSD7, Linux + FUSE, and OSX 10.5 Leopard. Here I would say I'd agree w/you to wait a little bit till it is proven stable. However, w/offline backup, the issue is mitigated.
Slashdot Mirror
It really pisses me off to read about childish nerd with his toy car (SUV) which is dangerous for both pedestrians and environment. Then I feel some kind of jealousy about the low tax the childish American pays for his fuel. Then I'm reminded how well that works here in Europe and btw here in the Netherlands cars which are bad for the environment are going to pay substantially more regarding road-tax while the opposite is true for environment-friendly cars. The fuel itself ofcourse, costs still approx the same.
True, its too expensive to replace SATA. Nevermind performance compares. Just run your OS on SSD and keep your data seperate. It'll make your computer much faster, won't cost much, yet makes your setup also more flexible. I wanted to run an OS on an USB stick, but alas it was bad supported by my OS so I ended up with a SSD/PATA. Ironically, those have to reside inside the case whereas a USB stick is plugged on the outside. The latter is in some setups more flexible: have a backup copy or 1:1 copy of USB easily available for the server. If one needs replacement its sticking out and in a key (and pressing a button). Even the secretary can do such simple tasks.
Well, for now I've decided I want to be less depending on all kind of gadgets but later I might buy a model if it suits my needs (it has to be hackable to my taste). Hence, I've opted for an agenda with pen and paper together with efforts to improve my writing skills. One factor which forces me to write in bad quality is stress. So far, I've established wonders. I also have a laptop, but normally not used for this purpose. I don't have or need a cellphone. YMMV. IMO it all starts with organising skills. If you don't have them, no matter how good your PDA will be, it won't work well. Perhaps I simply did not need a PDA at all.
Makes no sense. The Zaurus is a niche, and the product line is EOL. Dead. Sharp quit developing Zaurus. I sold my Zaurus this summer because of that.
A genius IS often (if not always) someone who is 'developmentally disabled'. First, at some point on the IQ ladder, there is the definition "insane". IIRC its at 150, which is the same as the definition 'genius'. Then there is autism (especially asperger's), OCD, and other disabilities. Hence, geniuses shouldn't be excluded from this program!
My first mouse was a Genius, btw.
Its not my field, but afaik SGI solved this in MIPS with a 'cross' architecture. Imagine it like an X.
The problem for an outsider is that they will pick "Linux". Now "Linux" is not "Linux" anymore because we want to punish Novell. So we have to make some political ramblings about patent deals and such. Even though the patent claims are FUD. Even though that BS doesn't exist for large parts of the world.
Hence your statement "A common enemy unites, not fractures, a community" is incorrect hoompapa. What is happening is that its fractured by 2 groups while there used to be one. Yes, the friends and foes are identified and yes the 'good' group may seem tight and yes you may argue it is the 'true group standing true to the original ideals' but... fractured, it has.
If you threaten a group of people, they will have something in common. Some will betray and backstab while the others will work together beating that enemy. Novell some of the smaller businesses are the backstabbers. While all this is happening Microsoft wants us to believe they are all ohhh and ahhh to open source, that there is no 'us and them'. I believe that while they realize its inevadable they'll lose ground they do damaga control in all kind of ways. This is an example. Its also a PR stunt to pull some sort of trick from a seemingly innocent yet incredibly dangerous kid.
As for Lenovo: they probably opted for those voucher thingies from Microsoft. Whenever a customer truelly wants to switch to Linux, Microsoft will say: "ohh here have this SUSE druggie." This is damage control, whereas Novell is Microsoft's puppy anyway.
Lockpicking? I don't know. What was used before lockpicking? Guards? Common sense from neighbors et al? What about hunting for mammoths? I don't have a good answer but I suspect there are better ones. Willing to hear yours.
For the last time: Apple (and e.g. IBM) do not simply 'do hardware'. They make (slick) end solutions glueing together hardware (some proprietary), software (some proprietary) and service (in-house or outsourced). This comes with a price tag, ofcourse. I'd call it simply "end-solution". Apple has both patents on their hardware as well as their software.
AMD, they do primarily hardware. VIA, hardware. Intel, hardware as core business too. Microsoft and Google, mostly software but I'd differentiate Google primarily does service (advertisement) and some hardware whereas Microsoft also do service and hardware while their traditional core business was software (and IMO still is but less dependant on it). If Apple did hardware or software you'd be able to run MacOSX on your own AMD64 w/o a hardware vendor lock-in. Or an Apple computer w/o having to buy MacOSX w/it. I mean, compare it with the old SGI (though price diff is apparent).
Many businesses which sell software as core business are pro software patents. Those involving mostly service (Google, Red Hat) and those which are underdogs are usually against software patents.
Besides, one can be pro patents while anti software patents.
My point is that a UNIX certification doesn't say much about the portability at a whole IOW not the dreamworld where everyone writes portable UNIX applications (that does not exist) but the practical world where MacOSX applications are often not portable. This reality check questions what the certification is worth. I admit I do not know the full answer, but regarding MacOSX GUI applications (which many run) there is a clear vendor lock-in. Perhaps you would understand it better if you were actually forced to dive in the sea with SGI 10+ years ago.
.NET applications on Solaris as you were running on Linux. Although you might not be able to run all .NET applications because the Mono is not a complete .NET implementation yet.
The compare is also wrong. Mono is opensource, ported to other platforms, and also portable. If you decide to switch away from Linux to Solaris, Mono would ensure you're able to use the same
Why, you think, Java is popular? One of the reasons is the one outlined above. The fact (Sun's) Java is open sourced also allows any other OS to port Java to their OS further increasing the practical situation of portability.
Well I got Vista w/my laptop (it was still a cheap deal via my work) but I can assure you I only picked the laptop cause it was cheap hardware-wise not because of Vista. If I could have bought XP or Ubuntu I would have bought that for day a $50 cut. Alas. This makes me wonder whether they know what their customers want or just shove Vista up their throat. I'm not impressed with Vista's performance, btw.
30 USD for for a 4 GB USB is indeed cheap, but 50 USD is probably more expensive than they are. Costs me about 45 EUR here.
1) I want to run my OS _fast_.
2) I want to _fix_ my OS from different computer easily when its broken.
3) I don't _need_ 20 GB. I need approx 4-8 GB for my OS.
USB is a good solution. A disadvantage is indeed that it sticks out.
10 years? Amazing. Where did you buy your crystal ball? Especially its detailed compare with WINE was in one word outstanding! My crystal ball says it'll take 10 years and 2 months, and I think mine is lying. :(
I was more thinking about a situation where you actually have to copy the data over to a local storage device without watching the data itself while protecting the source. S/Key and OPIE only do the latter. A bugged monitor or bugged keyboard is not a viable attack against such situation. A bugged USB port or bugged I/O controller could be mitigated by having the data encrypted (but the key may not be send over the SSH connection in any way whatsoever) for example with GPG. Issueing a few remote commands on an otherwise secure server is also not a problem as long as you trust your SSH client (e.g. USB stick) which is another weak link in the whole story. Now, once you get your USB stick to a secure place with a secure computer you'll be able to read your data e.g. by entering your GPG password.Good disk encryption solutions such as LUKS and GELI support key management. However, they don't support S/Key or OPIE.
Working component is very relative since it doesn't work well in some situations at all.
The FreeBSD folks are also writing new schedulers for FreeBSD 7 (as you know, since you actually talked a bit about this in the other thread), and finally are including a GEOM based journaling implementation. Neat, but again the 7.0 will also follow revisions quickly (esp if not well tested) just like 2.6.24.1 is more stable than 2.6.24. The Linux kernel follows the 'release early, release often' path and the development goes much faster than other kernels (or cores/bases). Plus, I happen to know some folks simply prefer to stick with FreeBSD 4 or 5 on production servers. Once 7 is out, the same will be true for 6 until 7 has matured a lot. You see this behaviour elsewhere too; e.g. Windows world. New != best to run. Same is true for Linux kernel. Not news, sorry.
Insecure lines, potentially bugged monitors and keyboards and such are mitigated by using S/Key or OPIE. *BSD and Linux support these w/SSH.
Look these are the options. It doesn't get much simpler than how I outline it here. Take it or leave it, or throw some code/cash at developers.Create the NTFS partition on Windows then. You run Windows as well, right?NTFS-3G is known as write safe. For a long time.Too bad for you, then. Contribute to the NTFS-3G developers by code and/or cash and/or bounties. If you can afford a Mac...See above; FUD. The userspace NTFS driver via FUSE (NTFS-3G) is known to be write-safe.Ext2 drivers (the 2 most known Windows ones) are also known as write safe. For a long time as well. Ext3 is simply Ext2 + journaling. Journaling is not supported however the drivers do respect journaling logs. This means that if there are entries in your journal (ie. due to unclean unmount) the Windows driver will not support journaling however your FS will remain consistent. No big deal I'd say.
I fail to see how creating Ext2/Ext3 on Windows or MacOSX matters. Surely you do have a Linux machine?
If your argument is that you're scared for data loss then only use native drivers. Hence, an option could be to use NAS / networking FS such as CIFS and NFS). Also, simply create offline backups as well. It seems people who are too scared to use a 3rd party driver simply don't have a backup of their data. Trust me, if the driver would be unstable, it would have been known and the website of the driver would state such in red colours, as would Google. This is not the case. There might be rare bugs, and you may stumble upon them. Thats true for any FS or driver. So, make an offline (and offsite) backup.
Me, I use CIFS + UFS on a RAID fileserver, although previously I ran CIFS + Ext3 and used Ext2 r/w on Windows 2+ years ago. Succesfully. Windows supports UFS too. If I ever get a Macbook, I won't regret this situation, as I still have CIFS and NFS. Over gigabit they're quite fast. Besides, I won't get a Macbook anyway.
FAT32 is no solution. Yes it is ancient and well supported, but it doesn't have journaling or softupdates. Which means you don't want to run it on a large partition or harddisk. Would you enjoy doing that on a 500 GB HDD? You argue you don't like data loss and such, wish to be pedantic about being able to create FS on any of your 3 OSes, yet you don't mind long (and not always consistent) fsck/scandisk? Strange.
Btw, keep an eye on ZFS. It has byte-swapping, and runs on Solaris, FreeBSD7, Linux + FUSE, and OSX 10.5 Leopard. Here I would say I'd agree w/you to wait a little bit till it is proven stable. However, w/offline backup, the issue is mitigated.