Question: if going from 15% market share to 2% in a straight line over 15 years is success, what would failure look like?
Commodore, Be, Atari, Digital Research, IBM...
Note that last one. Microsoft managed to beat the biggest and most successful company in the computer industry in this market. While I don't agree with a lot of what Jobs has done, I have to say that just surviving is pretty good.
Also, that's 2.5% worldwide, 4% in the US, and it's hardly a straight line.
The customer is going to see two identically specified machines.
You really think the customer is stupid enough that they can't already compare the specs on a Mac and a Windows box, and note that the Windows box costs less for better specs? You've never actually talked to any ordinary PC buyers if you think they're going to suddenly realise that Macs are expensive PCs. Why do you think Mac apologists have been trying (unsuccessfully) to argue that there isn't a "Mac Tax" for years? This isn't going to change any of that.
From Paul's article: Because Microsoft built a search box into the Start menu, you can no longer use keyboard shortcuts to navigate around. To launch the Control Panel in XP, for example, you simply hit the Windows key and then the "C" key and, voila, the Control Panel opens. In Windows Vista, however, when you hit the "C" key, the system assumes you're searching for an application (Figure). Sigh.
For me, the user interface of Windows peaked with Windows 3.11 and NT 3.51. In these systems, virtually every control in every program could be easily navigated to using only the keyboard, with consistent shortcuts everywhere. This was a significantly better environment than Apple has managed to provide even now, and probably the best feature of the Windows UI. In 95/NT4 the Start Menu and Task Bar required new shortcuts. Then companies started shipping keyboards with extra keys (making the spacebar shorter and a harder target to hit, and not really solving the problem for people who have to work on multiple computers with a variety of keyboards). Newer versions of Office applications removed the ability to keyboard-navigate through toolbars (with or without he new keyboards). What's next?
So rather than implement a sandbox in the browser, they run the whole browser in a sandbox. This means that hostile scripts and ActiveX components can still be used to attack other systems, compromise the user's personal information on other web sites, steal passwords and credit card numbers, and take part in zombie networks.
On a lighter note, I'm not sure that having self-aware COM objects is a good idea. Apart from this being a dubious application of strong AI technology, won't this make shutting down your computer equivalent to murder?
Doesn't Cisco use a number of different chips in their routers and other devices? For example, there's a port of NetBSD to Cisco's 680x0-based routers. And wasn't the PIX originally x86-based?
Re:Why FreeBSD is not good for most businesses
on
Why FreeBSD
·
· Score: 1
I understand why Linux does things differently.
What I'm trying to get across to you is that it's NOT the only way to do things, and there are advantages to doing things other ways.
Linux has one kernel and maybe a couple of dozen distros, more if you include the embedded variants. BSD has four major open source variants and maybe as many minor ones, and there's more similarities between any of these than there is between the major Linux distros.
Linux has one person who has commit rights on the kernel, and other packages have different groups but typically a handful of people at most who are actually managing each individual package. FreeBSD has a core team and dozens of comitters, any of whom have commit rights on the core.
You prefer the Linux model. Fine. I don't expect you to change that opinion, but while I prefer the BSD model I can see where Linux's strengths are as well as its weaknesses. Can you make the corresponding jump, and instead of reacting with "BSD's different and therefore wrong", or "BSD can only be the way it is because it's unpopular", consider that there's strengths to both designs and actual good reasons that keep BSD the way it is?
Re:Why FreeBSD is not good for most businesses
on
Why FreeBSD
·
· Score: 1
And packages aren't?
I haven't run into many that are more than minimally configured, let alone modified to the extent that the stuff that's gone into the FreeBSD source tree has been.
It's not the dodgy greenbacks I'm worried about, it's the other uses of the information.
"Oh look, this tax return was printed with the same printer used to print this flyer complaining about the administration's policies on Iraq, better add him to the blacklist too."
New CD at Best Buy, containing two songs you want to listen to,six you're willing to listen to, and seven you'll just fast-forward through or you already have on another CD: $12.00
Buying the two songs you actually like at iTunes: $2.00.
The fact is that most of the folks complaining that OS X won't run on generic hardware are folks who want to run OS X on their own hardware.
You mean like my Powermac 7500? Which I've been running OS X on? Until Apple came up with the Mini, I was planning on keeping on running OS X on my 7500 (or a Beige G3) because I didn't like any of Apple's newer hardware (none that I could afford, anyway)... back in January even a Powermac G4/400 cost MORE than a Mac mini!
My guess is that most of these same people wouldn't bother paying for the OS either. They'd just pirate it.
I bought both Puma and Jaguar for my 7500. Even though Apple didn't support it and I had to buy a third party processor upgrade and use a third party hack to make it run.
I'd buy Leopard for my Thinkpad, if Apple would sell it to me. Unless Apple teams up with IBM (err, Lenovo) again and produces a Thinkpad-influenced Powerbook, though, I'm out of luck. Because none of the current crop of Apple's laptops appeal to me at all.
Seriously, these are people who think Linux is ready for the desktop.
Dude, the ones who think Linux is ready for the desktop aren't the ones who are salivating over a generic Mac OS X box... they're already running Linux on the desktop, remember?
Let's say Apple releases OS X x86 for generic x86 hardware. It's a box right next to Windows XP.
Let's say they don't. Let's say Apple releases OS X x86 through HP and Lenovo, so you can start buying Thinkpads and Pavillions running OS X and don't have to deal with Apple's Bloody Stupid Hardware. Yes, stupid. Powerbook keyboards suck, and I'd have paid more for a real desktop Mac than I did for a Mini but Apple thinks I'm independently wealthy and have two grand to splurge on a G5. No, the eMac and iMac aren't even in the running.
You go on about this possibility as if it's some big secret that ONLY YOU could have thought of. It's not. There's a whole boatload of people out there who are really bloody tired of Steve Jobs boutique designs and just want a regular desktop computer that doesn't suck running an OS that doesn't suck.
Apple hasn't made one since Steve Jobs took over. The only good thing about the Intel switch is that it provides a possibility of that happening again. Whether that involves buying a copy of OSX Intel at Best Buy and sticking it on my generic PC, or buying an HPMac running it, I don't care, and neither do most of the other people who speculate about what Apple might do next.
Unfortunately, I find it hard to believe that the New Apple is willing to let that happen.
Palm's problem is that Hawkins doesn't actually give a shit for his customers, and just like making cool shit until he gets bored with it and dumps it for the next cool shit he comes up with.
Palm should still be making cheap (really cheap, as cheap as Moore's-law lets them, as much cheaper than a Zire as a Shuffle is cheaper than an iPod Mini... $40 or less) low-power long-battery-life 68000-based handhelds, and they should be on the checkout line of every Kroger and Walmart in the country.
Plus, they should be doing whatever Sony wants to get them to bring back *their* 68000-based Palms in the US market.
Or, if they're going to put an ARM in there, upgrade the OS so the apps are running native ARM code instead of mostly running under a 68000 emulator... so they can get decent performance on a low end low power LOW PRICE processor running at 40 or 50 MHz.
Palm is like Apple would be if Apple was still running Mac OS 8.1 under 68000 emulation on their G5s. There's nothing Apple or anyone else can learn from the mess Palm has made of what used to be the only credible handheld platform out there.
I bought a Mac Mini for my daughter, and I'd happily buy OSX Intel for my Thinkpad if Apple was willing to sell it to me... but I'm still using my Clie SJ22 because there's only been one handheld made since then that's even attracted my interest... and it's discontinued too...
If Mercedes-Benz was finding that having 1% market share meant that you could only refuel at 1% of the gas stations, they'd do something to increase their market share in a hurry. Like, oh, buying a generic US auto-maker like Chrysler.
Hold on, that happened, didn't it? I guess even a luxury car maker needs a commodity product to keep the bottom line looking good.
The day that Apple starts allowing MacOS to run on any old computer with the right CPU is the day that I stop buying Apple products.
Promise? Will you stop posting bad analogies to/. as well?
Re:Why FreeBSD is not good for most businesses
on
Why FreeBSD
·
· Score: 1
I'm using core in the same sense FreeBSD does - the group of packages that are maintained and released as a single block.
That's *not* what it means.
The FreeBSD core is not built out of packages.
Let me repeat that, because you seem to be missing it, and it's a really really important point.
The FreeBSD core is not built out of packages.
It's a single source tree. Even the stuff under "contrib" is modified to fit the BSD build and configuration conventions. This gives it a level of consistency that's simply impossible in a package-based system.
It's like the difference between "off the shelf" and "tailored".
The only place Apple really could get away with locking you into Apple hardware is in Quartz. The kernel is open source and people are already running Darwin/XNU on generic PC hardware. Most of the basic drivers, likewise, are open source. But they could easily make the Quartz acceleration only work on Apple video cards. They kind of do already... ATI's video cards for Apple don't just have PPC-compatible firmware, they have a bunch of Apple-specific OpenGL extensions.
Re:Why FreeBSD is not good for most businesses
on
Why FreeBSD
·
· Score: 1
I prefer Linux because the stable core is much larger than BSDs.
Linux doesn't *have* a core system. It's packages all the way down to the kernel itself.
I wouldn't consider an authority someone who's so out of touch that he's working to REDUCE the security of his product by having everyone run as root, when even Microsoft is trying to go the other way and Apple has made "root" unnecessary.
Re:Why FreeBSD is not good for most businesses
on
Why FreeBSD
·
· Score: 1
I think that software that many users depend upon should be a part of, and maintained with, the stable core.
And yet you prefer Linux because it's all packages, with no stable core.
Re:Why FreeBSD is not good for most businesses
on
Why FreeBSD
·
· Score: 1
For BSD to become popular, there is every reason for core to expand.
Why?
Would you expect people to move the shell into the Linux kernel just because Linux is getting more popular? The whole point to a stable core and optional packages is to make it possible to ship glitzy systems without screwing up the core.
It's not "the amount of software users depend on", it's "the amount of software that other software depends on". The "other software that users depend on" goes in optional packages, so users can choose what they depend on. Or not.
Re:Why FreeBSD is not good for most businesses
on
Why FreeBSD
·
· Score: 1
As I've said, that might change in the next release.
There's a very low possibility of that. The core of BSD is deliberately conservative, and there's no reason for it to change. There's MUCH less reason for it to change than a system built out of packages.
What was that you said about BSDs consistency?
The core is consistent and stable. Optional packages shipped *with* the OS are just that, optional. None of them are installed by default, not even a "desktop" - even if you DO select X11.
But the 'part of OS / not part of OS' thing I think is artificial in an OSS environment
It's no more or less artificial than in a commercial OS. "part of the OS" means "you can depend on it being there" and "you can depend on it working the same way". Applications are only moved into the core when they are already sufficiently widely used that people have already come to expect them. Moving a version of Perl into the core (even in contrib) was a major step.
Sometimes I'd rather a new version of that app be installed. It minimises redundancy.
Yeh, but a program that's written to deal with a quirk in/usr/bin/perl can continue to use/usr/bin/perl. It won't break when you install a newer one in/usr/local/bin.
Redundancy can be a concern, but given the relative sizes of a default FreeBSD install and a default Red Hat install a bit of redundancy can be permitted.
I had expected to see G4s going for under $500 very soon after the Mac mini came out, but it didn't happen. The price of G4s stayed steady, dropping no faster than they had been, until the Intel announcement.
Then they dropped like a bomb. I've been offered a dual G4/550 for $350, or a stripped G4/400 for $150. I wish the Mini had had that effect, because I was trying to get a cheap G4 a couple of months back and finally went for the Mini instead.
But your G4/867 (MDD, I assume)? It's got a faster hard drive than the Mini, it supports twice the RAM, it supports Core Graphics in the GPU with a Radeon 9600 or better video card. You can upgrade it (thanks to Sonnet and their pals) all the way to a dual 2 GHz G4. Depending on how you have it actually loaded, it could be quite comparable to a $600 Mini.
But not until the aftershocks of the Intel bomb settle down, I suspect.
Although your point has some teeth, please realize that there is still a big performance disconnect between static and dynamic RAM.
Yeh, but are you going to get that performance out of this gadget? They benchmarked it as only 6x the speed of an actual hard drive.
Also, flash RAM tends to have limited write cycles.
I wouldn't want to use it for swap, and I'd mount noatime, but for a boot disk for a solid-state server this would be fine. People have been using flash drives for embedded systems, even embedded UNIX and Linux systems, for years.
For the Gigabyte product? I wouldn't use it for swap... I'd get better performance out of just using the same RAM chips directly on the motherboard. In fact just about anything you coudl use it for other than a solid-state boot disk you'd be better off just adding RAM.
4GB, populated, probably fits in a 2.5" drive bay let alone 3.5". There's actually versions that have 2.5" connectors and mounting brackets so you can use them directly in a laptop, but I couldn't quickly google a price on them.
From the summary it sounded like a 2.5" or 3.5" 4GB IDE drive using flash instead of an IDE emulator and battery-backed up RAM using a PCI slot for power... and no memory included!
I'd pay $100 for 4GB of flash in a PCI or hard drive form factor, for a solid-state BSD or Linux webserver.
I don't think I'd pay $100 for a 0GB hard drive emulator that takes up both a PCI slot AND an SATA cable, and I still have to populate with RAM, and that will lose all its data if you leave it off too long.
Given that you can get a 2GB Compact Flash drive for $100 or 4GB for around $200 and you can hook those up to PATA with a $40 adapter, and populating this thing to 4GB will set me back more than that... I don't see the point.
Slashdot Mirror
Question: if going from 15% market share to 2% in a straight line over 15 years is success, what would failure look like?
Commodore, Be, Atari, Digital Research, IBM...
Note that last one. Microsoft managed to beat the biggest and most successful company in the computer industry in this market. While I don't agree with a lot of what Jobs has done, I have to say that just surviving is pretty good.
Also, that's 2.5% worldwide, 4% in the US, and it's hardly a straight line.
The customer is going to see two identically specified machines.
You really think the customer is stupid enough that they can't already compare the specs on a Mac and a Windows box, and note that the Windows box costs less for better specs? You've never actually talked to any ordinary PC buyers if you think they're going to suddenly realise that Macs are expensive PCs. Why do you think Mac apologists have been trying (unsuccessfully) to argue that there isn't a "Mac Tax" for years? This isn't going to change any of that.
From Paul's article: Because Microsoft built a search box into the Start menu, you can no longer use keyboard shortcuts to navigate around. To launch the Control Panel in XP, for example, you simply hit the Windows key and then the "C" key and, voila, the Control Panel opens. In Windows Vista, however, when you hit the "C" key, the system assumes you're searching for an application (Figure). Sigh.
For me, the user interface of Windows peaked with Windows 3.11 and NT 3.51. In these systems, virtually every control in every program could be easily navigated to using only the keyboard, with consistent shortcuts everywhere. This was a significantly better environment than Apple has managed to provide even now, and probably the best feature of the Windows UI. In 95/NT4 the Start Menu and Task Bar required new shortcuts. Then companies started shipping keyboards with extra keys (making the spacebar shorter and a harder target to hit, and not really solving the problem for people who have to work on multiple computers with a variety of keyboards). Newer versions of Office applications removed the ability to keyboard-navigate through toolbars (with or without he new keyboards). What's next?
So rather than implement a sandbox in the browser, they run the whole browser in a sandbox. This means that hostile scripts and ActiveX components can still be used to attack other systems, compromise the user's personal information on other web sites, steal passwords and credit card numbers, and take part in zombie networks.
On a lighter note, I'm not sure that having self-aware COM objects is a good idea. Apart from this being a dubious application of strong AI technology, won't this make shutting down your computer equivalent to murder?
Doesn't Cisco use a number of different chips in their routers and other devices? For example, there's a port of NetBSD to Cisco's 680x0-based routers. And wasn't the PIX originally x86-based?
I understand why Linux does things differently.
What I'm trying to get across to you is that it's NOT the only way to do things, and there are advantages to doing things other ways.
Linux has one kernel and maybe a couple of dozen distros, more if you include the embedded variants. BSD has four major open source variants and maybe as many minor ones, and there's more similarities between any of these than there is between the major Linux distros.
Linux has one person who has commit rights on the kernel, and other packages have different groups but typically a handful of people at most who are actually managing each individual package. FreeBSD has a core team and dozens of comitters, any of whom have commit rights on the core.
You prefer the Linux model. Fine. I don't expect you to change that opinion, but while I prefer the BSD model I can see where Linux's strengths are as well as its weaknesses. Can you make the corresponding jump, and instead of reacting with "BSD's different and therefore wrong", or "BSD can only be the way it is because it's unpopular", consider that there's strengths to both designs and actual good reasons that keep BSD the way it is?
And packages aren't?
I haven't run into many that are more than minimally configured, let alone modified to the extent that the stuff that's gone into the FreeBSD source tree has been.
It's not the dodgy greenbacks I'm worried about, it's the other uses of the information.
"Oh look, this tax return was printed with the same printer used to print this flyer complaining about the administration's policies on Iraq, better add him to the blacklist too."
New CD at Best Buy, at a cut-rate price: $12.00
New CD at Best Buy, containing two songs you want to listen to,six you're willing to listen to, and seven you'll just fast-forward through or you already have on another CD: $12.00
Buying the two songs you actually like at iTunes: $2.00.
The fact is that most of the folks complaining that OS X won't run on generic hardware are folks who want to run OS X on their own hardware.
You mean like my Powermac 7500? Which I've been running OS X on? Until Apple came up with the Mini, I was planning on keeping on running OS X on my 7500 (or a Beige G3) because I didn't like any of Apple's newer hardware (none that I could afford, anyway)... back in January even a Powermac G4/400 cost MORE than a Mac mini!
My guess is that most of these same people wouldn't bother paying for the OS either. They'd just pirate it.
I bought both Puma and Jaguar for my 7500. Even though Apple didn't support it and I had to buy a third party processor upgrade and use a third party hack to make it run.
I'd buy Leopard for my Thinkpad, if Apple would sell it to me. Unless Apple teams up with IBM (err, Lenovo) again and produces a Thinkpad-influenced Powerbook, though, I'm out of luck. Because none of the current crop of Apple's laptops appeal to me at all.
What is wrong with this argument is that it confuses a strategy with a market segment.
That's a really nice way of putting it. It's the perfect answer to the people who go on about Apple being the Mercedes of computing.
Because Mercedes didn't confuse a strategy with a market segment.
That's why they bought Chrysler.
Seriously, these are people who think Linux is ready for the desktop.
Dude, the ones who think Linux is ready for the desktop aren't the ones who are salivating over a generic Mac OS X box... they're already running Linux on the desktop, remember?
Let's say Apple releases OS X x86 for generic x86 hardware. It's a box right next to Windows XP.
Let's say they don't. Let's say Apple releases OS X x86 through HP and Lenovo, so you can start buying Thinkpads and Pavillions running OS X and don't have to deal with Apple's Bloody Stupid Hardware. Yes, stupid. Powerbook keyboards suck, and I'd have paid more for a real desktop Mac than I did for a Mini but Apple thinks I'm independently wealthy and have two grand to splurge on a G5. No, the eMac and iMac aren't even in the running.
You go on about this possibility as if it's some big secret that ONLY YOU could have thought of. It's not. There's a whole boatload of people out there who are really bloody tired of Steve Jobs boutique designs and just want a regular desktop computer that doesn't suck running an OS that doesn't suck.
Apple hasn't made one since Steve Jobs took over. The only good thing about the Intel switch is that it provides a possibility of that happening again. Whether that involves buying a copy of OSX Intel at Best Buy and sticking it on my generic PC, or buying an HPMac running it, I don't care, and neither do most of the other people who speculate about what Apple might do next.
Unfortunately, I find it hard to believe that the New Apple is willing to let that happen.
Palm's problem is that Hawkins doesn't actually give a shit for his customers, and just like making cool shit until he gets bored with it and dumps it for the next cool shit he comes up with.
Palm should still be making cheap (really cheap, as cheap as Moore's-law lets them, as much cheaper than a Zire as a Shuffle is cheaper than an iPod Mini... $40 or less) low-power long-battery-life 68000-based handhelds, and they should be on the checkout line of every Kroger and Walmart in the country.
Plus, they should be doing whatever Sony wants to get them to bring back *their* 68000-based Palms in the US market.
Or, if they're going to put an ARM in there, upgrade the OS so the apps are running native ARM code instead of mostly running under a 68000 emulator... so they can get decent performance on a low end low power LOW PRICE processor running at 40 or 50 MHz.
Palm is like Apple would be if Apple was still running Mac OS 8.1 under 68000 emulation on their G5s. There's nothing Apple or anyone else can learn from the mess Palm has made of what used to be the only credible handheld platform out there.
I bought a Mac Mini for my daughter, and I'd happily buy OSX Intel for my Thinkpad if Apple was willing to sell it to me... but I'm still using my Clie SJ22 because there's only been one handheld made since then that's even attracted my interest... and it's discontinued too...
If Mercedes-Benz was finding that having 1% market share meant that you could only refuel at 1% of the gas stations, they'd do something to increase their market share in a hurry. Like, oh, buying a generic US auto-maker like Chrysler.
/. as well?
Hold on, that happened, didn't it? I guess even a luxury car maker needs a commodity product to keep the bottom line looking good.
The day that Apple starts allowing MacOS to run on any old computer with the right CPU is the day that I stop buying Apple products.
Promise? Will you stop posting bad analogies to
I'm using core in the same sense FreeBSD does - the group of packages that are maintained and released as a single block.
That's *not* what it means.
The FreeBSD core is not built out of packages.
Let me repeat that, because you seem to be missing it, and it's a really really important point.
The FreeBSD core is not built out of packages.
It's a single source tree. Even the stuff under "contrib" is modified to fit the BSD build and configuration conventions. This gives it a level of consistency that's simply impossible in a package-based system.
It's like the difference between "off the shelf" and "tailored".
The only place Apple really could get away with locking you into Apple hardware is in Quartz. The kernel is open source and people are already running Darwin/XNU on generic PC hardware. Most of the basic drivers, likewise, are open source. But they could easily make the Quartz acceleration only work on Apple video cards. They kind of do already... ATI's video cards for Apple don't just have PPC-compatible firmware, they have a bunch of Apple-specific OpenGL extensions.
I prefer Linux because the stable core is much larger than BSDs.
Linux doesn't *have* a core system. It's packages all the way down to the kernel itself.
I wouldn't consider an authority someone who's so out of touch that he's working to REDUCE the security of his product by having everyone run as root, when even Microsoft is trying to go the other way and Apple has made "root" unnecessary.
I think that software that many users depend upon should be a part of, and maintained with, the stable core.
And yet you prefer Linux because it's all packages, with no stable core.
For BSD to become popular, there is every reason for core to expand.
Why?
Would you expect people to move the shell into the Linux kernel just because Linux is getting more popular? The whole point to a stable core and optional packages is to make it possible to ship glitzy systems without screwing up the core.
It's not "the amount of software users depend on", it's "the amount of software that other software depends on". The "other software that users depend on" goes in optional packages, so users can choose what they depend on. Or not.
As I've said, that might change in the next release.
/usr/bin/perl can continue to use /usr/bin/perl. It won't break when you install a newer one in /usr/local/bin.
There's a very low possibility of that. The core of BSD is deliberately conservative, and there's no reason for it to change. There's MUCH less reason for it to change than a system built out of packages.
What was that you said about BSDs consistency?
The core is consistent and stable. Optional packages shipped *with* the OS are just that, optional. None of them are installed by default, not even a "desktop" - even if you DO select X11.
But the 'part of OS / not part of OS' thing I think is artificial in an OSS environment
It's no more or less artificial than in a commercial OS. "part of the OS" means "you can depend on it being there" and "you can depend on it working the same way". Applications are only moved into the core when they are already sufficiently widely used that people have already come to expect them. Moving a version of Perl into the core (even in contrib) was a major step.
Sometimes I'd rather a new version of that app be installed. It minimises redundancy.
Yeh, but a program that's written to deal with a quirk in
Redundancy can be a concern, but given the relative sizes of a default FreeBSD install and a default Red Hat install a bit of redundancy can be permitted.
I had expected to see G4s going for under $500 very soon after the Mac mini came out, but it didn't happen. The price of G4s stayed steady, dropping no faster than they had been, until the Intel announcement.
Then they dropped like a bomb. I've been offered a dual G4/550 for $350, or a stripped G4/400 for $150. I wish the Mini had had that effect, because I was trying to get a cheap G4 a couple of months back and finally went for the Mini instead.
But your G4/867 (MDD, I assume)? It's got a faster hard drive than the Mini, it supports twice the RAM, it supports Core Graphics in the GPU with a Radeon 9600 or better video card. You can upgrade it (thanks to Sonnet and their pals) all the way to a dual 2 GHz G4. Depending on how you have it actually loaded, it could be quite comparable to a $600 Mini.
But not until the aftershocks of the Intel bomb settle down, I suspect.
Although your point has some teeth, please realize that there is still a big performance disconnect between static and dynamic RAM.
Yeh, but are you going to get that performance out of this gadget? They benchmarked it as only 6x the speed of an actual hard drive.
Also, flash RAM tends to have limited write cycles.
I wouldn't want to use it for swap, and I'd mount noatime, but for a boot disk for a solid-state server this would be fine. People have been using flash drives for embedded systems, even embedded UNIX and Linux systems, for years.
For the Gigabyte product? I wouldn't use it for swap... I'd get better performance out of just using the same RAM chips directly on the motherboard. In fact just about anything you coudl use it for other than a solid-state boot disk you'd be better off just adding RAM.
Show me a version that doesn't take up valuable PCI real-estate, make it mountable in an internal 3.5" bay, and then I might be interested.
2 E16820160137&CMP=OTC-Froogle&ATT=Transcend+4GB+Com pact+Flash+(CF)+Card+Model+TS4GCF45
http://www.elx.com.au/item/CFIDE1
http://www.newegg.com/product/Product.asp?Item=N8
4GB, populated, probably fits in a 2.5" drive bay let alone 3.5". There's actually versions that have 2.5" connectors and mounting brackets so you can use them directly in a laptop, but I couldn't quickly google a price on them.
From the summary it sounded like a 2.5" or 3.5" 4GB IDE drive using flash instead of an IDE emulator and battery-backed up RAM using a PCI slot for power... and no memory included!
I'd pay $100 for 4GB of flash in a PCI or hard drive form factor, for a solid-state BSD or Linux webserver.
I don't think I'd pay $100 for a 0GB hard drive emulator that takes up both a PCI slot AND an SATA cable, and I still have to populate with RAM, and that will lose all its data if you leave it off too long.
Given that you can get a 2GB Compact Flash drive for $100 or 4GB for around $200 and you can hook those up to PATA with a $40 adapter, and populating this thing to 4GB will set me back more than that... I don't see the point.