I'd love to see sales stats of those Walmart Lindows PCs. (i.e. a preinstalled version of Linux)
Most of these reviews seem to forget that Windows installations can be equally nightmarish to install Hell, it took me *two years* before I figured out how to get Win98 to run on an old system of mine. (I ran 95 until then. It turned out in the end that 98 needed a motherboard driver update to work with my SCSI controller, until then the system would hang on bootup.
No way is plugging your keyboard in via USB rather than PS/2 improving things. Do you know how many signal conversions a signal from a USB keyboard goes through? Then all of the driver complexity needed to read that info? The latency through a USB connection is far greater than that through a PS/2 connection (which is pretty "low-level" and goes through almost no signal conversions whatsoever before hitting the PC's bus.)
"this makes a big difference for people who type at speeds above 2000 words per minute."
Maybe it would make a difference if you're typing that fast, but for anyone typing less than 197 wpm (the current world record according to http://www.greyowltutor.com/essays/typing.html ), PS/2 has more than enough throughput, and has lower latency than USB.
Did you know that one clustered computing project used parallel ports for one of their interconnects? Yup. Ethernet handled the bandwidth-critical stuff, but for some information latency was the bottleneck and not throughput, and they used a custom parallel-based scheme for that. Faster doesn't always mean better for the job.
Also, you don't have to use the legacy ports when you don't want to. But when you have to use a $40 USB->serial converter for some serial device, you'll wish you'd spent the extra $5 for "the works" as far as ports on your motherboard.
$40 for a serial port or $5 for 2S 1P + PS/2
on
Legacy-Free PCs
·
· Score: 1
Which would you take?
Like you, I'd take "the works" for $5. USB->serial adapters with all control lines are $30-40 minimum, and even they won't even work with some serial devices.
By ditching legacy ports on a motherboard, you'll save a few bucks (less than $5 most likely, NatSemi SuperIO chips handle "the works" and are cheap - The most expensive one is $5.50 in 1k quantities and has serial, parallel, PS/2, 40 GPIO pins for blinkenlights/etc., and voltage/temperature monitoring.
Given that the above chip also handles non-legacy functions like system status monitoring, the cost of the parallel/serial/PS/2 ports is probably only $1-2 silicon-wise. $3 should cover the connectors if bought in quantity.
Intel's move to making FXCH non-free is the first sign that they're trying to phase out x87 while being compatible. Essentially, that FPU is only there for compatibility reasons, and for modern apps Intel doesn't want you to use it - They want you to use SSE. (Note: You can use SSE without vectorizing your code. In the past this would not be of any benefit, but with the crippled x87 FPU, SSE is the way to go even when non-vectorized.)
As to other ports: I say leave em' on there. Due to the massive volumes involved, they only increase system cost by $5 or less for "the works" (2S 1P, both PS/2).
That means I can use my old AT&T keyboard (IBM Model M clone), new keyboards just don't "feel" right. Yes, there are PS/2-USB adapters, but they cost $20-30. $20-30 for one port when I could've gotten "the works" for $5 or less? I don't think so.
Same for my serial devices. Adapters that implement all RS232 control lines cost $30-40 minimum. I can get two onboard 16550s for a fraction of that price.
USB is nice, but for some operations it's massive overkill, and for those operations I want my nice cheap 16550s!
Without an RS232 port, 90% of Atmel's AVR microcontroller line can no longer be used for various PC-interfacing projects I have. The remaining AVRs are much more expensive, harder to obtain, and a bitch to work with.
Amen. I'll take a legacy RS232 port any day
on
Legacy-Free PCs
·
· Score: 1
I have a decent amount of hardware that just simply Doesn't Work with most USB->Serial adapters. I have some that won't even work with ones that implement all control lines due to the nature of how they use the port. (i.e. directly twiddling the control lines in oddball ways. Think timing-critical devices like IR remote receivers.)
I'd rather pay $2 extra in the price of my machine for a proper 16550-compatible RS232 port that will work with anything designed for a serial port than $40 for a USB->Serial adapter that doesn't even work with half the serial devices I own. (Which is why I'm glad my Dell Inspiron 8200 has the damned port.)
And as an EE major - I will be very sad when parallel ports disappear completely, they're wonderful for misc. PC-controlled electronics projects. Again, something USB can do but not without a large investment in money and time.
It's not half the calories. It's 1/100 or less. Aspartame (NutraSweet) has approx. the same calories per unit weight as sugar, but is 100-200 times sweeter per unit weight. (It's THAT strong. NutraSweet sugar replacements are 99% inert powder.)
One would have to drink a few liters of diet soda to even reach 10 calories.
I don't like Guinness. I like stouts, but not that particular one.
Most U.S. beers are horrendous, but it all changes when you discover the good microbrews. There's a brewery near where I live in NJ that specializes in Belgian-style wheat beers and does a damn good job of it. Wagner Valley in Lodi, NY (Same owner and building as Wagner's winery) and Ithaca Beer Co. in Ithaca, NY are also excellent. Wagner's doppelbock and Ithaca's seasonal stout are my two favorite beers.
Nothing preventing some of the watercooling system from being outside of the case. I'm sure you could come up with some unique looking reservoirs, for example. And you can dress up the radiator to look cool. (Acrylic housing, UV-sensitive glowing fan, etc etc.)
Except for the strength issue (don't use balsa...), all of those issues are easily solvable by gluing copper or aluminum foil on the inside of each wood panel, or some sort of wire mesh (window screening will work...)
I would take the nice look of a wood case over a light one any day for my main desktop system. (A LAN party system would be a different story...) With the right wood, a wood case can be MUCH stronger than a metal one.
Unfortunately, the fidelity of DVDs isn't all that hot. 720x480, approx. 2.5 Mbps MPEG-2 MP@ML.
The highest fidelity video you can get at home right now is 1080i (1920x1080 interlaced) ATSC HDTV or JVC D-Theater. ATSC HD streams are MPEG-2 MP@HL (Main Profile at High Level) streams at 19.2 Mbps, D-Theater is at around 25-28 Mbps.
At a given bitrate, MPEG-4 (WM9 and DivX/XviD are both MPEG-4 variants/derivatives) will do significantly better than MPEG-2. I'm sure that with some hardware tweaking, WM9 can do 1080p (1920x1080 progressive) video at a high bitrate. The codec supports it, unfortunately that high of a resolution will strain even the latest CPUs. That's the same resolution that I believe AoTC was filmed at. (I believe AoTC was delivered to digital theaters as a high-bitrate MPEG-2 stream.)
In short, WM9 does not mean that Landmark can't provide fidelity beyond what you can obtain at home. In fact, chances are that they will provide fidelity beyond the best available for home use, JVC D-Theater. (Or not. Landmark specializes in independent films, and the independent film industry makes heavy use of DV video these days, which itself is only 720x480, albeit VERY lightly compressed in its source form. But most likely Landmark will be able to provide higher fidelity than what you can get at home, with the potential for much more with high-quality source material.)
MPEG-4 is pretty flexible. By changing various codec parameters, it can do extremely well at high bitrates. (Or if you don't want to manually tweak codec parameters, the optimum method is two-pass encoding.)
MPEG-4 can do anything MPEG-2 can and more. At a sufficiently high bitrate, you'll be hard pressed to tell it apart from whatever codec was used to encode the theatrical releases of Star Wars Episode 2. (I've read that it was filmed in 1080p, and most likely it was encoded using high-bitrate MPEG-2 MP@HL - Main Profile at High Level, similar to HDTV. DVDs are MP@ML, Main Profile at Main Level.)
While I think WM9 (A proprietary version of MPEG-4) is a bad idea when XviD/DivX (much more open implementations) are nearly as good, I think you're being unfairly harsh on WM9 here.
A video is only as good as its source. Unfortunately, 90% of the time WM9 is used because it excels at low bitrates.
What most people don't know is that WM9 (and DivX/Xvid) also do incredibly well at high bitrates. Try to find some WM9 clips encoded from HDTV sources, or get an HDTV tuner card and encode some clips yourself. (Or use DivX like I do.) You'll change your opinion of MPEG-4 very quickly.
At a given bitrate, MPEG-4 is better than MPEG-2. Unfortunately it takes more CPU to encode and decode. (There are VERY few hardware MPEG-4 decoders and so far I have not seen any hardware MPEG-4 encoders).
DVD will look like crap on the big screen, it's only 720x480. HDTV MIGHT be acceptable, the ATSC standard includes resolutions up to 1920x1080 interlaced (Actually, I'm not sure but I believe that it may also include 30 fps 1080p, but no one broadcasts that at the moment.) HDTV is encoded at 19.2 Mbps max (approx. 8.5 gigabytes/hour, as opposed to approx. 2.5 or so for DVD). "studio master" HD (used by the networks for distribution before final retransmission) is around 30-40 Mbps or more, and JVC "D-Theater", only playable on JVC D-VHS decks, is around 25-28 Mbps. All of these formats are MPEG-2. Encode using MPEG-4 (WM9, DivX, etc.) at those bitrates and the quality will be even better.
WM9 and high-bitrate DivX are popular formats for backing up HDTV recordings that will fit on a DVD. (Lots of info on that at http://www.avsforum.com/ in the HTPC section.)
DVD in the big screen is a Bad Idea. DVD has a max 720x480 resolution.
That said, it's possible to have high-bitrate (but still compressed) video that looks good on the big screen.
Try watching some 1080i HDTV content. (Like CSI or CSI: Miami). This will look pretty good on the big screen. Especially if you use a "studio master" bitstream, which is often at 40+ megabits/sec as opposed to the 19.2 Mbps of ATSC HD.
Now take that further, and use a better codec, like MPEG-4. With MPEG-4 at 1080p resolution and a high bitrate, you can still have a high compression ratio but have it look excellent on the big screen.
Probably MPEG-4 encoded at HDTV bitrates (19.2 Mbps) would be indistinguishable from pure film.
(BTW, there is already a looming format war over "high definition" DVDs, as HDTV users are beginning to realize that DVD isn't all it's cracked up to be. The two main competing techniques are standard DVD media but with MPEG-4 encoding, and Blu-Ray with MPEG2. There is also DVHS, which supports MPEG-2 at up to 25+ megabits/sec.)
This post needs a +5... BitTorrent is one of the ultimate cures to the Slashdot Effect... Just look at how well it's been holding up with the RedHat9 distribution.
I believe that if your firewall is closed up, then your client can actively connect to another to start an upload. But this requires that the other endpoint be open. (i.e. one endpoint or the other must be open, it can be either.)
Opening up a few firewall ports is the difference between not being able to connect to any other clients running a firewall whether for upload or download, or being able to connect to any client. Makes more of a difference for "small" torrents where there are fewer clients.
Yesterday even without opening up ports 6881-6889 on my firewall, I got download rates of around 270KB/sec. Fastest RedHat download I've ever done, especially on a cable modem.
If that's anything like the D-Link DWL-520 (Prism2 based PCI adapter, wouldn't be surprised if it's nearly identical.), then I'm not surprised you're having problems.
Linux does support the Prism2 PCIs, don't know if the default RHL kernel does. I used linux-wlan-ng, which has the best Prism2 support bar-none.
But the PCI Prism2s (like the DWL-520) are pieces of shit. Horrendous packet loss under Linux, very unreliable. Don't blame it on Linux though - At least it worked under Linux! While it had 25% packet loss despite a good signal under Linux, under Windows it was utterly incapable of communicating at all.
Needless to say, I returned that POS card. I had better luck with an Orinoco in a PCI adapter (Worked perfectly under Linux, unfortunately massive problems under Windows. Windows doesn't have the equivalent of the pci=biosirq setting that the Linux PCMCIA drivers have.)
In the end, stay away from PCI WLAN adapters. Get a WET11 and hook it to a nice reliable Ethernet card.
PCMCIA adapters in systems designed with PCMCIA bridges work fine - I can use any WLAN card on the market except the new TI ACX100 (D-Link AirPlus) and Broadcom chipsets (Very rare, only card I know of is the Dell TrueMobile 1180 - The 1150 is an Orinoco).
" Yes, future code could be closed but already GPL'ed versions would remain GPL'ed. So would any derivative works of a GPL licensee." Note that future code can only be closed IF the authors have been assigned copyright rights to all portions of the code or all contributors have agreed to the closing.
i.e. if there's one line of code that hasn't had its copyright assigned and the contributor hasn't given permission, then the code can't go GPL (Until that line is removed, a la BSD and the AT&T copyrights.)
I've been involved with a project that did this - We originally released under GPL, and for various reasons it was decided to create a closed fork. Fortunately, we only had 2 contributions from outside of our research lab, and the authors of both of those agreed to going commercial. (In fact, I think one is getting money out of the deal.)
FYI, the project is at http://www.sourceforge.net/projects/cu30/
Slashdot Mirror
I'd love to see sales stats of those Walmart Lindows PCs. (i.e. a preinstalled version of Linux)
Most of these reviews seem to forget that Windows installations can be equally nightmarish to install Hell, it took me *two years* before I figured out how to get Win98 to run on an old system of mine. (I ran 95 until then. It turned out in the end that 98 needed a motherboard driver update to work with my SCSI controller, until then the system would hang on bootup.
Can you say placebo effect?
No way is plugging your keyboard in via USB rather than PS/2 improving things. Do you know how many signal conversions a signal from a USB keyboard goes through? Then all of the driver complexity needed to read that info? The latency through a USB connection is far greater than that through a PS/2 connection (which is pretty "low-level" and goes through almost no signal conversions whatsoever before hitting the PC's bus.)
"this makes a big difference for people who type at speeds above 2000 words per minute."
Maybe it would make a difference if you're typing that fast, but for anyone typing less than 197 wpm (the current world record according to http://www.greyowltutor.com/essays/typing.html ), PS/2 has more than enough throughput, and has lower latency than USB.
Did you know that one clustered computing project used parallel ports for one of their interconnects? Yup. Ethernet handled the bandwidth-critical stuff, but for some information latency was the bottleneck and not throughput, and they used a custom parallel-based scheme for that. Faster doesn't always mean better for the job.
Also, you don't have to use the legacy ports when you don't want to. But when you have to use a $40 USB->serial converter for some serial device, you'll wish you'd spent the extra $5 for "the works" as far as ports on your motherboard.
Which would you take?
Like you, I'd take "the works" for $5. USB->serial adapters with all control lines are $30-40 minimum, and even they won't even work with some serial devices.
By ditching legacy ports on a motherboard, you'll save a few bucks (less than $5 most likely, NatSemi SuperIO chips handle "the works" and are cheap - The most expensive one is $5.50 in 1k quantities and has serial, parallel, PS/2, 40 GPIO pins for blinkenlights/etc., and voltage/temperature monitoring.
Given that the above chip also handles non-legacy functions like system status monitoring, the cost of the parallel/serial/PS/2 ports is probably only $1-2 silicon-wise. $3 should cover the connectors if bought in quantity.
Intel's move to making FXCH non-free is the first sign that they're trying to phase out x87 while being compatible. Essentially, that FPU is only there for compatibility reasons, and for modern apps Intel doesn't want you to use it - They want you to use SSE. (Note: You can use SSE without vectorizing your code. In the past this would not be of any benefit, but with the crippled x87 FPU, SSE is the way to go even when non-vectorized.)
As to other ports: I say leave em' on there. Due to the massive volumes involved, they only increase system cost by $5 or less for "the works" (2S 1P, both PS/2).
That means I can use my old AT&T keyboard (IBM Model M clone), new keyboards just don't "feel" right. Yes, there are PS/2-USB adapters, but they cost $20-30. $20-30 for one port when I could've gotten "the works" for $5 or less? I don't think so.
Same for my serial devices. Adapters that implement all RS232 control lines cost $30-40 minimum. I can get two onboard 16550s for a fraction of that price.
USB is nice, but for some operations it's massive overkill, and for those operations I want my nice cheap 16550s!
Without an RS232 port, 90% of Atmel's AVR microcontroller line can no longer be used for various PC-interfacing projects I have. The remaining AVRs are much more expensive, harder to obtain, and a bitch to work with.
I have a decent amount of hardware that just simply Doesn't Work with most USB->Serial adapters. I have some that won't even work with ones that implement all control lines due to the nature of how they use the port. (i.e. directly twiddling the control lines in oddball ways. Think timing-critical devices like IR remote receivers.)
I'd rather pay $2 extra in the price of my machine for a proper 16550-compatible RS232 port that will work with anything designed for a serial port than $40 for a USB->Serial adapter that doesn't even work with half the serial devices I own. (Which is why I'm glad my Dell Inspiron 8200 has the damned port.)
And as an EE major - I will be very sad when parallel ports disappear completely, they're wonderful for misc. PC-controlled electronics projects. Again, something USB can do but not without a large investment in money and time.
It's not half the calories. It's 1/100 or less. Aspartame (NutraSweet) has approx. the same calories per unit weight as sugar, but is 100-200 times sweeter per unit weight. (It's THAT strong. NutraSweet sugar replacements are 99% inert powder.)
One would have to drink a few liters of diet soda to even reach 10 calories.
Calorie-wise, diet soda = water.
Yes, I like the Belgian beers too. Quite good.
I don't like Guinness. I like stouts, but not that particular one.
Most U.S. beers are horrendous, but it all changes when you discover the good microbrews. There's a brewery near where I live in NJ that specializes in Belgian-style wheat beers and does a damn good job of it. Wagner Valley in Lodi, NY (Same owner and building as Wagner's winery) and Ithaca Beer Co. in Ithaca, NY are also excellent. Wagner's doppelbock and Ithaca's seasonal stout are my two favorite beers.
Nothing preventing some of the watercooling system from being outside of the case. I'm sure you could come up with some unique looking reservoirs, for example. And you can dress up the radiator to look cool. (Acrylic housing, UV-sensitive glowing fan, etc etc.)
Except for the strength issue (don't use balsa...), all of those issues are easily solvable by gluing copper or aluminum foil on the inside of each wood panel, or some sort of wire mesh (window screening will work...)
I would take the nice look of a wood case over a light one any day for my main desktop system. (A LAN party system would be a different story...) With the right wood, a wood case can be MUCH stronger than a metal one.
Yeah... Going proprietary is a bad idea. But video quality is one place where WM9 is not lacking and it shouldn't be bashed in that arena.
Anyone going to be "torrenting" this one?
:)
I've been thinking of trying FreeBSD, and I definately will grab it if it's torrented.
Unfortunately, the fidelity of DVDs isn't all that hot. 720x480, approx. 2.5 Mbps MPEG-2 MP@ML.
The highest fidelity video you can get at home right now is 1080i (1920x1080 interlaced) ATSC HDTV or JVC D-Theater. ATSC HD streams are MPEG-2 MP@HL (Main Profile at High Level) streams at 19.2 Mbps, D-Theater is at around 25-28 Mbps.
At a given bitrate, MPEG-4 (WM9 and DivX/XviD are both MPEG-4 variants/derivatives) will do significantly better than MPEG-2. I'm sure that with some hardware tweaking, WM9 can do 1080p (1920x1080 progressive) video at a high bitrate. The codec supports it, unfortunately that high of a resolution will strain even the latest CPUs. That's the same resolution that I believe AoTC was filmed at. (I believe AoTC was delivered to digital theaters as a high-bitrate MPEG-2 stream.)
In short, WM9 does not mean that Landmark can't provide fidelity beyond what you can obtain at home. In fact, chances are that they will provide fidelity beyond the best available for home use, JVC D-Theater. (Or not. Landmark specializes in independent films, and the independent film industry makes heavy use of DV video these days, which itself is only 720x480, albeit VERY lightly compressed in its source form. But most likely Landmark will be able to provide higher fidelity than what you can get at home, with the potential for much more with high-quality source material.)
MPEG-4 is pretty flexible. By changing various codec parameters, it can do extremely well at high bitrates. (Or if you don't want to manually tweak codec parameters, the optimum method is two-pass encoding.)
MPEG-4 can do anything MPEG-2 can and more. At a sufficiently high bitrate, you'll be hard pressed to tell it apart from whatever codec was used to encode the theatrical releases of Star Wars Episode 2. (I've read that it was filmed in 1080p, and most likely it was encoded using high-bitrate MPEG-2 MP@HL - Main Profile at High Level, similar to HDTV. DVDs are MP@ML, Main Profile at Main Level.)
Those numbers are the number of channels, not version numbers.
5.1 indicates 5.1 channels of sound. 5 full-frequency-range channels of sound (Front left/right/center, rear left/right) and one low-frequency-only (subwoofer) sound channel.
7.1 indicates an extra two channels. Maybe front left/right/center, middle left/right, rear left/right plus subwoofer.
While I think WM9 (A proprietary version of MPEG-4) is a bad idea when XviD/DivX (much more open implementations) are nearly as good, I think you're being unfairly harsh on WM9 here.
A video is only as good as its source. Unfortunately, 90% of the time WM9 is used because it excels at low bitrates.
What most people don't know is that WM9 (and DivX/Xvid) also do incredibly well at high bitrates. Try to find some WM9 clips encoded from HDTV sources, or get an HDTV tuner card and encode some clips yourself. (Or use DivX like I do.) You'll change your opinion of MPEG-4 very quickly.
Remember, there are multiple variants of MPEG.
VCD and DVD use MPEG-2.
WM9 is an MPEG-4 variant, as is DivX/XviD.
At a given bitrate, MPEG-4 is better than MPEG-2. Unfortunately it takes more CPU to encode and decode. (There are VERY few hardware MPEG-4 decoders and so far I have not seen any hardware MPEG-4 encoders).
DVD will look like crap on the big screen, it's only 720x480. HDTV MIGHT be acceptable, the ATSC standard includes resolutions up to 1920x1080 interlaced (Actually, I'm not sure but I believe that it may also include 30 fps 1080p, but no one broadcasts that at the moment.) HDTV is encoded at 19.2 Mbps max (approx. 8.5 gigabytes/hour, as opposed to approx. 2.5 or so for DVD). "studio master" HD (used by the networks for distribution before final retransmission) is around 30-40 Mbps or more, and JVC "D-Theater", only playable on JVC D-VHS decks, is around 25-28 Mbps. All of these formats are MPEG-2. Encode using MPEG-4 (WM9, DivX, etc.) at those bitrates and the quality will be even better.
WM9 and high-bitrate DivX are popular formats for backing up HDTV recordings that will fit on a DVD. (Lots of info on that at http://www.avsforum.com/ in the HTPC section.)
DVD in the big screen is a Bad Idea. DVD has a max 720x480 resolution.
That said, it's possible to have high-bitrate (but still compressed) video that looks good on the big screen.
Try watching some 1080i HDTV content. (Like CSI or CSI: Miami). This will look pretty good on the big screen. Especially if you use a "studio master" bitstream, which is often at 40+ megabits/sec as opposed to the 19.2 Mbps of ATSC HD.
Now take that further, and use a better codec, like MPEG-4. With MPEG-4 at 1080p resolution and a high bitrate, you can still have a high compression ratio but have it look excellent on the big screen.
Probably MPEG-4 encoded at HDTV bitrates (19.2 Mbps) would be indistinguishable from pure film.
(BTW, there is already a looming format war over "high definition" DVDs, as HDTV users are beginning to realize that DVD isn't all it's cracked up to be. The two main competing techniques are standard DVD media but with MPEG-4 encoding, and Blu-Ray with MPEG2. There is also DVHS, which supports MPEG-2 at up to 25+ megabits/sec.)
When I need them?
This post needs a +5... BitTorrent is one of the ultimate cures to the Slashdot Effect... Just look at how well it's been holding up with the RedHat9 distribution.
Is the best approach.
Spark plugs are a proven, reliable, and mature technology.
I believe that if your firewall is closed up, then your client can actively connect to another to start an upload. But this requires that the other endpoint be open. (i.e. one endpoint or the other must be open, it can be either.)
Opening up a few firewall ports is the difference between not being able to connect to any other clients running a firewall whether for upload or download, or being able to connect to any client. Makes more of a difference for "small" torrents where there are fewer clients.
You probably have some sort of technical issues.
Yesterday even without opening up ports 6881-6889 on my firewall, I got download rates of around 270KB/sec. Fastest RedHat download I've ever done, especially on a cable modem.
Obviously someone didn't read the /. article about RH9's release and BitTorrent. :)
I downloaded all three binary ISOs in about 2-3 hours. Not sure, had it in the background with screen. Also leeched MDK9.1 in about the same time.
I'm waiting for the source ISOs to become available on BT so I can get them and burn everything onto a DVD.
If that's anything like the D-Link DWL-520 (Prism2 based PCI adapter, wouldn't be surprised if it's nearly identical.), then I'm not surprised you're having problems.
Linux does support the Prism2 PCIs, don't know if the default RHL kernel does. I used linux-wlan-ng, which has the best Prism2 support bar-none.
But the PCI Prism2s (like the DWL-520) are pieces of shit. Horrendous packet loss under Linux, very unreliable. Don't blame it on Linux though - At least it worked under Linux! While it had 25% packet loss despite a good signal under Linux, under Windows it was utterly incapable of communicating at all.
Needless to say, I returned that POS card. I had better luck with an Orinoco in a PCI adapter (Worked perfectly under Linux, unfortunately massive problems under Windows. Windows doesn't have the equivalent of the pci=biosirq setting that the Linux PCMCIA drivers have.)
In the end, stay away from PCI WLAN adapters. Get a WET11 and hook it to a nice reliable Ethernet card.
PCMCIA adapters in systems designed with PCMCIA bridges work fine - I can use any WLAN card on the market except the new TI ACX100 (D-Link AirPlus) and Broadcom chipsets (Very rare, only card I know of is the Dell TrueMobile 1180 - The 1150 is an Orinoco).
" Yes, future code could be closed but already GPL'ed versions would remain GPL'ed. So would any derivative works of a GPL licensee."
Note that future code can only be closed IF the authors have been assigned copyright rights to all portions of the code or all contributors have agreed to the closing.
i.e. if there's one line of code that hasn't had its copyright assigned and the contributor hasn't given permission, then the code can't go GPL (Until that line is removed, a la BSD and the AT&T copyrights.)
I've been involved with a project that did this - We originally released under GPL, and for various reasons it was decided to create a closed fork. Fortunately, we only had 2 contributions from outside of our research lab, and the authors of both of those agreed to going commercial. (In fact, I think one is getting money out of the deal.)
FYI, the project is at http://www.sourceforge.net/projects/cu30/
IIRC, that Sony drive is 4x capable.
:)
:)
Go to www.rima.com and buy some 4x discs. $2 each for Ritek G04s in a 25-pack, $1.80 each in a 100-pack.
Burn in 15 minutes.
(I have a Pioneer DVR-105... These RH9 CDs are going on a DVD!