Besides the fact that Linux can access more than 4GB of RAM, this thing isn't a single server; it's a cluster. If each of the nodes has 4GB of RAM, that adds up to a lot.
I looked up how much IPv6 addresses cost (at least in North America), and it's $2,500/year for a/35. Since each customer is supposed to get a/48, that's enough room for 8192 customers. That works out to about 30 cents per customer per year.
This is called Universal Plug-n-Play; it's already supported by Windows XP and the spec will come out in a few months. Makers of cheap NAT boxes have already pledged their support.
It looks like Linux supports camcorders, storage devices, and even FireWire networking. (Apple invented FireWire, yet they don't even support FireWire networking!) I'm sure there are bugs, but in general I can't think of anything missing.
Sure, but you lose flexibility. When your hard-wired RC4/WEP chip turns out to have a huge flaw, you can't fix it. I'm not really opposed to putting it in hardware, but I think it would be good if NICs allowed certain hardware features to be turned off and done in software.
[P]eople in the know... have repeatedly said that an open source version of BeOS will basically never happen. The system depends on licensed code that Be apparently couldn't give away even if they wanted to.
Gee, this sounds familiar...
People in the know have repeatedly said that an open source version of Netscape will basically never happen. The system depends on licensed code that Netscape apparently couldn't give away even if they wanted to.
Why do people need portable addresses (unless they are multihomed)? Are there routing protocols that could handle a Net where everyone used portable addresses?
Slashdot Mirror
Palm has announced that they are moving from 68K to ARM, so PowerPC code won't be that useful. Luckily BeOS is portable.
Neither 1394 nor USB 2.0 have enough bandwidth to drive a monitor.
How much horsepower does each of those virtual servers get? It can't be that much. $500/server would be too much if it was only 100 MIPS.
Besides the fact that Linux can access more than 4GB of RAM, this thing isn't a single server; it's a cluster. If each of the nodes has 4GB of RAM, that adds up to a lot.
It's a cluster, not a server. It's hard to tell how this is that different from a rack full of 1U servers, but I didn't read their Web site carefully.
A non-anti-aliased, hand-hinted font is also patented.
I don't think so. Apple has patents on one method of hinting, but other methods (e.g. plain old bitmaps) are not patented.
Why does this use LD_PRELOAD? Why not just patch GDK directly? Heck, why hasn't Xft support been integrated into a released version of GDK yet?
Use 6to4; it's more efficient than tunnel-based systems like 6bone and freenet6.
Instructions: BSD, Debian, Windows.
I looked up how much IPv6 addresses cost (at least in North America), and it's $2,500/year for a /35. Since each customer is supposed to get a /48, that's enough room for 8192 customers. That works out to about 30 cents per customer per year.
I'm guessing MS can buy all the IPv4 addresses they need. Why should they move to IPv6?
This is called Universal Plug-n-Play; it's already supported by Windows XP and the spec will come out in a few months. Makers of cheap NAT boxes have already pledged their support.
Component video is analog.
First HavenCo, then this; what will they think of next?
SPEC rejected those benchmarks because Intel used a special proprietery compiler with the tests and not a normal compiler a developer would use.
Do you have evidence of this? I see plenty of SPEC CPU benchmarks using Intel compilers.
Target disk mode is completely different from FireWire networking.
Future DVD players will be able to have FireWire outputs as long as they use DTCP.
It looks like Linux supports camcorders, storage devices, and even FireWire networking. (Apple invented FireWire, yet they don't even support FireWire networking!) I'm sure there are bugs, but in general I can't think of anything missing.
That doesn't seem right; the other day I did a file transfer at 4.8 Mbps over an Aironet card with 128-bit WEP enabled.
Sure, but you lose flexibility. When your hard-wired RC4/WEP chip turns out to have a huge flaw, you can't fix it. I'm not really opposed to putting it in hardware, but I think it would be good if NICs allowed certain hardware features to be turned off and done in software.
Doing RC4 or AES at 11 Mbps in software is no problem.
[P]eople in the know ... have repeatedly said that an open source version of BeOS will basically never happen. The system depends on licensed code that Be apparently couldn't give away even if they wanted to.
Gee, this sounds familiar...
People in the know have repeatedly said that an open source version of Netscape will basically never happen. The system depends on licensed code that Netscape apparently couldn't give away even if they wanted to.
Sun's Grid Engine doesn't seem nearly as powerful as the Globus toolkit used by the Grid.
Why do people need portable addresses (unless they are multihomed)? Are there routing protocols that could handle a Net where everyone used portable addresses?
AFAIK, scientists aren't asking for their work to be copyright-free, just available without paying.
I haven't heard anything like that. Does the XP kernel run on MIPS, ARM, SH, and PowerPC? Is XP Embedded ROMable?