Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
Get HideMyAss! VPN, PC Mag's Top 10 VPNs of 2016 for 55% off for a Limited Time ×

Comment Re:If Windows is so bad, why use it? (Score 1) 328

No one compiles their own kernel these days.

That's not true. I still compile custom kernels and I know a few others that do so too. However to do so is a choice these days and you better have a very good reason to do so because distribution provided kernels are perfect for 99% of the use scenarios. I have a very good reason: I think it's fun. I'm weird that way.
Plus I create custom hardware. But many times even that is not really a reason for compiling a custom kernel with the user space driver options.

Submission + - Dutch minister refuses to stop using private mail for government business (www.nu.nl)

Melkman writes: Despite being victim of phishing the Dutch minister of economic affairs Henk Kamp will not stop using his private mail for government business. Even after being warned that this is against regulations Kamp said he will continue with this practice because "It's just easier for me, and that's the way it is". Aside from being insecure the messages in private accounts are exempt from WOB requests, the Dutch equivalent of FOIA.

Comment Not everything (Score 1) 460

When an friend an I got started with Linux he wanted to remove his Slackware install from a dual boot PC. For fun he ran rm -rf / on that install. We had a good laugh when the message scrolled by of the OS trying and failing to remove files from the CDROM. That was until he realized that he had mounted his Windows partition too. It didn't fail to remove files there :-)
 

Comment Re:Wrong choice (Score 0) 282

Sigh, the reason Juniper etc use BSD is because of the licensing, not because it's better at networking.
You can take BSD, modify everything and not have to share those modifications with anyone. So it's a great base to start building your own proprietary system. The real networking stuff in machines made by the vendors you name isn't done by the normal FreeBSD kernel but by software custom written by those manufacturers. Of course they will not back port this software to FreeBSD proper as it's the product they sell. If real heavy networking is going on its done by ASICs like Junipers Trio chipset.
If Microsoft doesn't have the intention to distribute their Linux version the licensing isn't a big deal. And there may be reasons why Linux is a better fit in their circumstance like the much more expansive hardware support.
Maybe you've heard of Arista ? Their kit runs on Linux. They must be stupid.
 

Comment Re:Cannot scale anyway (Score 4, Informative) 399

The only -current- viable source of tritium is fission. However fusion can produce its own tritium in breeder blankets. This is one of the concepts that will be researched in ITER: https://www.iter.org/mach/trit...

So the last part of your post "but it's not a viable power source unless the need for tritum is eliminated" is just wrong.

Comment Re:Why IPv6 is broken (Score 1) 595

Easy: Just add prefixes to the numbers and everybody is happy. The old numbers stay valid, you can still connect within the old network(s), nobody has to remember new numbers.

You have no knowledge of IP have you ? To follow your telephony analogy: an "IPv4 telephone" can only dial numbers with exactly 10 digits. If you are going to expand the address space by adding a digit you will have to change all those "IPv4 phones".

Without the analogy: IPv4 addresses are 32 bit and every IPv4 stack defines them as 32 bit numbers. To address more than 2^32 nodes you will have to adapt every IPv4 stack and redefine addresses as something bigger, say 2^128. And that is exactly what IPv6 does.

When you redefine your address you will get incompatibility. A node which still has IP addresses defined as 32 bit will not be able to send replies to a node with a 2^128 address. The destination address simply does not fit in the defined address space. So while a node with an updated stack might be able to send traffic to a node without an updated stack that last one cannot send data back and you won't have meaningful communication.

That being the case it is better to make it clear that an expanded address space is incompatible with the current stacks. And that is exactly what has been done with IPv6.

Calling people morons without have any significant knowledge about the problem domain yourself is the real stupidity in your post. Dunningâ"Kruger in full effect

Comment Nice resolution (Score 2, Interesting) 96

But remembering interviews with Occulus developers there is more to VR than a good resolution and tracking. Things like ridiculous low latency needed to prevent motion sickness and screen artifacts caused by rapid panning. Has Valve solved these things in record time in secret or will this be a better specs on paper but worse in practice product ? Or maybe I'm just falling for Oculus marketing: http://arstechnica.com/gaming/...

Comment Re:Hope it has GigE. (Score 1) 180

There is no 100MHz ethernet. There is 100Mbps ethernet that runs at 125MHz on copper with 4B/5B encoding (and MLT-3 to limit the bandwidth of the signal). The overhead is in ethernet and IP headers, preamble and interpacket gaps etc. Although the theoretical maximum efficiency for TCP on IP over ethernet with an IP MTU of 1500 is about 95% this is almost never reached due to latencies and suboptimal implementations etc. It's quite normal to get 90% maximum throughput in optimal situations.

Slashdot Top Deals

"Consistency requires you to be as ignorant today as you were a year ago." -- Bernard Berenson

Working...