AES256 encyption can be done at speeds of 48+ MB/s.
http://www.eskimo.com/~weidai/benchmarks.html This is on a Opteron 1.6 GHz, and it's a stock library. When I last looked, a Skype stream takes 5 kB/s, so 10 of those is 50 kB/s...
Wait until 2.16 so you can actually see some changes when you upgrade:) Point releases are hardly distinguishable from each other in the parts that matter.
It would have been even better if MySQL got it earlier (like PostgreSQL did from the start). It's a shame that MySQL started getting cool (and necessary!) features only now when the market pressure on its company increased. Prior to this, it was practically stagnating.
What MySQL *really* needs is a nice BSD-licensed connector/driver/client library compatible with existing solutions for popular languages such as PHP, Perl, Python, Java & others.
So, you don't just get to create 8 threads all of which can do arbitrary jobs, but you must explicitely split and divide the jobs and assigned them to specific processors (and do this from the thread that runs on the master processor). In a SMP system, the OS can move threads across processors freely, but not so here, and because of that programming would be harder than usual.
I think 1+7 CPUs deal like that would be ideal for huge batch processing jobs, such as encoding/decoding and mass calculations on similar data: just divide input into chunks, and process 7 or 8 chunks at a time. The need to constantly copy chunks of memory between individual processors would require either a very clever OS, or much hard work for the programmer. Also, it would probably be pessimal to make this kind of division of work: lets make cpus #1-#5 do monsters, cpu #6 process input, and cpu #7 sound, because at the end all processing results must be integrated into one single thing: the screen. Unless memory copies are remarkably fast and thread synchronisation routines very optimised, performance will suffer.
Note: I'm not an expert and don't know actual architecture of PS3 or Cell, so at best this is an educated guess.:) On the other hand, the "256k of onboard memory per processor" thing could also mean the author of the article is somehow confused about L2 cache:)
It looks like that people woting for that list were not big corporate users. In such environments Outlook is immensly popular, especially with management staff, because it's a nice integrated environment for everything from e-mail to group calendars, todo lists and similar organisational features. Of course, all this depends on Exchange servers.
I've heard several times that offices could switch to Linux, and even tolerate OpenOffice, but they simply cannot do without Outlook+Exchange.
Yes, there may be better solutions (such as using separate applications for e-mail and calendaring, possibly web applications) but none are as polished, easy to use and comprehensive in just the areas people like this need.
1. You may copy and distribute verbatim copies of the Program's
source code as you receive it, in any medium, provided that you
conspicuously and appropriately publish on each copy an appropriate
copyright notice and disclaimer of warranty; keep intact all the
notices that refer to this License and to the absence of any warranty;
and give any other recipients of the Program a copy of this License
along with the Program.
The DMCA makes it illegal to break copyright protection mechanisms. If the movie was NOT protected, and it probably wasn't, then the DMCA does not come into play in any way.
Can you quote paragraph against paragraph, side by side, why are those two licenses incompatible? Which paragraph in the GPL license prevents use of BSD licenseed code inside a GPL project?!
BSD License is strictly less demanding, and you CAN use any BSD library or project inside a GPL project without ANY problems with licenses. This is because BSD License doesn't prevent distribution of source code with a product, and GPL requires it. BSD License doesn't care about the software-is-free dogma thingy in the GPL, and if you want to apply it, so be it - it won't violate the BSDL. Think of it this way: GPL is kind of a superset of the BSDL - it adds things about which the BSDL just doesn't care.
If it were the other way around - using GPL code inside a BSD-licensed project would create major problems because either the BSD code would be automatically (virally) upgraded to the GPL or the GPL code would be stripped of some of its clauses (for example "you MUST distribute source code" semantics would be downgraded to "you MAY distribute source code if you wish so").
If it is, it's a good thing to have, though all that "commercial firm pitching a free version of their product into baseline kernel" thing sounds a bit dodgy.
I'm not an expert in anonymizing, but: to receive any information (really *any* - network packets, postal packages, etc.) efficiently, you have to have a unique address, and the party that sends the information must know it. Therefore, the path of that information can be tracked.
The only way I see to guarantee anonymous receiving is some kind of broadcast - for example as exists with satellite downloading systems: the information is always broadcast by the satellite to a really wide area, in which any party can receive it (and discard it if it's not meant for them to have it) without the abbility to detect who and where has received it. The reason this scheme works is because satellite receiving is a "read only" system - the receivers are passive and don't send information to satellites, they only filter the received content (i.e. channels or download streams). This could be useful with a public key encryption scheme.
(btw. the way satellite downloads currently works is that the receiver must have a separate "ground" line to a regular ISP that's used to send requests to the satellite company to broadcast the desired information, so there's still a traceable line, but in one direction only)
Oh not again about the "BSD is bad because MS can steal BSD-licensed code":((
Here it is again: THAT'S PRECISELY THE POINT! Don't you think people who put their code under the BSD license know it?
The license is liberal because:
maybe the software is made on an academic institution and it wouldn't be fair to restrict it (as the GPL does)
maybe the author writes the software for fame not for fortune (mostly closely tied with previous point - academics tend to make software that will increase their academic status [hint: BSD unix] and maybe land them consultant jobs)
maybe the author is a really good guy
maybe the author is really optimistic and thinks that if the companies can use quality BSD-licensed code they wouldn't have to develop their own crappy versions (hint: MS)
Slashdot Mirror
Sadly, no, not even to DragonflyBSD. Don't know why, but maybe it's because it uses kernel threads internally...
Also, there's "GEOM gate" on FreeBSD: http://garage.freebsd.pl/GEOM_Gate.pdf
For other cool stuff with GEOM see here and here. See also this discussion thread about ggate's limits.
AES256 encyption can be done at speeds of 48+ MB/s. http://www.eskimo.com/~weidai/benchmarks.html This is on a Opteron 1.6 GHz, and it's a stock library. When I last looked, a Skype stream takes 5 kB/s, so 10 of those is 50 kB/s...
Wait until 2.16 so you can actually see some changes when you upgrade :)
Point releases are hardly distinguishable from each other in the parts that matter.
It would have been even better if MySQL got it earlier (like PostgreSQL did from the start). It's a shame that MySQL started getting cool (and necessary!) features only now when the market pressure on its company increased. Prior to this, it was practically stagnating.
What MySQL *really* needs is a nice BSD-licensed connector/driver/client library compatible with existing solutions for popular languages such as PHP, Perl, Python, Java & others.
Immediate mode was not even available in DX until version 5.
So, you don't just get to create 8 threads all of which can do arbitrary jobs, but you must explicitely split and divide the jobs and assigned them to specific processors (and do this from the thread that runs on the master processor). In a SMP system, the OS can move threads across processors freely, but not so here, and because of that programming would be harder than usual.
I think 1+7 CPUs deal like that would be ideal for huge batch processing jobs, such as encoding/decoding and mass calculations on similar data: just divide input into chunks, and process 7 or 8 chunks at a time. The need to constantly copy chunks of memory between individual processors would require either a very clever OS, or much hard work for the programmer. Also, it would probably be pessimal to make this kind of division of work: lets make cpus #1-#5 do monsters, cpu #6 process input, and cpu #7 sound, because at the end all processing results must be integrated into one single thing: the screen. Unless memory copies are remarkably fast and thread synchronisation routines very optimised, performance will suffer.
Note: I'm not an expert and don't know actual architecture of PS3 or Cell, so at best this is an educated guess. :) On the other hand, the "256k of onboard memory per processor" thing could also mean the author of the article is somehow confused about L2 cache :)
I know about Evolution, but when I tried to use it, it was lacking features, slow and AFAIK only played nice with Novell's groupware server.
I've heard several times that offices could switch to Linux, and even tolerate OpenOffice, but they simply cannot do without Outlook+Exchange.
Yes, there may be better solutions (such as using separate applications for e-mail and calendaring, possibly web applications) but none are as polished, easy to use and comprehensive in just the areas people like this need.
I'm still waiting for a quote from the GPL where it says it's incompatible with the advertising clause...
http://www.gnu.org/philosophy/bsd.html
Though it's inconvenient, I don't see how the clause makes it strictly incompatible with GPL. Anyone know why it says so in http://www.gnu.org/philosophy/license-list.html ?
BSD License is strictly less demanding, and you CAN use any BSD library or project inside a GPL project without ANY problems with licenses. This is because BSD License doesn't prevent distribution of source code with a product, and GPL requires it. BSD License doesn't care about the software-is-free dogma thingy in the GPL, and if you want to apply it, so be it - it won't violate the BSDL. Think of it this way: GPL is kind of a superset of the BSDL - it adds things about which the BSDL just doesn't care.
If it were the other way around - using GPL code inside a BSD-licensed project would create major problems because either the BSD code would be automatically (virally) upgraded to the GPL or the GPL code would be stripped of some of its clauses (for example "you MUST distribute source code" semantics would be downgraded to "you MAY distribute source code if you wish so").
If it is, it's a good thing to have, though all that "commercial firm pitching a free version of their product into baseline kernel" thing sounds a bit dodgy.
I only used Iridium as an example of 2-way communication with a satellite system :)
Yes, it's possible - see http://en.wikipedia.org/wiki/Iridium_(satellite), but I don't think it's widespread.
Mod parent up :)
The only way I see to guarantee anonymous receiving is some kind of broadcast - for example as exists with satellite downloading systems: the information is always broadcast by the satellite to a really wide area, in which any party can receive it (and discard it if it's not meant for them to have it) without the abbility to detect who and where has received it. The reason this scheme works is because satellite receiving is a "read only" system - the receivers are passive and don't send information to satellites, they only filter the received content (i.e. channels or download streams). This could be useful with a public key encryption scheme.
(btw. the way satellite downloads currently works is that the receiver must have a separate "ground" line to a regular ISP that's used to send requests to the satellite company to broadcast the desired information, so there's still a traceable line, but in one direction only)
Here it is again: THAT'S PRECISELY THE POINT! Don't you think people who put their code under the BSD license know it?
The license is liberal because: