Very insightful. I have a few "starving musician" friends who'd appreciate a bigger audience. And lately, even the CDs I've been buying have been
imports, and small labels (like Century Media).
Then again, I am following the sheep to this
summer's Ozzfest. But with a re-united Judas Priest, I couldn't resist.
Oh, well, I guess, small steps.
Oh! I lovedSilent Running. Huey,
Dewey, and Louie (the robots) were indeed cute. But
I thought the implied environmental disaster behind
the story was pretty scary.
And I enjoy Peter Schickele. Both PDQ Bach, and
his "Schickele Mix" program which I used to listen
to on NPR (not sure if it's still on, but I'm not
usually up that early on a weekend any more).
I don't know how low-level of network access Java
allows
Java has two socket classes: Socket (a TCP socket),
and DatagramSocket (a UDP socket)... and a subclass
of DatagramSocket for multicasting. No raw sockets
available (well, in theory you could implement one
in native code, and invoke it via RMI, but then it
wouldn't be available from an applet due to security
restrictions).
The kernel is so far diverged from the main thread of Linux that it's a dead-end
Do you think that well be less true of Fedora? It
sounds like Core 2 will have a 2.6 kernel,
apparently ahead of any of the Red Hat releases.
Has anyone compared the Core 2 Test 2 code to the
vanilla 2.6 kernel? Or perhaps it will start off
close, then instead of following new vanilla releases, just "backport" fixes to the kernel version they used for that release?
I bet you could build a low-speed, high-torque paddle wheel (it would need to have a lot of surface
area being pushed on by the river). Then, using gear
ratios, you can convert that to high-speed,
low-torque that may be needed by your generator.
Not being a mechanical engineer, I'll leave it at
that.:-)
Someone disparages my education by telling me I
should "rad" more? What should I "rad"? "Boks"?:-)
That may (or may not) have been the usage of those
terms in early 19th century France, but not in the
20th and 21st century United States.
I think the term you are looking for is "authoritarian", or "dictatorship". Or just plain
old "revolutionary". And need I remind you of a
reasonably successful revolution which took place
in the late 1770's and early 1780's?
Going back to the Nazis, while I suppose you could
call the "Beerhall Putsch" a mini-revolution, it
was unsuccessful, and it wasn't until the Nazis
had secured more power through elections that they
were able to complete their rise to complete
dictatorship.
Soon thereafter, the "Night of the Long Knives" removed any pretense of "socialism" in the National Socialist party.
The Nazis were anti-communism as their invasion
of Russia in 1941 showed.
Bringing this back to our country, the group who
I think inherits the idealogical heritage of the
Nazis, the conservative Christians, is also in
favor of having the government regulate citizens'
private lives, and is very anti-homosexual (brings
back images of the Nazis placing pink triangles
on gay men, just as they put yellow stars on
Jews).
I think you'd have trouble convincing most people
on earth that the Nazis were leftists. You
seem to be making the assumption that anything you
don't like must be part of the "opposition".
The Nazis had a regulated society, true, but not
regulated in an economic sense as a true communistic
society would. You may be confusing "rightist" with
"no government regulation at all" (usually associated with the label Libertarian in the US).
When most people think of "the right", they think
of economic anarchy, but government imposed moral
regulation of people's private lives.
Typically, the "left" is considered to be in favor
of a "managed economy", so no-one gets "too poor"
or "too rich".
Actually, while you can't get much further "right"
than the Nazis, the name Nazi is a contraction of
the German words for National and Socialist. I'm
sure someone who's better at German than me can
fill in the actual words. Yeah, it may seem like
a contradiction, but there it is.
You think doing without arrays is a pain - try working all day with no command-line editing beyond the terminal driver.
I had that situation back in '89-92 when I worked on
Suns from a VT100. My solution was to run Emacs, and
open a shell inside that. It also gave me crude windowing, and other cool stuff.
A far better measure of how well an OS has been ported is an Ultra 2.
I have an Ultra 1 (UltraSparc @ 167MHz), and I have
to say that my experience installing NetBSD 1.6.1
was quite similar to his experience (panic during
install, no binary packages available, etc.) And,
besides, the review seemed to focus on the app side
of things, not the driver/hardware side.
That said, I don't think there's anything major that can't be fixed. Once installed, it's run flawlessly, and the only package I couldn't compile
from pkgsrc is Apache 2.x (I installed Apache 1.x
instead).
I've had the (32-bit) sparc port of NetBSD running
solidly on a SparcStation-2 for over a year and a
half (it's my DHCP/NTP/DNS server).
I inherited an old SparcStation2 with 400MB disk
space and 32MB of RAM. Not many OS's could install
on such a small system, let alone make it a valuable
resource. But I put NetBSD on it, and now it's one
of my most stable systems. I made it my DHCP server
and primary DNS for 7 domains.
Ok, I went back and checked some of my books. You
may be confusing "delayed ACK" with what is called the congestion window. What I was calling "the window" is actually the send window. The
congestion window is indeed measured in packets (or
segments in TCP parlance).
The congestion window is expanded by an algorithm
called "slow start", which will only transmit a small number of packets before getting an ACK (either linearly (e.g. add 1), or exponentially (e.g. double each time), depending on the implementation).
I believe delayed ACK means waiting a few milliseconds after you would ordinarily send an ACK, in case a packet is being sent from the
receiving side on which the ACK bit can "piggyback".
It's complicated, and I clearly don't understand
it as well as I thought I did (but it will make
for some interesting reading). OS developers (like Linux) are tweaking these things as time goes by, but it sounds like the article is describing a more extreme form of tweaking of these alorithms.
First, thanks for correcting my use of MTU vs MSS.
I don't think the packet size makes much of a difference as far as when TCP will stop to wait for an ACK (remember, the TCP "window size" is specified in bytes (or "octets" for you networking weenies)).
But, yes, you're correct in that it seems that the
new work is addressing the effect when bandwidth
allows you to transmit more data than fits in a window during the RTT (round-trip time). Perhaps they're increasing the number of bits used in specifying the window, but they definitely seem to
be addressing the speed at which the default window size gets adjusted upward.
Good summary. But I have to correct you on one minor
point:
DSL is a link-layer protocol.
Actually, it's not link layer (layer 2 in OSI, which separates a bit stream into units called "frames", and identfies frame types, addressing, etc.). It's
physical layer (using voltages to encode bits), in
the case of DSL, using frequency-domain multiplexing
to encode 3 different links (inbound data, outbound
data, voice) onto the same wires.
Usually DSL uses PPP and Ethernet (PPPoE) or PPP and ATM (PPPoA) as its link layers.
Yes, you can adjust the max. receive (MRU) and transmit (MTU) packet sizes. The MRU isn't adjusted,
usually, just accept as big a packet as you can. The
MTU can be adjusted manually (by the sysadmin), or
automatically (PMTUD - path MTU discovery).
However, adjusting the MTU has little to do with speed, as the Window Size (how much data can be
transmitted before being acknowledged by the far end)
is specified in number of bytes (in TCP). I suppose
it could have some effect on speed, as when
you send a packet that exceeds the MTU, it gets
"segmented" into multiple IP packets, each with its
own packet header overhead (and if any get lost,
the whole bunch have to get retransmitted).
What this new protocol deals with, however, is
dynamically varying the window-size. Current TCP
does that, but apparently not in as efficient a
manner as this.
So all this "x thousand times faster than DSL" is
just complete bullshit. You'll never get
any faster speeds than the slowest link between
point A and point B. This new protocol simply tries
to use the Y/bits-per-second available more
efficiently. And you won't notice the inefficiency
of the current TCP at speeds most DSL/cable/dialup
users have available.
Actually, even while the military is scrambling to get more Arab linguists,
they recently discharged at least seven of them.
Talk about shooting yourself in the foot. So, I have
no sympathy for the military needing experts
in languages (human or computer).
Sorry, but the fact that they are
creationists implies that they are
not "quite intelligent, and very good
in the fields of biology, physics, and
mathematics". So, feel free to start ignoring
their opinions.
No, it was quite clear. Quite clearly the result of
denial on your part. If someone is correct
on everything outside of politics, the likelihood
of them also being right on politics is greater
than yours. Clearly.
Some of the most blindingly leftist people I know are otherwise quite bright
So, the smart people you know have "left" political
leanings? Maybe this should tell you something. Has
it occurred to you that they are correct, and you
are wrong? I suggest you ask them who you should vote
for this fall, and follow their advice.
And, if you'd bothered to read the article, the scientists in question are not trying to influence
politics, but trying to keep from politicians from
influencing their research.
Devices are a source of interrupts. Error conditions in the processor are another (divide by
zero, for example). And finally, explicit instructions to generate interrupts are used for
various purposes: break points inserted by a
debugger, and calling the operating system (which is
usually not done via a simple subroutine call due to
priviledge levels and virtual memory issues).
I started using BSD (v4.2 on a VAX) in 1983, years before I started using MS Windows (v1.2 in 1987)... and even more years before Windows came up with the BSoD (Win95 or WinNT?)
Slashdot Mirror
Then again, I am following the sheep to this summer's Ozzfest. But with a re-united Judas Priest, I couldn't resist. Oh, well, I guess, small steps.
And I enjoy Peter Schickele. Both PDQ Bach, and his "Schickele Mix" program which I used to listen to on NPR (not sure if it's still on, but I'm not usually up that early on a weekend any more).
I bet you could build a low-speed, high-torque paddle wheel (it would need to have a lot of surface area being pushed on by the river). Then, using gear ratios, you can convert that to high-speed, low-torque that may be needed by your generator. Not being a mechanical engineer, I'll leave it at that. :-)
Sorry, couldn't resist. :-)
That may (or may not) have been the usage of those terms in early 19th century France, but not in the 20th and 21st century United States.
I think the term you are looking for is "authoritarian", or "dictatorship". Or just plain old "revolutionary". And need I remind you of a reasonably successful revolution which took place in the late 1770's and early 1780's?
Going back to the Nazis, while I suppose you could call the "Beerhall Putsch" a mini-revolution, it was unsuccessful, and it wasn't until the Nazis had secured more power through elections that they were able to complete their rise to complete dictatorship.
Soon thereafter, the "Night of the Long Knives" removed any pretense of "socialism" in the National Socialist party. The Nazis were anti-communism as their invasion of Russia in 1941 showed.
Bringing this back to our country, the group who I think inherits the idealogical heritage of the Nazis, the conservative Christians, is also in favor of having the government regulate citizens' private lives, and is very anti-homosexual (brings back images of the Nazis placing pink triangles on gay men, just as they put yellow stars on Jews).
I think that's about all I need to say.
The Nazis had a regulated society, true, but not regulated in an economic sense as a true communistic society would. You may be confusing "rightist" with "no government regulation at all" (usually associated with the label Libertarian in the US).
When most people think of "the right", they think of economic anarchy, but government imposed moral regulation of people's private lives. Typically, the "left" is considered to be in favor of a "managed economy", so no-one gets "too poor" or "too rich".
Actually, while you can't get much further "right" than the Nazis, the name Nazi is a contraction of the German words for National and Socialist. I'm sure someone who's better at German than me can fill in the actual words. Yeah, it may seem like a contradiction, but there it is.
I hate to reply to my own comment, but it looks like the binary packages are available for 1.6.2 for sparc64.
That said, I don't think there's anything major that can't be fixed. Once installed, it's run flawlessly, and the only package I couldn't compile from pkgsrc is Apache 2.x (I installed Apache 1.x instead).
I've had the (32-bit) sparc port of NetBSD running solidly on a SparcStation-2 for over a year and a half (it's my DHCP/NTP/DNS server).
A big "Thank you!" to all the NetBSD developers!
The congestion window is expanded by an algorithm called "slow start", which will only transmit a small number of packets before getting an ACK (either linearly (e.g. add 1), or exponentially (e.g. double each time), depending on the implementation).
I believe delayed ACK means waiting a few milliseconds after you would ordinarily send an ACK, in case a packet is being sent from the receiving side on which the ACK bit can "piggyback".
It's complicated, and I clearly don't understand it as well as I thought I did (but it will make for some interesting reading). OS developers (like Linux) are tweaking these things as time goes by, but it sounds like the article is describing a more extreme form of tweaking of these alorithms.
I don't think the packet size makes much of a difference as far as when TCP will stop to wait for an ACK (remember, the TCP "window size" is specified in bytes (or "octets" for you networking weenies)).
But, yes, you're correct in that it seems that the new work is addressing the effect when bandwidth allows you to transmit more data than fits in a window during the RTT (round-trip time). Perhaps they're increasing the number of bits used in specifying the window, but they definitely seem to be addressing the speed at which the default window size gets adjusted upward.
Usually DSL uses PPP and Ethernet (PPPoE) or PPP and ATM (PPPoA) as its link layers.
However, adjusting the MTU has little to do with speed, as the Window Size (how much data can be transmitted before being acknowledged by the far end) is specified in number of bytes (in TCP). I suppose it could have some effect on speed, as when you send a packet that exceeds the MTU, it gets "segmented" into multiple IP packets, each with its own packet header overhead (and if any get lost, the whole bunch have to get retransmitted).
What this new protocol deals with, however, is dynamically varying the window-size. Current TCP does that, but apparently not in as efficient a manner as this.
So all this "x thousand times faster than DSL" is just complete bullshit. You'll never get any faster speeds than the slowest link between point A and point B. This new protocol simply tries to use the Y/bits-per-second available more efficiently. And you won't notice the inefficiency of the current TCP at speeds most DSL/cable/dialup users have available.
Some tech journalists are just idiots.
Actually, even while the military is scrambling to get more Arab linguists, they recently discharged at least seven of them. Talk about shooting yourself in the foot. So, I have no sympathy for the military needing experts in languages (human or computer).
I usually test new code from my desk, but the equipment is in the lab where it's (usually) connected via the LAN.
GAAAAAAHHHHH!!!!
Sorry, but the fact that they are creationists implies that they are not "quite intelligent, and very good in the fields of biology, physics, and mathematics". So, feel free to start ignoring their opinions.
No, it was quite clear. Quite clearly the result of denial on your part. If someone is correct on everything outside of politics, the likelihood of them also being right on politics is greater than yours. Clearly.
And, if you'd bothered to read the article, the scientists in question are not trying to influence politics, but trying to keep from politicians from influencing their research.
Devices are a source of interrupts. Error conditions in the processor are another (divide by zero, for example). And finally, explicit instructions to generate interrupts are used for various purposes: break points inserted by a debugger, and calling the operating system (which is usually not done via a simple subroutine call due to priviledge levels and virtual memory issues).
I started using BSD (v4.2 on a VAX) in 1983, years before I started using MS Windows (v1.2 in 1987)... and even more years before Windows came up with the BSoD (Win95 or WinNT?)