A bridge will broadcast all ethernet broadcast packets to all hosts on the network.
I don't know what kind of IGMP you were seeing, but I'd be astounded if it wasn't either bound for the ethernet address of something behind your cable modem or the ethernet broadcast address.
This doesn't sound credible to me. In a Cable Network the CM ( Cable Modem ) receives on a downstream frequency band and sends on an upstream frequency band to the CMTS ( Cable Modem Termination
System). The spec requires the CM CMTS system to act as a bridge. It is NOT hubbed. You can listen on your ethernet port until you're blue in the face and you will only see your own traffic and the broadcast traffic on the network. Period. Ever.
Now, this does not rule out ARP spoofing, but the only really interesting ARP to spoof would be the one for the default gateway on the network. Since
the gateway for the network is living on the CMTS and since any ARP request must pass through the CMTS before getting to our spoofer, I would expect the spoofed replies to arrive after the legitimate ones from the CMTS. Additionally, I would not be surprised to find out that the CMTS suppresses attempts to ARP spoof it's addresses ( and if it doesn't now, it will in the near future ).
Additionally, I also wonder if these articles take into account the admin costs. Ignoring the misleading numbers the article gives. Is it easier to admin 3 Windows servers or 3 *nix servers? In my experience windows seems to be more hands off than *nix, or Solaris in particular. Maybe I'm comparing Apple's and Oranges though given my experience.
Hmm. Admining three Windows servers vs three *nix servers is an interesting comparison. I know that in all the places I've ever worked ( where the scale was much larger than 3 servers ) it was incredible how much more manpower it took to look after Windows than *nux. I've been in a couple of places where you had one dude looking after a half a dozen *nix servers and several hundered *nix workstations and a group of 12 people looking after a similar number of Windows workstations and their supporting servers.
But at the three server level I'm not so sure. My guess would be it would be more work to set up the *nix servers, but less work to keep them running.
Do people really read everything in AIP journals? Nobody I knew did. But I did know several people who would go to look at what was new this morning on arxiv.org in their field ( and possibly related fields ). They'd scan through a couple of pages of new preprints looking for people whose work they knew to be worth reading, or for abstracts that sounded promising.
Perhaps the AIP has gained some celerity recently. I would hope so, but I suspect the value proposition offered by the traditional AIP journals is wearing very thin in many subdisciplines ( like hep-* etc. ).
Perhaps I'm simply misinformed... but when I was last in the High Energy Physics community the
e-print archive
was the definitive reference.
One never actually fell back to the journals unless the paper predated the early 90s. People would reference other peoples work via the e-print archive reference number. e-prints circulated so widely that most major papers had already been read and reviewed by the relavent people long before it actually hit the official peer review process at the journals. By the time a paper made it into the journals it was VERY old news.
Yes, people still submitted their papers to the standard APS journals for publication, but nobody read them. Everybody read the e-print archive. Most people couldn't even tell you what journal most of the articles had been published in, nobody cared.
I switched from RedHat to Debian about 2 years ago,
largely because of a series of technical decisions
that RedHat made that I didn't want to have to put
up with anymore ( gcc-2.96 anyone ).
I did not switch from RedHat because I question their ethics. To the best of my knowledge they have always opensourced anything they've done.
They have eventually open sourced anything they've acquired. All under the GPL. I don't see how
we can fault them for that.
RedHat has done things that I feel are stupid
( gcc-2.96, recent behavior towards KDE ). But
NONE of these things are in anyway unethical. Some of them have been handled badly from a PR perspective. But I have yet to see RedHat do anything that even slighly had nefarious intent.
RedHat provides a very valuable service. They provide a familiar interface to the commercial world. Large companies want a standard distribution with support contracts to help them sleep well at night. Large commercial software producers who right rather overly rigid software NEED a platform to implement to ( because they can't implement to standards, or deal with minor variations ). RedHat provides all of these interests with what they need.
People should really leave RedHat alone on the Microsoft comparison front. Kick them around over some of the dumb technical decisions they make if you like. That's fair and decent criticism, but don't FUD them.
The laws on this matter tend to vary from state to
state ( as murder, like most crimes, is a state matter in the US ). In two of the states I have resided in ( Indiana and North Carolina ) there is a presumption that if some one breaks into your home they mean you bodily harm. This renders any use of force against them self defense against bodily harm in the eyes of the law. I tend to think this is reasonable. I can't speak to the laws in Colorado, but I would be shocked ( and dismayed ) if defense of property figured into the right to use force to defend yourself against a burglar in anyway.
What are the laws like in Sweden regarding the use of force against someone who has broken into your home?
In general you have just as much authority to use force to defend another person from violence as you do to defend yourself. Even if you don't know the person.
Sure, no problem there. I don't see anything in my statements that suggested that you didn't have as much right to use violence to defend someone else from bodily harm as you do to defend yourself from bodily harm.
I live in Colorado where I may shoot a person dead if he is both 1. on my property and 2. I have reasonable cause to beleive he is or is about to commit another crime (against a person or property.)
Interesting. In most of the states who's laws I am familiar with the right to shoot an intruder in your home dead is rested firmly on the assupmtion in the law that someone who is breaking into your home if perfectly willing to use lethal force against you, thus reducing it to a defense against bodily harm case. In most states I believe the simple act of them breaking into your home is sufficient cause for you to reasonably believe they intend to harm you. I've never seen any state provide justification for the use of lethal force based on a justification of defense of property. Perhaps Colorado is different.
I think your opinion is based more on your pacifistic world-view than on any actual facts.
I think perhaps I've not communicated to you clearly. You are perhaps the first person I've encountered who has ever accused me of pacifism. I have no problems whatsoever with the application of force within reasonable limits, as proscibed by law. I also happen to believe that the right to use lethal force against an intruder in your home based upon the assumption that they intended to do you harm is reasonable. That is hardly the point of view of a pacifist.
Well, you have really twisted my example around. Someone actively attacking your computer (network) or actively breaking into your house is not related to your vigilante revenge scenario in any way, so I'll dismiss it out of hand.
Ah... I think I see where some of the confusion is now. Please note the tense I used with the word burgled. Someone currently, actively, burgling your home is a direct threat to your person for which you can reasonably respond with deadly force in most states. Belief that someone has, at some point in the past burgled is quite different as it carries no threat of bodily harm.
The point I was attempting to make is this: those senarios in which the criminal conduct of another person are grounds justifying retaliatory action which is normally proscribed by law are generally limited to cases involving the threat of bodily harm to a person. I know of no examples in US law permitting actions normally proscribed by law being justified by crimes or threats against property ( with the possible exception of your assertion with regard to Colorado state law).
Vigilante justice is not the solution. When I discover someone has burgled my house, and I have reason to believe I KNOW who did it, that does not entitle me to go break into their house to take my stuff back and avenge myself upon them.
It's important to remember WHY vigilante actions are generally illegal:
They are highly error prone
They effectively invalidate all of the accused rights summarily.
They lead to chains of criminal behavior that can be hard to unravel.
I can only think of one set of circumstances in which our culture and law condone vigilante justice: self defense of a human being against bodily harm.
It is important to remember that computer crime is almost universally property crime. With rare exceptions there is absolutely no danger to the person of a human being posed by computer cracking, and thus no reasonable basis for authorizing vigilante justice.
Let's do you video calculation again.
If memory serves compressed HDTV is about
19MB/second. Let's call it 20MB/second
to make the calculations easier. So 1 terabyte
gives us about 50,000 seconds. This is about
13-14 hours. So we are looking at about
13 hours of HDTV.
Trying to seriously drive a gig ethernet card.
Trying to do network capture and analysis for a
gig ethernet card. Trying to use multiple
fast ethernet ports to simultaneously monitor
different network points. Any of these will
blow a PCI bus really fast. I currently have
to spend a great deal of money on specialized
equipment to do poorly what I could do with
a Linux box if it the applications didn't
choke on the PCI bus speed.
And don't even think about the new 10 gig ethernet
you're SOL out of the box on that, curtesy of your
crappy PCI bus.
The Hard Drive may be the slowest component, but
the PCI bus is the big bottleneck for MANY applications. I can cache enough of the things I
need to hit the disk for in memory to render the
disk bottleneck irrelavent to many applications.
What I can't do is drive stuff across the crappy
PCI bus any faster. Could we please move away
from PCI towards something better!!
I know that there are a lot of rotten PMs out there, but if you ever have the good fortune to work on a project with a good PM you will never willingly work on a project without one again.
I miss my PM. Her job was basically:
Beat up other people to get us the resources we needed to succeed
Block outside people from bothering us about things so we could work on the project
Keep track of the pedestrian details of the project so that we were free to actually get the work done
When prototypes for the project were running late, I didn't have to spend endless hours chasing people down and tracking the issues delaying them. My PM did that.
When the project had slipped 6 weeks, I wasn't the one on the calls getting yelled at and yelling back about the fact that more than 50% of the TYPES of prototypes we needed hadn't even been delivered yet. My PM did that. I was down in the lab working.
When I had to attend technical calls ( like bug
scrubs ) I didn't have to go dig up the bugs being
covered so I could review them for the meeting. My PM always met with us 30 minutes prior and went over the list so that we could get things clearly in mind going into the call.
And when the shit hit the fan, and we were death marching till 2am for weeks on end, my PM was there making sure we got fed ( on the company dime ), and staying late to make sure we did eventually go home and sleep.
None of this really requires much technical skill on the part of the PM. All it requires is a respect for the team and an understanding that the most effective way to get your project in on time is to support the team. By the middle of the project we ( the technical guys ) where willing to kill ourselves to meet the project objectives for this PM.
Most of the problems of the developing world are
politcal, not economic. Scratch a poor third world
nation that is suffering from starvation, and you will generally find a despotic ruling regime, not
a transnational corporation enslaving the people.
This is NOT to say that there are not nasty transnational corporations doing bad things, but
for the most part in most places where large numbers of people are facing starvation today it is due to their despotic rulers.
Examples:
Zimbabwe, formerly a breadbasket country, is facing famine in large part do to Mugabe's disastrous land use policies.
North Korea, suffering famine recently ( although
I haven't checked the status lately ) due largely to the mismanagement of their communist government.
If you really want to know who is oppressing the people of the third world, look not to the developed world and it's consuption, but rather to the pointless excesses of the rulers of the third world.
I've worked with both DOCSIS 1.0 and 1.1. The MAP
MAC message is an integral part of both 1.0 and
1.1. It is not new in 1.1. The cable modem
needs to specify a COS ( class of service )
during it's registration process to the CMTS
( cable modem termination system ) in both versions
of the standard. The CMTS enforces the COS in
both version of the standard. The only major
changes I recall between 1.0 and 1.1 with regard
to how COS was handled was the introduction of
dynamic classes of service for cable modems to
accomidate telephony services.
Funny that. Cable actually has really quite good
QOS/COS features if your cable provider is willing
to TURN THEM ON. DOCSIS ( www.cablelabs.com )
has had the ability to set the max up/down per
customer since version 1.0.
Additionally it has
also had the capacity to set a minimum guaranteed
upstream/downstream bandwidth. I've played
with cable modems and CMTSes. The QOS works.
But guess what, your average Cable Operator
doesn't use them much. You 'premium service'
is likely NOT buying you any higher a max
on your bandwidth ( or any minimum guarantee )
even though the technology can provide it.
So
don't whine about the college students, complain
to your cable company about refusing to provide
the class of service you wish to purchase, even
though the existing technology supports it.
This is probably a DOCSIS variant. You can look
into the DOCSIS specs at:
http://www.cablemodem.com/specifications.html
if you are interested. Basically DOCSIS is the
Data over Cable System Interface Specification.
It's how your cable modem works.
DOCSIS puts your downstream data inside MPEG
frames on a normal 6Mhz television channel.
For cable modem normaly this is the only digital
data on the downstream channel and so the MPEG
framing is largely irrelavent. I'm pretty
sure that your standard digital TV signal is
exactly the same as your DOCSIS signal up until
we get into the contents of the MPEG frame
( which probably is MPEG for digital TV ). Since
You can get about 28Mb/s in a QAM-64 carrier and
compressed HDTV only takes about 19Mb/s, I'd
say this is a creative use of the leftover
bandwidth.
As to the phone dial back... getting a return
signal even over cable is trickier, over
air it is likely to be downright prohibitive
( can you image how much power you would need
to broadcast from your home to get back to the
station? ). Because not all cable providers
originally had cable plants that could provide for
upstream channels, the DOCSIS spec already lays
out neatly specifications for providing for
a dial return path.
I'll almost bet this is what they are doing in
Cincinnati. Doing anything to different would
just make their equipment at their headends
really expensive for no good reason.
Funny,
I've had to hack around at least
one bug in Solaris 8 that has been present
at least as far back as Solaris 2.5.1 ( that
I can verify ) but has never been fixed.
That would be the line length bug in 'sed'.
Solaris's version of sed simply pukes with
command lines that are two long. This in
turn trips up libtool, and thus a whole
lot of other software. I have had to install
the gnu version of sed because of this for
YEARS, and no one at Sun seems to care.
As far as I can tell Sun has decided that some
bugs are simply acceptable and will not be
fixed, and somehow the market translates this
into a notion of stability.
Don't get me wrong, there are some applications
for which I would pick Sun over linux, serving
NFS in a serious way would be one. Another would
be for running on Sun's really big hardware if
I had an application that needed that kind of
juice in a central place. Yet another would
be if I was looking at running a cache bound
application ( Sun has MUCH nicer cache than
you will find in PC hardware ).
But PC hardware is catching up fast.
In any senario in which Linux + x86 hardware
will do the job it will almost always do it
cheaper, and frequently better ( although not
always, see previous examples ).
Who could have predicted THIS nightmare
on
Electronic Abacus
·
· Score: 2, Insightful
Excuse me?
If anything we are LACKING sufficient computing to
allow for efficient operation of our businesses.
I recently went to refinance my car ( at a bank that will remain nameless ). This bank held my original car loan. I spent an hour filling out
paperwork ( all of which had been filled out with
my original loan ), having the loan officer call my
insurance company to get my insurance information
( even though they had all of my insurance info on
file ), etc. All of this totally redundant. I had to come back the NEXT day because the guy
the loan officer calls to do credit checks and fax
them to him was busy.
All of this should have been accomplishable over the
web. There is NO reason that it had to be that
hard.
Oh yes, and then they proceeded to automagically
debit my old loan payment ( several days after
the old loan had been paid off in full by the new
loan ) because it takes about a week for the PAPER
to work it's way through channels.
It took almost six weeks for the bank to
restore order to my account ( I will not recount the full ins and outs of that, but it was bad ).
All of this was unecessary. If they had proper computer systems handling the back end of the bank I should have been able to go to the web page for my account arrange refinancing there in under 10 minutes.
Note that every one of the bank employees I dealt
with had a computer on their desk. What made this
experience so inefficient ( and frustrating ) was
not their lack of computers, but the lack of competent back end systems for them to access with those computers. That is were the efficiency comes.
Talk to some folks at Fermi Lab about how to
handle REALLY big data streams. Last I
heard they produce about 250GB/s of data
when the beam is on ( after the first stage
hardware filtering has reduced the data set ).
These are the folks to talk to if you have any
really heavy duty data needs.
This was first seriously proposed by
Gerald K. Oneill of Princeton University in
1975! It was feasable ( and even profitable )
then, but the capitalization was to high for
any organization on earth but the US Government
to undertake. The only reason we haven't done
it already is because of a defect of will, a
myopy of purpose, and inability to look further
ahead than the next election.
When will we, the citizens of the United States,
have the vision to demand these sorts of
projects from our government? Oneill's initial
proposal had an estimated 20 year pay back time, for
the first powersat. Subsequent powersats would
have been much cheaper. If the proposal Oneill
made had been taken up seriously in 1976, and taken
say 2 years to get it's political legs so that
actual work began in 1978, and it took ten
years to build, we would have had cheap abundant
energy by 1988.
Given cheap abundant energy it would be feasible to
produce, for example, metal hydride or fuel cell
powered cars. Given a 10 year ramp up and phase in
for those technologies we would have in 1998
been largely petroleum free ( at least for
power ).
Does anyone question that this would be a better
place to be... and we could be there by now, if
only we had the vision, and the will.
I've recently switched from Enlightenment/Gnome
as a desktop to KDE. Two major driving
factors where Konqueror and Konsole ( tabbed
shells... mmm... ). I had previously
used Galeon, and I prefer Konqueror in most
respects ( but please for the love of God bring
tabbed browsing to Konqueror!! ).
I still use Evolution for email
( kmail is just not as feature rich or usable )
and Gnumeric for spreadsheets. I would say that
Kmail and Kspread still need a fair bit more work
to catch up (particularily kmail).
Ah... but they have not in fact copyrighted the
numbers. They have copyrighted the musical
representation of these numbers as DTMF tones.
Additionally, like hell numbers aren't copyrightable.
What do you think an mp3 file is? It's a very
large number. In fact EVERYTHING digital is a
number. So if you can't copyright a number, how
then is software, source code, digital music,
digital video copyrightable?
Slashdot Mirror
A bridge will broadcast all ethernet broadcast packets to all hosts on the network. I don't know what kind of IGMP you were seeing, but I'd be astounded if it wasn't either bound for the ethernet address of something behind your cable modem or the ethernet broadcast address.
Now, this does not rule out ARP spoofing, but the only really interesting ARP to spoof would be the one for the default gateway on the network. Since the gateway for the network is living on the CMTS and since any ARP request must pass through the CMTS before getting to our spoofer, I would expect the spoofed replies to arrive after the legitimate ones from the CMTS. Additionally, I would not be surprised to find out that the CMTS suppresses attempts to ARP spoof it's addresses ( and if it doesn't now, it will in the near future ).
Hmm. Admining three Windows servers vs three *nix servers is an interesting comparison. I know that in all the places I've ever worked ( where the scale was much larger than 3 servers ) it was incredible how much more manpower it took to look after Windows than *nux. I've been in a couple of places where you had one dude looking after a half a dozen *nix servers and several hundered *nix workstations and a group of 12 people looking after a similar number of Windows workstations and their supporting servers.
But at the three server level I'm not so sure. My guess would be it would be more work to set up the *nix servers, but less work to keep them running.
Do people really read everything in AIP journals? Nobody I knew did. But I did know several people who would go to look at what was new this morning on arxiv.org in their field ( and possibly related fields ). They'd scan through a couple of pages of new preprints looking for people whose work they knew to be worth reading, or for abstracts that sounded promising.
Perhaps the AIP has gained some celerity recently. I would hope so, but I suspect the value proposition offered by the traditional AIP journals is wearing very thin in many subdisciplines ( like hep-* etc. ).
One never actually fell back to the journals unless the paper predated the early 90s. People would reference other peoples work via the e-print archive reference number. e-prints circulated so widely that most major papers had already been read and reviewed by the relavent people long before it actually hit the official peer review process at the journals. By the time a paper made it into the journals it was VERY old news.
Yes, people still submitted their papers to the standard APS journals for publication, but nobody read them. Everybody read the e-print archive. Most people couldn't even tell you what journal most of the articles had been published in, nobody cared.
I did not switch from RedHat because I question their ethics. To the best of my knowledge they have always opensourced anything they've done. They have eventually open sourced anything they've acquired. All under the GPL. I don't see how we can fault them for that.
RedHat has done things that I feel are stupid ( gcc-2.96, recent behavior towards KDE ). But NONE of these things are in anyway unethical. Some of them have been handled badly from a PR perspective. But I have yet to see RedHat do anything that even slighly had nefarious intent.
RedHat provides a very valuable service. They provide a familiar interface to the commercial world. Large companies want a standard distribution with support contracts to help them sleep well at night. Large commercial software producers who right rather overly rigid software NEED a platform to implement to ( because they can't implement to standards, or deal with minor variations ). RedHat provides all of these interests with what they need.
People should really leave RedHat alone on the Microsoft comparison front. Kick them around over some of the dumb technical decisions they make if you like. That's fair and decent criticism, but don't FUD them.
The laws on this matter tend to vary from state to state ( as murder, like most crimes, is a state matter in the US ). In two of the states I have resided in ( Indiana and North Carolina ) there is a presumption that if some one breaks into your home they mean you bodily harm. This renders any use of force against them self defense against bodily harm in the eyes of the law. I tend to think this is reasonable. I can't speak to the laws in Colorado, but I would be shocked ( and dismayed ) if defense of property figured into the right to use force to defend yourself against a burglar in anyway.
What are the laws like in Sweden regarding the use of force against someone who has broken into your home?
Sure, no problem there. I don't see anything in my statements that suggested that you didn't have as much right to use violence to defend someone else from bodily harm as you do to defend yourself from bodily harm.
I live in Colorado where I may shoot a person dead if he is both 1. on my property and 2. I have reasonable cause to beleive he is or is about to commit another crime (against a person or property.)
Interesting. In most of the states who's laws I am familiar with the right to shoot an intruder in your home dead is rested firmly on the assupmtion in the law that someone who is breaking into your home if perfectly willing to use lethal force against you, thus reducing it to a defense against bodily harm case. In most states I believe the simple act of them breaking into your home is sufficient cause for you to reasonably believe they intend to harm you. I've never seen any state provide justification for the use of lethal force based on a justification of defense of property. Perhaps Colorado is different.
I think your opinion is based more on your pacifistic world-view than on any actual facts.
I think perhaps I've not communicated to you clearly. You are perhaps the first person I've encountered who has ever accused me of pacifism. I have no problems whatsoever with the application of force within reasonable limits, as proscibed by law. I also happen to believe that the right to use lethal force against an intruder in your home based upon the assumption that they intended to do you harm is reasonable. That is hardly the point of view of a pacifist.
Well, you have really twisted my example around. Someone actively attacking your computer (network) or actively breaking into your house is not related to your vigilante revenge scenario in any way, so I'll dismiss it out of hand.
Ah... I think I see where some of the confusion is now. Please note the tense I used with the word burgled. Someone currently, actively, burgling your home is a direct threat to your person for which you can reasonably respond with deadly force in most states. Belief that someone has, at some point in the past burgled is quite different as it carries no threat of bodily harm.
The point I was attempting to make is this: those senarios in which the criminal conduct of another person are grounds justifying retaliatory action which is normally proscribed by law are generally limited to cases involving the threat of bodily harm to a person. I know of no examples in US law permitting actions normally proscribed by law being justified by crimes or threats against property ( with the possible exception of your assertion with regard to Colorado state law).
It's important to remember WHY vigilante actions are generally illegal:
I can only think of one set of circumstances in which our culture and law condone vigilante justice: self defense of a human being against bodily harm.
It is important to remember that computer crime is almost universally property crime. With rare exceptions there is absolutely no danger to the person of a human being posed by computer cracking, and thus no reasonable basis for authorizing vigilante justice.
You are correct... mia culpa... I always forget :)
the bytes to bits thing
Let's do you video calculation again. If memory serves compressed HDTV is about 19MB/second. Let's call it 20MB/second to make the calculations easier. So 1 terabyte gives us about 50,000 seconds. This is about 13-14 hours. So we are looking at about 13 hours of HDTV.
Trying to seriously drive a gig ethernet card. Trying to do network capture and analysis for a gig ethernet card. Trying to use multiple fast ethernet ports to simultaneously monitor different network points. Any of these will blow a PCI bus really fast. I currently have to spend a great deal of money on specialized equipment to do poorly what I could do with a Linux box if it the applications didn't choke on the PCI bus speed.
And don't even think about the new 10 gig ethernet you're SOL out of the box on that, curtesy of your crappy PCI bus.
The Hard Drive may be the slowest component, but the PCI bus is the big bottleneck for MANY applications. I can cache enough of the things I need to hit the disk for in memory to render the disk bottleneck irrelavent to many applications. What I can't do is drive stuff across the crappy PCI bus any faster. Could we please move away from PCI towards something better!!
I miss my PM. Her job was basically:
When prototypes for the project were running late, I didn't have to spend endless hours chasing people down and tracking the issues delaying them. My PM did that.
When the project had slipped 6 weeks, I wasn't the one on the calls getting yelled at and yelling back about the fact that more than 50% of the TYPES of prototypes we needed hadn't even been delivered yet. My PM did that. I was down in the lab working.
When I had to attend technical calls ( like bug scrubs ) I didn't have to go dig up the bugs being covered so I could review them for the meeting. My PM always met with us 30 minutes prior and went over the list so that we could get things clearly in mind going into the call.
And when the shit hit the fan, and we were death marching till 2am for weeks on end, my PM was there making sure we got fed ( on the company dime ), and staying late to make sure we did eventually go home and sleep.
None of this really requires much technical skill on the part of the PM. All it requires is a respect for the team and an understanding that the most effective way to get your project in on time is to support the team. By the middle of the project we ( the technical guys ) where willing to kill ourselves to meet the project objectives for this PM.
This is NOT to say that there are not nasty transnational corporations doing bad things, but for the most part in most places where large numbers of people are facing starvation today it is due to their despotic rulers.
Examples:
Zimbabwe, formerly a breadbasket country, is facing famine in large part do to Mugabe's disastrous land use policies.
North Korea, suffering famine recently ( although I haven't checked the status lately ) due largely to the mismanagement of their communist government.
If you really want to know who is oppressing the people of the third world, look not to the developed world and it's consuption, but rather to the pointless excesses of the rulers of the third world.
I've worked with both DOCSIS 1.0 and 1.1. The MAP MAC message is an integral part of both 1.0 and 1.1. It is not new in 1.1. The cable modem needs to specify a COS ( class of service ) during it's registration process to the CMTS ( cable modem termination system ) in both versions of the standard. The CMTS enforces the COS in both version of the standard. The only major changes I recall between 1.0 and 1.1 with regard to how COS was handled was the introduction of dynamic classes of service for cable modems to accomidate telephony services.
Funny that. Cable actually has really quite good QOS/COS features if your cable provider is willing to TURN THEM ON. DOCSIS ( www.cablelabs.com ) has had the ability to set the max up/down per customer since version 1.0.
Additionally it has also had the capacity to set a minimum guaranteed upstream/downstream bandwidth. I've played with cable modems and CMTSes. The QOS works.
But guess what, your average Cable Operator doesn't use them much. You 'premium service' is likely NOT buying you any higher a max on your bandwidth ( or any minimum guarantee ) even though the technology can provide it.
So don't whine about the college students, complain to your cable company about refusing to provide the class of service you wish to purchase, even though the existing technology supports it.
This is probably a DOCSIS variant. You can look
into the DOCSIS specs at:
http://www.cablemodem.com/specifications.html
if you are interested. Basically DOCSIS is the
Data over Cable System Interface Specification.
It's how your cable modem works.
DOCSIS puts your downstream data inside MPEG
frames on a normal 6Mhz television channel.
For cable modem normaly this is the only digital
data on the downstream channel and so the MPEG
framing is largely irrelavent. I'm pretty
sure that your standard digital TV signal is
exactly the same as your DOCSIS signal up until
we get into the contents of the MPEG frame
( which probably is MPEG for digital TV ). Since
You can get about 28Mb/s in a QAM-64 carrier and
compressed HDTV only takes about 19Mb/s, I'd
say this is a creative use of the leftover
bandwidth.
As to the phone dial back... getting a return
signal even over cable is trickier, over
air it is likely to be downright prohibitive
( can you image how much power you would need
to broadcast from your home to get back to the
station? ). Because not all cable providers
originally had cable plants that could provide for
upstream channels, the DOCSIS spec already lays
out neatly specifications for providing for
a dial return path.
I'll almost bet this is what they are doing in
Cincinnati. Doing anything to different would
just make their equipment at their headends
really expensive for no good reason.
Funny,
I've had to hack around at least
one bug in Solaris 8 that has been present
at least as far back as Solaris 2.5.1 ( that
I can verify ) but has never been fixed.
That would be the line length bug in 'sed'.
Solaris's version of sed simply pukes with
command lines that are two long. This in
turn trips up libtool, and thus a whole
lot of other software. I have had to install
the gnu version of sed because of this for
YEARS, and no one at Sun seems to care.
As far as I can tell Sun has decided that some
bugs are simply acceptable and will not be
fixed, and somehow the market translates this
into a notion of stability.
Don't get me wrong, there are some applications
for which I would pick Sun over linux, serving
NFS in a serious way would be one. Another would
be for running on Sun's really big hardware if
I had an application that needed that kind of
juice in a central place. Yet another would
be if I was looking at running a cache bound
application ( Sun has MUCH nicer cache than
you will find in PC hardware ).
But PC hardware is catching up fast.
In any senario in which Linux + x86 hardware
will do the job it will almost always do it
cheaper, and frequently better ( although not
always, see previous examples ).
Excuse me?
If anything we are LACKING sufficient computing to
allow for efficient operation of our businesses.
I recently went to refinance my car ( at a bank that will remain nameless ). This bank held my original car loan. I spent an hour filling out
paperwork ( all of which had been filled out with
my original loan ), having the loan officer call my
insurance company to get my insurance information
( even though they had all of my insurance info on
file ), etc. All of this totally redundant. I had to come back the NEXT day because the guy
the loan officer calls to do credit checks and fax
them to him was busy.
All of this should have been accomplishable over the
web. There is NO reason that it had to be that
hard.
Oh yes, and then they proceeded to automagically
debit my old loan payment ( several days after
the old loan had been paid off in full by the new
loan ) because it takes about a week for the PAPER
to work it's way through channels.
It took almost six weeks for the bank to
restore order to my account ( I will not recount the full ins and outs of that, but it was bad ).
All of this was unecessary. If they had proper computer systems handling the back end of the bank I should have been able to go to the web page for my account arrange refinancing there in under 10 minutes.
Note that every one of the bank employees I dealt
with had a computer on their desk. What made this
experience so inefficient ( and frustrating ) was
not their lack of computers, but the lack of competent back end systems for them to access with those computers. That is were the efficiency comes.
For graphing:
For Numerical Analysis:
-
GSL ( Gnu Scientific Library ):
language bindings for perl,python, and C++ for GSL are also available.Check out the Scientific Computing FAQ: which I've been having trouble reaching so you might want to try the Google cache of it.
Talk to some folks at Fermi Lab about how to
handle REALLY big data streams. Last I
heard they produce about 250GB/s of data
when the beam is on ( after the first stage
hardware filtering has reduced the data set ).
These are the folks to talk to if you have any
really heavy duty data needs.
This was first seriously proposed by
Gerald K. Oneill of Princeton University in
1975! It was feasable ( and even profitable )
then, but the capitalization was to high for
any organization on earth but the US Government
to undertake. The only reason we haven't done
it already is because of a defect of will, a
myopy of purpose, and inability to look further
ahead than the next election.
When will we, the citizens of the United States,
have the vision to demand these sorts of
projects from our government? Oneill's initial
proposal had an estimated 20 year pay back time, for
the first powersat. Subsequent powersats would
have been much cheaper. If the proposal Oneill
made had been taken up seriously in 1976, and taken
say 2 years to get it's political legs so that
actual work began in 1978, and it took ten
years to build, we would have had cheap abundant
energy by 1988.
Given cheap abundant energy it would be feasible to
produce, for example, metal hydride or fuel cell
powered cars. Given a 10 year ramp up and phase in
for those technologies we would have in 1998
been largely petroleum free ( at least for
power ).
Does anyone question that this would be a better
place to be... and we could be there by now, if
only we had the vision, and the will.
I've recently switched from Enlightenment/Gnome
as a desktop to KDE. Two major driving
factors where Konqueror and Konsole ( tabbed
shells... mmm... ). I had previously
used Galeon, and I prefer Konqueror in most
respects ( but please for the love of God bring
tabbed browsing to Konqueror!! ).
I still use Evolution for email
( kmail is just not as feature rich or usable )
and Gnumeric for spreadsheets. I would say that
Kmail and Kspread still need a fair bit more work
to catch up (particularily kmail).
Ah... but they have not in fact copyrighted the
numbers. They have copyrighted the musical
representation of these numbers as DTMF tones.
Additionally, like hell numbers aren't copyrightable.
What do you think an mp3 file is? It's a very
large number. In fact EVERYTHING digital is a
number. So if you can't copyright a number, how
then is software, source code, digital music,
digital video copyrightable?