Before the Newton, nobody knew what a PDA was or what it could do. It sucks being first, because no amount of market research can tell you what will actually sell (see the entry under "Walkman").
It's comments like these from smug hindsighters that kill the future. Have you ever been on the leading edge of anything? I think not.
Tags are item-specific, if you want them to be
on
NYT on RFID Tags
·
· Score: 3, Informative
The EPC spec has all those bits so the instances of objects can be tracked. An EPC is broken down into four sections:
bits 00-07 = header bits 08-35 = manufacturer (EPC Manager) bits 36-59 = Object Class bits 60-95 = Serial Number
There's another EPC, the Compact EPC, that's only 64 bits long, because the longer bit length translates into higher-cost tags.
So saying that RFID tags are -not- instance specific is incorrect. They can be (and the EPC is designed to be) instance specific, but it's up to the manufacturer.
Google sends a copy of all the indexed documents to the US Government for tracking, as well as an updated realtime list of everone who visits those pages. They also employ hackers to break into company intranets, suck as much confidential web content as they can using their high-tech web spiders, and sell that data to the Chinese.
If you're not a good programmer none of that will make sense. If you are, then it's stuff you already know. Go figure, huh?
As a programmer, the three most important things you need to be able to do are:
debugging optimization implementation
probably in that order. Note that there isn't much difference in the skillset needed to debug and to optimize. The basic techniques are the same, it's just the specific tools for each are a bit different. Both require that you keep the model of your thing in your head for extended periods of time, as well as the execution environment.
Implementation is easy once you look at it from a debugging and optimization standpoint. Design becomes a bottom-up process; whether that's good or bad is subjective, but it sure does make it easier to debug when a program is written to be easy to debug and optimize:)
The book is good, but it doesn't get into the really hard stuff - how does something get from point A to point B in a loosely coupled network, or any network?
His description of the neuron network in the brain, for example, talks about how some neurons link some parts of the brain with others, and that random links help the brain (and networks) function. But nowhere does he say how a signal actually gets from point A to point B - just that the loose coupling and random connections between brain areas make everything closer together.
. Maybe he doesn't know? Maybe nobody knows? But the whole point of the book is "connect tightly at the micro level, connect the micro groups with their immediate neighbors, and connect each micro grouping randomly with other non-local micro groupings for better connectivity."
My god, at this rate RAM production will consume all of Earth's resources! They must be stopped before Earth is transformed into a floating mass of DRAM!
Lets say you want too go steal a bunch of equipment from a lab in your university. Wouldn't you rather use a method that implicates someone on the inside (entering with no signs of forced entry)?
Plus, breaking in can attract attention. Why would you want to do that? Just turn your key and go.
Or, say, you want to break into your neighbor's apartment. Well now it's easy - just walk on in any time you want, and nobody'd be able to tell.
Or even better, you now have access to every office in your building. Go crazy in the HR department, and as a bonus the HR workers have their usernames and passwords in their desks for easy access.
Your attitude just shows a lack of imagination. It's a real concern if you live in the Real World.
What are you, deaf? On the bands in NYC there's a station almost every odd frequency or four.
neat - no upper limit discussed
on
Reflections
·
· Score: 1
The guy didn't want to say what the upper limit on BLAST would be - that implies that it might be substantially higher than 19.2MB/s, a lot higher.
The current BLAST implementation is designed to recycle existing equipment. If you designed your stuff from the ground up with BLAST in mind, would that imply a much higher throughput?
I wonder how the receiver puts the signals back together...
I'd guess retailers see these things as incredible timesavers. Imagine a wal-mart store - now imagine that consumers have left items all over the place.
How do you know something is out of place? Or where it's supposed to go?
Just counting inventory is a PITA.
Of course, the tags have to work, but if they do, they'll be an incredible time saver.
Apparently we're all wrong when it comes to the common ideas about residual radiation. Here are two links that seem appropriate:
[In regards to the radiation at Hiroshima and Nagasaki] "At present, more than 50 years after the bombings, one needs ultrasensitive equipment to measure the induced radioactivity"
http://www.rerf.or.jp/eigo/radefx/dosereco/resid ua l.htm
"an hour spent at the site of 113 nuclear explosions over a 40-year period ending in l989 has about the same negative health effects from radiation as a trip from San Francisco to New York in an ordinary jet airliner"
Do you really want to carry around another modem? If you're wireless, you've got 802.11b already, or will be getting bluetooth so you can hook into your mobile carrier's 3g/2.5g data network. Why in god's name would you want another access device/provider?
What I remember from those days is that pen computing was killed by FUD spread by Microsoft and Pen Windows. In those days, like the early days of the net, people were saying how Pen Computing was going to wipe out normal desktop computing. This would have left Microsoft SOL, so the FUD machine went to work and essentially killed the market by luring the gullible with the promise of commodity (cheaper) hardware running - tada! Pen Windows.
This, as well as a bunch of well-documented problems in some of the startups (read "Startup," it's a classic) dropped pen computing into the embedded/vertical application ghetto.
Will the tablet PC bring pen computing back? Not really...or if it does, well, too bad for everyone, because it's running Win32, the poster child for inappropriate code reuse. I can hear the conversation now:
"We need a new OS for a blender/coffeemaker combo."
And you base the "no precedent" on what data, exactly?
Too many errors on one line. Make fewer.
on
Gnarly Error Messages
·
· Score: 2, Funny
This is from the classic MPW C 68k compiler. There were lots more messages, most of which I've forgotten. Apparently the writer was an englishman with a truly droll sense of humor.
Apparently standard practice in the advertising industry is to pay everyone who appears in a commercial, no matter what the principal says. Those actors in commercials get paid too!
It's some sort of liability thing, because by being paid they have "consideration", in the legal sense of the word. Basically it means that they took the money, so they can't sue later.
Nice try, though! Maybe if Apple had actually made someone up out of whole cloth, you might have had a point. But unfortunately, everyone in the "Switch" campaign is real.
one hacker can write an "I Hate Saddam" virus and knock out the entire US Navy infrastructure by exploiting outlook/w2k problems! Talk about liability for EDS!
Let's face it, don't open source it, and you don't have enough $$ to patent it. Go to RSA, and give them a demo. RSA has its hooks into every encryption user. You, as a joe nobody, don't have the infrastructure needed to sell, market, develop, and support encryption.
The open source folks just want it for free. With RSA, you'll get paid. No question RSA would be better. The open source folks will bitch and moan, but that's what they do.
Take the $$ for the patent, and get a lawyer to NDA RSA.
One of the issues not addressed by Berman in the article or the bill itself is this: the mere presence of an item on a p2p network does not prove infringement. How are the copyright holders supposed to prove that someone is infringing on their copyright?
Most of the time, determination that a violation of law has taken place is done by the court system. Berman's bill has no provision for determining whether a given copyright is actually being infringed, nor does it specify a process that a copyright holder would go through to determine that a copyright is being infringed.
This and this alone should be enough to cause problems. By allowing private entities to determine and prosecute violations of law, the bill essentially places police power in the hands of private entities.
Write your congressman to complain! Most of them are lawyers, so should understand reasonable arguments. One plus is the bill is relatively short, so should be easy to comprehend.
Technically, almost every modern software program is a copy of something. There are only a few commercial applications that didn't exist before. A few examples:
Visicalc (spreadsheets) PageMaker (page layout) Mosaic (web browser) video games
Even free software isn't original. In the old days, it used to be called "Public Domain." The main difference is the FSF. Before, there was no organization publicizing "Public Domain", because, well, when you put something in the public domain you explicitly release all rights to it. That's no way to run an organization!
N1 is more than load balancing, it's treating your whole network of computers as a single, virtual, paritionable entity.
Step away from the network centric world that most of us live in, and go to the place where Real Work happens.
Imagine a database. Imagine the box that it runs on. OK, now imagine that database and all of its stuff not having to know that it's installed on any particular box - it can move itself across your installation to where there's extra CPU power, extra disk space, etc.
N1 is about virtualizing the hardware on which programs run. Or at least it should be.
Most programs really don't care about the underlying hardware - all they care about is access to their.conf files, data files, etc. They need to be able to locate and use resources, but it's only by happenstance that those resources are on a given box.
Think about it - if you're any good, you've standardized your filesystem layouts so they're identical across as many boxes as possible. That way you can log into a box and pretty much know what's there, without having to remember necessarily which box it is.
N1 takes that to another level, where the programs you run don't care or know what particular machine they're running on. You invoke gcc, but that gcc may run on a box miles away. Who cares where it runs, as long as its secure, low-latency, and it works?
The answer is that nobody cares.
Boxes are there as side-effect of the processes used today to add computing power to your infrastructure. Partitioning works within one box to create multiple virtual boxes, while N1 takes multiple physical boxes and creates one virtual box.
The benefit? Better resource utilization, and possibly better management.
Note: something like this is in plan9 already, if I remember correctly. It'll be interesting to see how N1 deals with code motion, etc.
well actually, the newspaper article said that they concluded that either c was changing or the 2nd law of thermodynamics was being violated. So they chose the first one, because it was socially acceptible.
But hey, what is everything if not a temporary violation of the 2nd law of thermodynamics?
Slashdot Mirror
Yeah, let's see how fast Dell moves off of those Tandem/Compaq Himalaya servers.
Blah blah blah blah.
Before the Newton, nobody knew what a PDA was or what it could do. It sucks being first, because no amount of market research can tell you what will actually sell (see the entry under "Walkman").
It's comments like these from smug hindsighters that kill the future. Have you ever been on the leading edge of anything? I think not.
The EPC spec has all those bits so the instances of objects can be tracked. An EPC is broken down into four sections:
- WH -002.pdfT -AUTOID-WH -008.pdf
bits 00-07 = header
bits 08-35 = manufacturer (EPC Manager)
bits 36-59 = Object Class
bits 60-95 = Serial Number
There's another EPC, the Compact EPC, that's only 64 bits long, because the longer bit length translates into higher-cost tags.
So saying that RFID tags are -not- instance specific is incorrect. They can be (and the EPC is designed to be) instance specific, but it's up to the manufacturer.
http://www.autoidcenter.org/research/MIT-AUTOID
http://www.autoidcenter.org/research/MI
Google sends a copy of all the indexed documents to the US Government for tracking, as well as an updated realtime list of everone who visits those pages. They also employ hackers to break into company intranets, suck as much confidential web content as they can using their high-tech web spiders, and sell that data to the Chinese.
If you're not a good programmer none of that will make sense. If you are, then it's stuff you already know. Go figure, huh?
:)
As a programmer, the three most important things you need to be able to do are:
debugging
optimization
implementation
probably in that order. Note that there isn't much difference in the skillset needed to debug and to optimize. The basic techniques are the same, it's just the specific tools for each are a bit different. Both require that you keep the model of your thing in your head for extended periods of time, as well as the execution environment.
Implementation is easy once you look at it from a debugging and optimization standpoint. Design becomes a bottom-up process; whether that's good or bad is subjective, but it sure does make it easier to debug when a program is written to be easy to debug and optimize
The book is good, but it doesn't get into the really hard stuff - how does something get from point A to point B in a loosely coupled network, or any network?
His description of the neuron network in the brain, for example, talks about how some neurons link some parts of the brain with others, and that random links help the brain (and networks) function. But nowhere does he say how a signal actually gets from point A to point B - just that the loose coupling and random connections between brain areas make everything closer together.
. Maybe he doesn't know? Maybe nobody knows? But the whole point of the book is "connect tightly at the micro level, connect the micro groups with their immediate neighbors, and connect each micro grouping randomly with other non-local micro groupings for better connectivity."
My god, at this rate RAM production will consume all of Earth's resources! They must be stopped before Earth is transformed into a floating mass of DRAM!
Lets say you want too go steal a bunch of equipment from a lab in your university. Wouldn't you rather use a method that implicates someone on the inside (entering with no signs of forced entry)?
Plus, breaking in can attract attention. Why would you want to do that? Just turn your key and go.
Or, say, you want to break into your neighbor's apartment. Well now it's easy - just walk on in any time you want, and nobody'd be able to tell.
Or even better, you now have access to every office in your building. Go crazy in the HR department, and as a bonus the HR workers have their usernames and passwords in their desks for easy access.
Your attitude just shows a lack of imagination. It's a real concern if you live in the Real World.
What are you, deaf? On the bands in NYC there's a station almost every odd frequency or four.
The guy didn't want to say what the upper limit on BLAST would be - that implies that it might be substantially higher than 19.2MB/s, a lot higher.
The current BLAST implementation is designed to recycle existing equipment. If you designed your stuff from the ground up with BLAST in mind, would that imply a much higher throughput?
I wonder how the receiver puts the signals back together...
I'd guess retailers see these things as incredible timesavers. Imagine a wal-mart store - now imagine that consumers have left items all over the place.
How do you know something is out of place? Or where it's supposed to go?
Just counting inventory is a PITA.
Of course, the tags have to work, but if they do, they'll be an incredible time saver.
Apparently we're all wrong when it comes to the common ideas about residual radiation. Here are two links that seem appropriate:
d ua l.htm
[In regards to the radiation at Hiroshima and Nagasaki]
"At present, more than 50 years after the bombings, one needs ultrasensitive equipment to measure the induced radioactivity"
http://www.rerf.or.jp/eigo/radefx/dosereco/resi
"an hour spent at the site of 113 nuclear explosions over a 40-year period ending in l989 has about the same negative health effects from radiation as a trip from San Francisco to New York in an ordinary jet airliner"
http://www.newaus.com.au/news4c.html
Do you really want to carry around another modem? If you're wireless, you've got 802.11b already, or will be getting bluetooth so you can hook into your mobile carrier's 3g/2.5g data network. Why in god's name would you want another access device/provider?
What I remember from those days is that pen computing was killed by FUD spread by Microsoft and Pen Windows. In those days, like the early days of the net, people were saying how Pen Computing was going to wipe out normal desktop computing. This would have left Microsoft SOL, so the FUD machine went to work and essentially killed the market by luring the gullible with the promise of commodity (cheaper) hardware running - tada! Pen Windows.
This, as well as a bunch of well-documented problems in some of the startups (read "Startup," it's a classic) dropped pen computing into the embedded/vertical application ghetto.
Will the tablet PC bring pen computing back? Not really...or if it does, well, too bad for everyone, because it's running Win32, the poster child for inappropriate code reuse. I can hear the conversation now:
"We need a new OS for a blender/coffeemaker combo."
"Why not use some MicroCruft?"
And you base the "no precedent" on what data, exactly?
This is from the classic MPW C 68k compiler. There were lots more messages, most of which I've forgotten. Apparently the writer was an englishman with a truly droll sense of humor.
Sorry, bzzzt!
Apparently standard practice in the advertising industry is to pay everyone who appears in a commercial, no matter what the principal says. Those actors in commercials get paid too!
It's some sort of liability thing, because by being paid they have "consideration", in the legal sense of the word. Basically it means that they took the money, so they can't sue later.
Nice try, though! Maybe if Apple had actually made someone up out of whole cloth, you might have had a point. But unfortunately, everyone in the "Switch" campaign is real.
one hacker can write an "I Hate Saddam" virus and knock out the entire US Navy infrastructure by exploiting outlook/w2k problems! Talk about liability for EDS!
Let's face it, don't open source it, and you don't have enough $$ to patent it. Go to RSA, and give them a demo. RSA has its hooks into every encryption user. You, as a joe nobody, don't have the infrastructure needed to sell, market, develop, and support encryption.
The open source folks just want it for free. With RSA, you'll get paid. No question RSA would be better. The open source folks will bitch and moan, but that's what they do.
Take the $$ for the patent, and get a lawyer to NDA RSA.
One of the issues not addressed by Berman in the article or the bill itself is this: the mere presence of an item on a p2p network does not prove infringement. How are the copyright holders supposed to prove that someone is infringing on their copyright?
Most of the time, determination that a violation of law has taken place is done by the court system. Berman's bill has no provision for determining whether a given copyright is actually being infringed, nor does it specify a process that a copyright holder would go through to determine that a copyright is being infringed.
This and this alone should be enough to cause problems. By allowing private entities to determine and prosecute violations of law, the bill essentially places police power in the hands of private entities.
Write your congressman to complain! Most of them are lawyers, so should understand reasonable arguments. One plus is the bill is relatively short, so should be easy to comprehend.
Only you can prevent Bad Law!
Technically, almost every modern software program is a copy of something. There are only a few commercial applications that didn't exist before. A few examples:
Visicalc (spreadsheets)
PageMaker (page layout)
Mosaic (web browser)
video games
Even free software isn't original. In the old days, it used to be called "Public Domain." The main difference is the FSF. Before, there was no organization publicizing "Public Domain", because, well, when you put something in the public domain you explicitly release all rights to it. That's no way to run an organization!
gnuts gto gthat!
Yes, you're missing something!
.conf files, data files, etc. They need to be able to locate and use resources, but it's only by happenstance that those resources are on a given box.
N1 is more than load balancing, it's treating your whole network of computers as a single, virtual, paritionable entity.
Step away from the network centric world that most of us live in, and go to the place where Real Work happens.
Imagine a database. Imagine the box that it runs on. OK, now imagine that database and all of its stuff not having to know that it's installed on any particular box - it can move itself across your installation to where there's extra CPU power, extra disk space, etc.
N1 is about virtualizing the hardware on which programs run. Or at least it should be.
Most programs really don't care about the underlying hardware - all they care about is access to their
Think about it - if you're any good, you've standardized your filesystem layouts so they're identical across as many boxes as possible. That way you can log into a box and pretty much know what's there, without having to remember necessarily which box it is.
N1 takes that to another level, where the programs you run don't care or know what particular machine they're running on. You invoke gcc, but that gcc may run on a box miles away. Who cares where it runs, as long as its secure, low-latency, and it works?
The answer is that nobody cares.
Boxes are there as side-effect of the processes used today to add computing power to your infrastructure. Partitioning works within one box to create multiple virtual boxes, while N1 takes multiple physical boxes and creates one virtual box.
The benefit? Better resource utilization, and possibly better management.
Note: something like this is in plan9 already, if I remember correctly. It'll be interesting to see how N1 deals with code motion, etc.
Geez, it's amusing how many comments here blast MS on the one hand, then bitch that MS removed its free fonts (that the same critics were using).
It just goes to show how ridiculous Linux advocates really are. Chhhhumpolas!
(Slap) I'm a user (Slap) I'm a developer (Slap) I'm a user (Slap) I'm a developer (Slap) I'm a user and a developer!
well actually, the newspaper article said that they concluded that either c was changing or the 2nd law of thermodynamics was being violated. So they chose the first one, because it was socially acceptible.
But hey, what is everything if not a temporary violation of the 2nd law of thermodynamics?