So a kid who walks up to your car and tries the doorhandles is not guilty of anything untoward?
I grew up in conservative Oklahoma. As a teenage kid, I was walking across a large parking lot with my friend and his girlfriend to a movie theater. My friend had long hair, so that probably tipped us off as obvious hoodlums, justifying some person calling the police to report "suspicious activity" of some kids messing with cars.
Maybe if we had been doing anything more than walking it would have been a good lesson. As it was, it just taught me the world definitely has scared, intolerant jackasses.
Before advocating low tolerance and hair-trigger fingers, consider the sociecty you're creating for everyone, not just the criminals and would-be criminals.
Do you really need thos icons at all ?
I just use Esc, F5, Backspace or the back mouse button and get rid of the buttons alltogether.
Then you're missing the drop down menus in back and forward. More than being faster for large jumps, I occasionally need them to get behind some pesky javascript that otherwise makes a single back operation useless.
There's lots of other subtle biases. For example, in the "grep" example, the sh code simply called grep. If he wanted to be pure about the scripting he'd not have had any way of doing a RE in sh...
"Pure"? What the... does that mean? The comparison is "how do you solve a problem in a given scripting environment". The comparison is not "what are the precise language features of each scripting language".
In shell, to find a substring in a file, you use grep, not some jackass case statements.
If you must, consider it as comparing the standard libraries between languages. In shell, the standard library is the canonical unix toolset.
... because once you single a number out for attention, it isn't random anymore.
Whoah. What?
Maybe I'm naive. But honestly, the only criteria I see for "random" is "utterly non-deterministic". Maybe I need enlightenment, but I see no justification for "random" to ever mean "unknowable" or "uncompressible".
My idea of a random number is that it, say, no deterministic process known or unknown created that number. There's no way a truly random number can be intentially recreated.
Furthermore, that doesn't mean that the process to create a random number cannot be known, just that it isn't in any way deterministic. And, it's perfectly reasonable to presume that a particular random number might be expressible as the outcome of a simple, determinisitic process -- but that process doesn't generate random numbers, just that one particular random number. This addresses the idea that a random number can still be compressed and known, but that it comes from a random process.
I have to strongly agree with the grandparent. Random numbers should be discussed in the context of process that creates them.
Knowledge for handling common situations is usually added to Compile itself instead of being coded in the script... It remains to be seen whether this experiment of gauging differently the tradeoff between flexibility and simplicity
Yes, simplicity is good, but only in the context of the whole system. Here, you're just shifting complexity from the per-package scripts to the overall Compile package itself -- creating a large, central, monolithic service.
Because it's centralized, over time, this is going to accumulate a lot crap and become opaque, obfuscated, and unmaintainable. Changes and maintenance to Compile will more significantly impact the contemporary set of recipes than, say, changes to Portage and ebuilds.
It's easy to apply a good idea, like "simplicity", in too narrow of a scope -- to the detrement of the overall design. Better to think about it as balance of "package maintainability", "system maintainability", "barrier of entry", etc.
Movie execs are known for their 'MORE, MORE, MORE!' mentality... can only work to drive down the price that the audience is willing to pay for the product.
Shhh.. don't tell them how to fix it. I like cheap things.
If you want to filter email politely, you must follow these rules.
One small quibble about a final point in those rules:
I reject almost any MIME attachment that could be Windows malware. Even.zip files now. I politely ask them to arrange with me another way to send it. (Sending binaries through email isn't such a good idea anyway; it's indirect, and base64 bloats files 50%.)
It's indirect? What's a good way to transfer binary files that is both direct and secure?... and archived with a personal note. One handy thing I do for large attachments is to upload them to a http server and send the link. But this is a pain in the ass for anything other than the biggest files. What are the good options otherwise?
The magic doesn't come from vectors. Vectors are just how you throw the numbers around
And your point is?
Ah, that's the main point. Both the article and your original post focus on the fact that vectors are being used. While true, this doesn't really impact the essense of the algorithm -- effectively addressing the lower-level data structures instead of the higher-level algorithms. Perhaps an analogy might be someone describing Google's search by explaining B-trees instead of getting into what process actually determines that one page is better than another for a given search.
I'm not going to address the finer details of image classification further than that the techniques you describe require a significant amount of preparation, selection, and manipulation up-front by a human before a computer can produce useful results. Rather, I used image classification as a motivation to describe why discussing only the notions of "vectors" and "clusters" misses a huge part of the story of what actually makes these sort of techniques work.
Ah, it uses vector math.... Image classification would be a no brainer for folks that spend their time examining images and multispectral datasets.
Ugh. The magic doesn't come from vectors. Vectors are just how you throw the numbers around. The reason the classification apparently works well is their choice of representation of the document: a word histogram -- the occurance count for each word. To measure the distance between two histograms, you usually use the chi-squared test. So, forget all about "vectors", the real work horse is the histogram. And, we can discuss about "clustering", but it's just as imporant to know how you're measuring the distance from one document to another.
Image clustering is hard, and the problem comes from picking a good representation of the image. Of course, a "word histogram" for an image makes no sense. Just considering pixel intensity or pixel color doesn't work either. You usually have to start looking at things like lines, curvatures, intersections, texture patterns, etc. Once you decide tools you're going to use to describe an image and algorithms to calculate them, you can starting talking about how far away one image is from another, which then naturally leads to clustering techniques. But, the hard part about the clustering is getting them into a space in which they actually, nicely cluster.
I had to stop reading the article because it was so clearly written by someone who had no comfort with the mathematical concepts or techniques. (Sorry, but seriously, it's the blind leading the blind.)
Actually we (Chipcon) launched a 802.15.4 compliant device in november last year.
And Moteiv is the first to ship a mote (wireless sensor node) using the CC2420, integrated PCB antenna, a TI MS430 microcontroller, and USB for reprogramming, debugging, and data collection.. all of course running TinyOS:).
Feh, nevermind the toaster oven, solder paste and a heat gun. I populated two boards yesterday like this. Just get some solder wick, too, because you'll most likely have to go back over and clean up any spurious solder bridges.
I found Quake3 to be one of the most accurate because...
But if you don't play any Quake3-based games, then it only amounts to a synthetic benchmark. When evaluating a card, you should have a set of applications (games you expect play), and you evaluate how the hardware you are interested in buying perform in those particular applications (games).
FWIW, I still play quake3, so those benchmarks do technically matter to me, except that everything since a geforce2 is more than sufficient for the game.
That's right. I perhaps incorrectly inferred that the parent expected increasing urban density to be categorically true as a city grows. Because of an existing pattern of urban sprawl, Houston can't significantly support an increase in urban density -- it'd just be a logjam of cars because everything is too far for walking.
When Emacs can do everything short of washing the dishes, it's every geeks favorite text editor. (Go vim!) But when Excel plays Pacman, it's a perversion of a spreadsheet (hell, for parallelism, "numeric editor").
As urban density increases, you'd think that the general short-haul travel requirements of individual citizens would diminish until their usual haunts ( the office, the local supermarket, the pub ) were within striking distance of their feet.
False. Counter-example: Houston.
Houston fact: Houston is the fourth most populous city in the
nation (trailing only New York, Los Angeles and Chicago), and is the largest in the southern U.S. and Texas.
It's a two mile drive (seriously) just to get out of brother's neighborhood, which is nothing but residential. That's a 30-40 minute walk just to get out of all the houses, and even then, nothing is nearby... probably another good hour walk to get to anything even resembling a Quikimart, which isn't exactly worth the three hour round trip.
Yes, it's true that some cities are or will be more convenient for foot traffic. But, that is neither categorically nor inevitably true for all cities. And, the people living in those cities not suited for foot travel would like a better way to get around short of tearing down, redesigning, and rebuilding 8,778 square miles of urban sprawl.
That's all real nice for the sake of the argument, but are there enough concrete examples of this happening to make it a valid argument? That assertion honestly feels as fluffy as some made by Microsoft, and I see it repeated again and again in this forum.
I favor an alternative approach, which the authors concede has some merits, but quickly dismiss, sender authentication.
If I start rejecting all email which is not from a verifiable sender, I'll quickly cut spam, and impose some costs onto those who wish to sent me email.
T. Loder, M. Van Alstyne, and R. Wash said:
A different, though promising approach is to incorporate
the use of strong identities using digital signatures [18]. This
allows authentication of unrecognized senders and explicit
granting of permission for email transactions. Authentication
has the advantage that it prevents "spoofing," deliberate
misuse of a third party identity to gain access, and it
will inevitably become part of any realistic solution.
However, there are a few difficulties with strong authentication
alone. The first is that the ease of obtaining new identities,
however verifiable, makes it possible to start over each
time the reputation capital of any given identity is spent beyond
repair. Friedman and Resnick [8] show that newcomers
will inevitably need to "pay their dues" in any open society
(one that does not charge per access) that has low cost
identities.
The paper doesn't exactly dismiss it, "... it will inevitably become part of any realistic solution." But, a number of general arguments made in the paper that apply:
1) Facilitating valuable exchange. To paraphrase their Key Insight: a model of exchange that encourages valuable communication will dominate a model that discourages unwanted communication. PKI only deters, not encourages, communication. The paper proposes a mechanism for the sender to sell their message to the recipient, promoting communication with value.
2) Technological arms race. With any technological solution, as spammers become more savvy, more machinery is emplaced to combat them. You concede this point yourself, "there would be an incentive for the spammers..., which we could counter with..." Where do the countermeasures end? Because the paper proposes an economic solution, there is no arms race but rather a free market on individual attention.
3) Individual (single person, single message) value. What I value in communication may differ from my neighbor. PKI concedes control to determine the value gained by the recipient of a message to a central authority that provides sender identities, presumably at some cost matched with the value of receipt of all messages to be sent in the lifetime of that identity. The paper proposes a system that effectively values each message (from unknown senders), not the identity of the sender.
I could cull a couple more arguments from the paper, but these are the strongest, and you get the idea, anyway. Note also that the authors concede the use of white lists so that known senders need not warranty their messages.
... Moore's Law. So long as components keep on falling in price,...
Ugh. That does not even remotely resemble Moore's Law, "the doubling of transistors every couple of years". Further, Moore's Law does not apply to the embeded domain. Whereas the driving motivation for desktops and supercomputers is nearly exclusively computation power, a huge motivation for wireless embeded devices is energy consumption and lifetime. Moore's Law makes no assertion about the power efficency of computers, and yes, they're getting more hungry, not less -- go look at the size of your heat sink (!) and fan (!!) on your CPU and imagine putting those in your phone (!!!).
You can't abstract Moore's Law to "falling prices" and apply it to embeded devices. Fewer transistors and less memory consume less energy than their more beefy counterparts. When designing and selecting the hardware and software in this domain, device lifetime must be considered and balanced against the processor power and available memory.
Re: I have some problems with at least ONE of his.
on
Make More Mistakes
·
· Score: 1
That's pure unadulterated bull- and anyone with business sense KNOWS this....
A market with no competition may not be a market afterall, but it may also be an untapped, unsold to market as well.
He was trying to be too cute with his lessons. If you read the story leading up to that rule, you find the real lesson in the last sentence, "RADish was designed to solve a problem that almost nobody has."
So, I'd say the real lesson from his story is "Don't try to sell solutions to problems that don't exist."... which is just plain not a market at all, not one that exists with no competition.
"ATMega128 7.3 MHz microcontroller" -- The MICA has an ATMega128 clocked at 4MHz (and the original version had an ATMega103).
"433 MHz wireless radio, CC1000 transciver" -- The MICA has a 433MHz RFM radio. The MICA2 comes with either a 433MHz or 916MHz Chipcon 1000 radio.
Though, corrections with respect to the MICA are not especially significant, since you'd really rather be working with the MICA2.
Even if they can get the size down to "dust", the antenna HAS TO BE 1/4 of the radio wave length. For the 433 MHz version this is around 8-10 inches! So these dust motes will have "tails". -- An antenna for 916MHz is half that size. An antenna for a 2.4GHz radio is small enough to be surface mounted onto a device the size of the MICA2.
To be precise, the company creating the dust mote is Dust Inc located in Berkeley, not UC Berkeley itself.
... I'd be quite happy to be restored from backup - especially if it was that or nothing.
Just to clarify: you'd be quite happy to be restored from backup in the same way you'd be quite happy to have your children persist beyond your death, correct?
Because, to extend the continuity argument to the information age: assuming there is a "you" that exists to sense what you sense (what you see, hear, touch, taste, smell), if there are simultaneously two of you, which one is *you*? Possible answers: A) neither, our sense of self is an illusion, B) just the original, self is defined beyond the informational or physical content of the body and brain, C) just the duplicate, I can't think of a good argument, but it does exist as an option, D) both, our sense of self is not explained by current science and there exists an aggregation of senses between entities, whereby "you" are both.
So, which one do you feel is most likely given your quite happiness in the duplicate? Or, even better, what other options do you fell there are to define self in this context of two simultaneous "yous"?
Slashdot Mirror
I grew up in conservative Oklahoma. As a teenage kid, I was walking across a large parking lot with my friend and his girlfriend to a movie theater. My friend had long hair, so that probably tipped us off as obvious hoodlums, justifying some person calling the police to report "suspicious activity" of some kids messing with cars.
Maybe if we had been doing anything more than walking it would have been a good lesson. As it was, it just taught me the world definitely has scared, intolerant jackasses.
Before advocating low tolerance and hair-trigger fingers, consider the sociecty you're creating for everyone, not just the criminals and would-be criminals.
Then you're missing the drop down menus in back and forward. More than being faster for large jumps, I occasionally need them to get behind some pesky javascript that otherwise makes a single back operation useless.
Though that doesn't address if the negative response is justified or not. In this case, Winstripe (the new default theme) is immature at best.
"Pure"? What the ... does that mean? The comparison is "how do you solve a problem in a given scripting environment". The comparison is not "what are the precise language features of each scripting language".
In shell, to find a substring in a file, you use grep, not some jackass case statements.
If you must, consider it as comparing the standard libraries between languages. In shell, the standard library is the canonical unix toolset.
Jesus christ. Did you just return a pointer to a stack allocated buffer? Oh. My. God. Please burn this C code and forget you ever wrote it.
Whoah. What?
Maybe I'm naive. But honestly, the only criteria I see for "random" is "utterly non-deterministic". Maybe I need enlightenment, but I see no justification for "random" to ever mean "unknowable" or "uncompressible".
My idea of a random number is that it, say, no deterministic process known or unknown created that number. There's no way a truly random number can be intentially recreated.
Furthermore, that doesn't mean that the process to create a random number cannot be known, just that it isn't in any way deterministic. And, it's perfectly reasonable to presume that a particular random number might be expressible as the outcome of a simple, determinisitic process -- but that process doesn't generate random numbers, just that one particular random number. This addresses the idea that a random number can still be compressed and known, but that it comes from a random process.
I have to strongly agree with the grandparent. Random numbers should be discussed in the context of process that creates them.
Yes, simplicity is good, but only in the context of the whole system. Here, you're just shifting complexity from the per-package scripts to the overall Compile package itself -- creating a large, central, monolithic service.
Because it's centralized, over time, this is going to accumulate a lot crap and become opaque, obfuscated, and unmaintainable. Changes and maintenance to Compile will more significantly impact the contemporary set of recipes than, say, changes to Portage and ebuilds.
It's easy to apply a good idea, like "simplicity", in too narrow of a scope -- to the detrement of the overall design. Better to think about it as balance of "package maintainability", "system maintainability", "barrier of entry", etc.
Shhh.. don't tell them how to fix it. I like cheap things.
One small quibble about a final point in those rules:
It's indirect? What's a good way to transfer binary files that is both direct and secure? ... and archived with a personal note. One handy thing I do for large attachments is to upload them to a http server and send the link. But this is a pain in the ass for anything other than the biggest files. What are the good options otherwise?
And your point is?
Ah, that's the main point. Both the article and your original post focus on the fact that vectors are being used. While true, this doesn't really impact the essense of the algorithm -- effectively addressing the lower-level data structures instead of the higher-level algorithms. Perhaps an analogy might be someone describing Google's search by explaining B-trees instead of getting into what process actually determines that one page is better than another for a given search.
I'm not going to address the finer details of image classification further than that the techniques you describe require a significant amount of preparation, selection, and manipulation up-front by a human before a computer can produce useful results. Rather, I used image classification as a motivation to describe why discussing only the notions of "vectors" and "clusters" misses a huge part of the story of what actually makes these sort of techniques work.
Ugh. The magic doesn't come from vectors. Vectors are just how you throw the numbers around. The reason the classification apparently works well is their choice of representation of the document: a word histogram -- the occurance count for each word. To measure the distance between two histograms, you usually use the chi-squared test. So, forget all about "vectors", the real work horse is the histogram. And, we can discuss about "clustering", but it's just as imporant to know how you're measuring the distance from one document to another.
Image clustering is hard, and the problem comes from picking a good representation of the image. Of course, a "word histogram" for an image makes no sense. Just considering pixel intensity or pixel color doesn't work either. You usually have to start looking at things like lines, curvatures, intersections, texture patterns, etc. Once you decide tools you're going to use to describe an image and algorithms to calculate them, you can starting talking about how far away one image is from another, which then naturally leads to clustering techniques. But, the hard part about the clustering is getting them into a space in which they actually, nicely cluster.
I had to stop reading the article because it was so clearly written by someone who had no comfort with the mathematical concepts or techniques. (Sorry, but seriously, it's the blind leading the blind.)
And Moteiv is the first to ship a mote (wireless sensor node) using the CC2420, integrated PCB antenna, a TI MS430 microcontroller, and USB for reprogramming, debugging, and data collection.. all of course running TinyOS :).
Feh, nevermind the toaster oven, solder paste and a heat gun. I populated two boards yesterday like this. Just get some solder wick, too, because you'll most likely have to go back over and clean up any spurious solder bridges.
Yay, they've also updated the graphical smileys from totally awful to reasonably attractive. Much better.
But if you don't play any Quake3-based games, then it only amounts to a synthetic benchmark. When evaluating a card, you should have a set of applications (games you expect play), and you evaluate how the hardware you are interested in buying perform in those particular applications (games).
FWIW, I still play quake3, so those benchmarks do technically matter to me, except that everything since a geforce2 is more than sufficient for the game.
That's right. I perhaps incorrectly inferred that the parent expected increasing urban density to be categorically true as a city grows. Because of an existing pattern of urban sprawl, Houston can't significantly support an increase in urban density -- it'd just be a logjam of cars because everything is too far for walking.
When Emacs can do everything short of washing the dishes, it's every geeks favorite text editor. (Go vim!) But when Excel plays Pacman, it's a perversion of a spreadsheet (hell, for parallelism, "numeric editor").
False. Counter-example: Houston.
It's a two mile drive (seriously) just to get out of brother's neighborhood, which is nothing but residential. That's a 30-40 minute walk just to get out of all the houses, and even then, nothing is nearby... probably another good hour walk to get to anything even resembling a Quikimart, which isn't exactly worth the three hour round trip.
Yes, it's true that some cities are or will be more convenient for foot traffic. But, that is neither categorically nor inevitably true for all cities. And, the people living in those cities not suited for foot travel would like a better way to get around short of tearing down, redesigning, and rebuilding 8,778 square miles of urban sprawl.
Ok, I'll ask. What's a "prototype-based object-oriented" language and how does it differ from C++ and Java?
1) Facilitating valuable exchange. To paraphrase their Key Insight: a model of exchange that encourages valuable communication will dominate a model that discourages unwanted communication. PKI only deters, not encourages, communication. The paper proposes a mechanism for the sender to sell their message to the recipient, promoting communication with value.
2) Technological arms race. With any technological solution, as spammers become more savvy, more machinery is emplaced to combat them. You concede this point yourself, "there would be an incentive for the spammers ..., which we could counter with ..." Where do the countermeasures end? Because the paper proposes an economic solution, there is no arms race but rather a free market on individual attention.
3) Individual (single person, single message) value. What I value in communication may differ from my neighbor. PKI concedes control to determine the value gained by the recipient of a message to a central authority that provides sender identities, presumably at some cost matched with the value of receipt of all messages to be sent in the lifetime of that identity. The paper proposes a system that effectively values each message (from unknown senders), not the identity of the sender.
I could cull a couple more arguments from the paper, but these are the strongest, and you get the idea, anyway. Note also that the authors concede the use of white lists so that known senders need not warranty their messages.
Ugh. That does not even remotely resemble Moore's Law, "the doubling of transistors every couple of years". Further, Moore's Law does not apply to the embeded domain. Whereas the driving motivation for desktops and supercomputers is nearly exclusively computation power, a huge motivation for wireless embeded devices is energy consumption and lifetime. Moore's Law makes no assertion about the power efficency of computers, and yes, they're getting more hungry, not less -- go look at the size of your heat sink (!) and fan (!!) on your CPU and imagine putting those in your phone (!!!).
You can't abstract Moore's Law to "falling prices" and apply it to embeded devices. Fewer transistors and less memory consume less energy than their more beefy counterparts. When designing and selecting the hardware and software in this domain, device lifetime must be considered and balanced against the processor power and available memory.
He was trying to be too cute with his lessons. If you read the story leading up to that rule, you find the real lesson in the last sentence, "RADish was designed to solve a problem that almost nobody has."
So, I'd say the real lesson from his story is "Don't try to sell solutions to problems that don't exist."... which is just plain not a market at all, not one that exists with no competition.
Some quick corrections to your list.
"ATMega128 7.3 MHz microcontroller" -- The MICA has an ATMega128 clocked at 4MHz (and the original version had an ATMega103).
"433 MHz wireless radio, CC1000 transciver" -- The MICA has a 433MHz RFM radio. The MICA2 comes with either a 433MHz or 916MHz Chipcon 1000 radio.
Though, corrections with respect to the MICA are not especially significant, since you'd really rather be working with the MICA2.
Even if they can get the size down to "dust", the antenna HAS TO BE 1/4 of the radio wave length. For the 433 MHz version this is around 8-10 inches! So these dust motes will have "tails". -- An antenna for 916MHz is half that size. An antenna for a 2.4GHz radio is small enough to be surface mounted onto a device the size of the MICA2.
To be precise, the company creating the dust mote is Dust Inc located in Berkeley, not UC Berkeley itself.
Just to clarify: you'd be quite happy to be restored from backup in the same way you'd be quite happy to have your children persist beyond your death, correct?
Because, to extend the continuity argument to the information age: assuming there is a "you" that exists to sense what you sense (what you see, hear, touch, taste, smell), if there are simultaneously two of you, which one is *you*? Possible answers: A) neither, our sense of self is an illusion, B) just the original, self is defined beyond the informational or physical content of the body and brain, C) just the duplicate, I can't think of a good argument, but it does exist as an option, D) both, our sense of self is not explained by current science and there exists an aggregation of senses between entities, whereby "you" are both.
So, which one do you feel is most likely given your quite happiness in the duplicate? Or, even better, what other options do you fell there are to define self in this context of two simultaneous "yous"?