I tell you what.. pour a few years of your life, lots of your own money, and all your ambition into inventing, developing, proving, and finally patenting your new invention, and then come back to us about "thinking things up".
or is it
Especially when it can cost so little to develop them.
Is the cost to patent something enourmous or is it trivial?
Calling me a fool? for what? I said that all the frivolous patents may destroy the patent system. How exactly are you interpreting that?
btw, here is the FUD:
As I said, a lot of patents are junk. They are the ones that get tossed out of court on a daily basis.
Patents are not often tossed out. In fact patents are rarely even challenged since the system is set up such that one has to prove ones innocence. Most patent cases are settled out of court, at enourmous cost to the consumer since they ultimately pay with increased prices.
And more of your double speak, oh I'm sorry "word tossing":
Jet engines are not per sepatented - rather, a single implementation and its thousands of pieces are patented.
What exactly are you saying? That "Jet engines" are not patented, but "a jet engine" is patented?
I tell you what.. pour a few years of your life, lots of your own money, and all your ambition into inventing, developing, proving, and finally patenting your new invention, and then come back to us about "thinking things up".
If you truly felt this way you would be more pissed off about all these frivolous patents as they dilute, and may possibly destroy the patent system.
And lets clear some other fud up okay, there are no patents on things with 4000 moving parts. You don't patent a jet engine. You patent all the things which make a jet engine possible.
Also, Japan has a patent system that is ridiculously out of control, it makes the US system look svelte.
Don't talk of asian countries in general, if you mean China say it.
I also do not see how you can possibly talk about patents taking years of toil and then talk about a task bar as an example. It was also probably done by one person and six other people signed it to boost their resumes. Happens the same with scientific papers.
The only positive thing about patents are that in maybe twenty years all the bullshit will have their patents expire and life can resume normally once again. It's just fucking sad that civilization will have this crutch for the next two decades.
I've been wanting to (re)try doom for awhile now, but there is a ton of stuff nowadays. You wouldn't happen to know of any doom-distros would you? I'm pretty sure the Linux community would be better served with a few less linux distros in favor of some category specific distros.
I recognize this argument, although not in relation to PDL. What is needed is the equivalent of an iso distribution except that the distribution would contain all the various math packages et al and run on any distro. Something like texlive for LaTeX.
In my experience, the document vectors containing the terms et al present must be stored in a sparse vector. The reason is that as new terms and such get added on it would require growing all exisiting vectors, only to be filled with zeros.
I believe the islands of pages are ignored because of abuse, you basically write some scripts that create an assload of interconnected webpages, thereby making the pagerank go up. Another way to think of it is, if the pages are all correlated (i.e., one source), then each page doesn't add as much information to the equation as an independently authored page would, thus you should discard them.
To me the biggest problem is the lack of contextual understanding. If the search engine could differentiate computer generated gibberish then they could eliminate most of the abuse going on.
My other peve is doing a search and just getting some guys list of random links containing my search terms in no relation to each other.
I am unfamilar with the Porter algorithm. Got a good link handy?
Sadly that is often the case. History is riddled with examples of mathematicians duplicating discoveries. That is my motivation for studying these types of techniques, a way to better compare scientific papers, to root out similarities. The history of the wavelet is a good example to discuss. The theory was almost hit upon several times, in slightly different fashions but the researchers involved never knew that others were making similar discoveries.
There are indeed many things that can be done with SOMs, but to get really good, human-like, results is quite difficult. The problem expands when you need to consider different words for the same thing (e.g., cat/feline), and exclude words that appear identical, but aren't. Then the need to consider context becomes increasingly important (e.g., this article is not about cats).
The complexity builds quite quickly.
All I said is that there is no reason for MS to know what apps you have.
Your reply starts debunking the phrase "all programs installed on the computer", which I care not about.
Irregardless, the argument you put forth is silly, MS keeps a tally of everything you install as you install it. You even mention it, the "Add/Remove Programs" panel.
Why does it not make sense? Since when did MS start distributing patches for QuickTime, or Adobe Photoshop, or any other non MS program?
It is your argument that does not make sense. Also, I have to point out that your posts are getting increasingly disjointed. Stretching ourselves a little thin perhaps?
If you do a full analysis, you'll find that the easiest and most reliable solution is to simply use a laptop. That's why this market will never take off. Especially as pdas and such are even easier than the laptop, and perform a lot of the tasks that you would've needed a laptop for in a car.
That is definately not what happens. A car battery has so much juice available that any wire that is shorting the terminals instantly turns into a line of fire with possiblity a plasma arc. The explosion happens if your battery is leaking hydrogen or there is gas leaking and accumulating under the hood (both rather unlikely). And even if did happen, you'd probably only get a nifty fireball then nothing.
I knew a guy once who spent thousands on a fancy hydralic system for his car sans the fuses. One day the power cables all ignited and the entire interior of his car caught fire. Pretty funny.
You assert that a MB cannot handle the spikes and such from the car's 12VDC, so you say, use an AC inverter instead. So why is it that this inverter can handle the spikes? Why wouldn't the same spikes destory the AC inverter? Whatever they use you could also use for a DC-DC converter. In all cases, you need to filter the power supply properly and there is no guarantee that the AC inverters (much less cheap ones) will protect anything from spikes.
Skimming the article, it seems to be quite similar to Self-Organizing Maps or Vector Quantizers. Care to comment?
Also, I've worked on SOMs previously and there is no way you can hold all the document vectors in RAM. Think google, when you do a search on 3 million documents. The part that really hurts is the number of possible keywords.
There is a huge difference between getting updates pushed to you versus having to poll for changes, especially with a server-challenged service such as blogger was/is. Not to mention that you can get much more accurate timining information.
You also must not read blogs very often, when the action is hot and heavy like it has been with iraq/inspections/etc.. many blogs start having conversations that jump from blog to blog to blog. If you and a couple hundred "perl guys" were to try track the action in "real time" I'm sure you would collectively cause enough bandwidth abuse to harm the very bloggers you are trying to read.
You're missing the point. If you utilized/.'s "highly-evolved Meta Moderation system" then you could easily mark all the low ID's as foes. Over time you get to the point where you should be able to trust any low IDs left (as the foes get massive penalties, ensuring you never see them again). Eventually those people would have to abandon their accounts and get shiny new IDs, as that is the only escape from a miserable karma with a ton of foes!
Not to mention the foe of friends, currently I don't penalize them, but I always scrutinize those comments a bit more than usual.
Perform an experiment sometime, save a couple of older stories at -1, then using grep, gawk, sort, unique, etc..., plot the distribution of user IDs. The number of posts coming from sub 100k is quite small.
PS I wish you had linked the posts you were referring to as I haven't seen any that match that description.
Ah, I was waiting for someone to post about that comment...
To which the reply is, evolved merely means that it is has gradually changed over time. Nothing in there says that the result of evolution is necessarily better or worse. This is by far the most frequently mistaken belief of anti-evolution religious zealots and newbie genetic algorithm afficionados.
Google probably bought them out of frustration of being unable to crawl the blog sites. Seriously, blogger has been having a crap load of server problems and many people who have lots of traffic were leaving it. Blogger was losing their reputation as being "the" blog place, they desperately needed something like google's server savy.
Most dotcom era hippies lauded the advent of the Internet as a strike against The Man and everyone would be a publisher.
This was and is a true statement, everyone can be a publisher. The problem is ego basically, what fun is it being a publisher if no one reads your stuff? Zipf's law is a distribution of the readers.
Your argument doesn't make sense, if you're talking about loading pixelmaps
from PNGs, then you could just as easily load pixelmaps from SVGs, with the
benefit that all required resolutions are automatically supported.
Also, the comment about having a postscript file that e.g. is of an infinite
resolution fractal is ridiculous. We are comparing two technologies attempting
to accomplish similar tasks, not how stupidly one could implement something. If
you want to go that route, then how slow would it be if I generated my icons at
1e9 x 1e9 pixels, scaling to the appropriate size on loading? Or is loading time
(and resources) really irrelevant?
It also doesn't take into account the task at hand. Suppose you're doing flow
charts, then having many images of arrows at various orientations would quickly
become obnoxious. Not to mention the task of combining the image of the arrow
with the rest of the elements on the screen, you cannot merely blit it to
screen.
Interesting comment. It seems as though you are assuming that an svg image would be composed of multiple objects. Perhaps as things evolve we'll have svg-squishers that optimize an image by removing occluded objects when it is efficient to do so.
The other aspect that needs to be mentioned is that comparing to PNG is rather ambiguous since it has more than one way of performing compression. The catch is that as PNG has tried to squeeze their file sizes further, the complexity of the encoder/decoder has subsequently increased. That all leads to slower performance. In fact, isn't one of the goals of image analysis to break up the scenes into it's base objects, i.e., turn a bitmap image into a vector image?
What I'd really like to know is how fast svg is compared to gif.
No one will ever need more than a 128x128 pixel icon.
IBM made a 400 dpi monitor once and said they had no plans on marketing it because MS Windows couldn't scale properly. The old 32x32 pixel icons would have been about 20 mm square. Think you can still click on it? To be fair, the larger 128x128 icons are almost a centimeter.
The goal is to one day have tablet computers with screens that have resolutions of around 600dpi so that they look like paper. All this talk of antialiasing is merely a method to alleviate the lack of a proper resolution. It's kind of like disk doubler in the old days of teeny harddrives.
When the floor is off kilter, that is the fault of the construction worker. An architects job is to design everything, knowing and compensating for the inevitable margins of error that will occur during "manufacturing".
A Software "Engineer" is someone who slaps some code up and says fuck it, the beta testers (read: customers) will find the bugs.
That's a bad example since according to his bio, he had a degree in Bachelor of Engineering Sciences, and was thus actually an engineer and not some software "engineer" who somehow graduates without taking any math courses.
Slashdot Mirror
Calling me a fool? for what? I said that all the frivolous patents may destroy the patent system. How exactly are you interpreting that?
btw, here is the FUD:
Patents are not often tossed out. In fact patents are rarely even challenged since the system is set up such that one has to prove ones innocence. Most patent cases are settled out of court, at enourmous cost to the consumer since they ultimately pay with increased prices.And more of your double speak, oh I'm sorry "word tossing":
What exactly are you saying? That "Jet engines" are not patented, but "a jet engine" is patented?And lets clear some other fud up okay, there are no patents on things with 4000 moving parts. You don't patent a jet engine. You patent all the things which make a jet engine possible.
Also, Japan has a patent system that is ridiculously out of control, it makes the US system look svelte. Don't talk of asian countries in general, if you mean China say it.
I also do not see how you can possibly talk about patents taking years of toil and then talk about a task bar as an example. It was also probably done by one person and six other people signed it to boost their resumes. Happens the same with scientific papers.
The only positive thing about patents are that in maybe twenty years all the bullshit will have their patents expire and life can resume normally once again. It's just fucking sad that civilization will have this crutch for the next two decades.
I've been wanting to (re)try doom for awhile now, but there is a ton of stuff nowadays. You wouldn't happen to know of any doom-distros would you? I'm pretty sure the Linux community would be better served with a few less linux distros in favor of some category specific distros.
Many packages also need lapack, fftw, etc....
I believe the islands of pages are ignored because of abuse, you basically write some scripts that create an assload of interconnected webpages, thereby making the pagerank go up. Another way to think of it is, if the pages are all correlated (i.e., one source), then each page doesn't add as much information to the equation as an independently authored page would, thus you should discard them.
To me the biggest problem is the lack of contextual understanding. If the search engine could differentiate computer generated gibberish then they could eliminate most of the abuse going on.
My other peve is doing a search and just getting some guys list of random links containing my search terms in no relation to each other.
I am unfamilar with the Porter algorithm. Got a good link handy?
As far as SOM stuff, Kohonen is the man.
There are indeed many things that can be done with SOMs, but to get really good, human-like, results is quite difficult. The problem expands when you need to consider different words for the same thing (e.g., cat/feline), and exclude words that appear identical, but aren't. Then the need to consider context becomes increasingly important (e.g., this article is not about cats). The complexity builds quite quickly.
Free means free, not "Free, except under the following conditions ..."
All I said is that there is no reason for MS to know what apps you have.
Your reply starts debunking the phrase "all programs installed on the computer", which I care not about.
Irregardless, the argument you put forth is silly, MS keeps a tally of everything you install as you install it. You even mention it, the "Add/Remove Programs" panel.
Why does it not make sense? Since when did MS start distributing patches for QuickTime, or Adobe Photoshop, or any other non MS program?
It is your argument that does not make sense. Also, I have to point out that your posts are getting increasingly disjointed. Stretching ourselves a little thin perhaps?
If you do a full analysis, you'll find that the easiest and most reliable solution is to simply use a laptop. That's why this market will never take off. Especially as pdas and such are even easier than the laptop, and perform a lot of the tasks that you would've needed a laptop for in a car.
I knew a guy once who spent thousands on a fancy hydralic system for his car sans the fuses. One day the power cables all ignited and the entire interior of his car caught fire. Pretty funny.
You assert that a MB cannot handle the spikes and such from the car's 12VDC, so you say, use an AC inverter instead. So why is it that this inverter can handle the spikes? Why wouldn't the same spikes destory the AC inverter? Whatever they use you could also use for a DC-DC converter. In all cases, you need to filter the power supply properly and there is no guarantee that the AC inverters (much less cheap ones) will protect anything from spikes.
The bottom line is that there is zero, nada, zilch, zippo reason for MS to know what apps you have.
Also, I've worked on SOMs previously and there is no way you can hold all the document vectors in RAM. Think google, when you do a search on 3 million documents. The part that really hurts is the number of possible keywords.
You also must not read blogs very often, when the action is hot and heavy like it has been with iraq/inspections/etc.. many blogs start having conversations that jump from blog to blog to blog. If you and a couple hundred "perl guys" were to try track the action in "real time" I'm sure you would collectively cause enough bandwidth abuse to harm the very bloggers you are trying to read.
Not to mention the foe of friends, currently I don't penalize them, but I always scrutinize those comments a bit more than usual.
Perform an experiment sometime, save a couple of older stories at -1, then using grep, gawk, sort, unique, etc..., plot the distribution of user IDs. The number of posts coming from sub 100k is quite small.
PS I wish you had linked the posts you were referring to as I haven't seen any that match that description.
To which the reply is, evolved merely means that it is has gradually changed over time. Nothing in there says that the result of evolution is necessarily better or worse. This is by far the most frequently mistaken belief of anti-evolution religious zealots and newbie genetic algorithm afficionados.
Google probably bought them out of frustration of being unable to crawl the blog sites. Seriously, blogger has been having a crap load of server problems and many people who have lots of traffic were leaving it. Blogger was losing their reputation as being "the" blog place, they desperately needed something like google's server savy.
Since your entire comment was redundant you should have similarly refrained from posting.
Also, the comment about having a postscript file that e.g. is of an infinite resolution fractal is ridiculous. We are comparing two technologies attempting to accomplish similar tasks, not how stupidly one could implement something. If you want to go that route, then how slow would it be if I generated my icons at 1e9 x 1e9 pixels, scaling to the appropriate size on loading? Or is loading time (and resources) really irrelevant?
It also doesn't take into account the task at hand. Suppose you're doing flow charts, then having many images of arrows at various orientations would quickly become obnoxious. Not to mention the task of combining the image of the arrow with the rest of the elements on the screen, you cannot merely blit it to screen.
The other aspect that needs to be mentioned is that comparing to PNG is rather ambiguous since it has more than one way of performing compression. The catch is that as PNG has tried to squeeze their file sizes further, the complexity of the encoder/decoder has subsequently increased. That all leads to slower performance. In fact, isn't one of the goals of image analysis to break up the scenes into it's base objects, i.e., turn a bitmap image into a vector image?
What I'd really like to know is how fast svg is compared to gif.
The goal is to one day have tablet computers with screens that have resolutions of around 600dpi so that they look like paper. All this talk of antialiasing is merely a method to alleviate the lack of a proper resolution. It's kind of like disk doubler in the old days of teeny harddrives.
A Software "Engineer" is someone who slaps some code up and says fuck it, the beta testers (read: customers) will find the bugs.
That's a bad example since according to his bio, he had a degree in Bachelor of Engineering Sciences, and was thus actually an engineer and not some software "engineer" who somehow graduates without taking any math courses.
I expect reapeatability and predictability from my coffee maker, ergo coffee is a scientific experiment.