Hey Joe, that moron is giving out conflicting info regarding the phone number listed on that page, and the voice sounds different when we call versus when we get a call. Check it out, will you?
K. Beep-boop-beep.
OMFG STOP CALLING ME.
Wait, I'm from Yahoo!. I called from my cellphone, and I get routed to you. When I call from the office phone, I get routed to the other person. HMMMM.
Entries that don't get reformed well won't OCR well. Entries that don't get OCRd well can automatically be flagged for manual inspection.
You can have a confidence interval based on the type of text. A novel will be all text and you would expect it to OCR perfectly. A coffee table book will have mostly photos that won't be OCRd. A text book will have symbols that will fuck up OCR.
Yes, what I suggested was simplistic. But it's not bullshit. It's the base idea on which you can build your tools. My suggestion is almost entirely software based - tripod, camera, book. My idea is cheaper, quicker, and far more distribution-friendly (even when you ignore the patents). The main idea is simple, and implementation can be as complex as you want it to be.
Google models each page in (stereoscopic) 3D. My idea models a few pages, and intelligently adjusts the others, knowing they'll be nearly identical.
Until Google ramps their tech up to include being able to scan pop-up books, the two methods will achieve nearly identical results.
Do you think all Google is doing is using it's magic infrared cameras? Do you think there is no quality control, or that they don't have a problem? We've already seen pages come out where we see the flipper's hands (with cute pink latex gloves no less).
I know they do - that's why you take a pic of the first and last pages first, and adjust those.
The 3D model for all pages is the same - a sheet of paper of a certain length and width.
The lay of the paper will be between the two extremes of the first and last pages.
Effectively, you can define the lay of the paper as a simple curve in the x/y plane.
The last page will be a flat line, the first page will be the most eccentrically curved.
Page 3(4) lays almost identical to page 1(2). The curve is just a little flatter. Page 699(700) lays almost identical to page 701(702).
Just tweak the first and last pages, and maybe some in the middle if you've got fucked up pages, books with different paper stock (maybe you're scanning Independence day and it's got stills from the movie in there on photo stock), etc.
It's cheaper, it's simpler, and the task can be easily distributed to other libraries.
Can't you just calculate the 3D model of the page based on a known stuff?
Make a generic flattener filter that takes in page height and length, as well as page number.
Manually tweak the output a bit for the first and last pages, and then intermediary pages can all be calculated with much more accuracy than you need.
Hell, with this method any book "scanned" (using a camera from overhead) could be processed. Let those college kids who love Google so much run their books through your filters (and do the manual tweaks and verification) for you. They won't need anything but a tripod, a camera, and the book.
What huge number of people who are obese are being reassured that they are perfectly healthy? I'm calling bullshit on this.
How is the definition of obese being used to make laws? There is a proposed tax in NY on games, movies, junk food, etc. There is no proposed tax on obese people.
If your claim is that the law is based on faulty numbers overestimating the problem of obesity, see:
"The government and health industry are declaring huge numbers of people obese who are not, and reassuring a huge number of people who are that they are perfectly healthy."
"Huge number" seems to me to cancel out "huge number". Do you have ACTUAL numbers? No? Okay.
No doctor worth a damn will not label you as obese based solely on BMI if you are actually muscular. A muscular person who is overweight/obese according to BMI will be very obviously muscular.
Are you being labeled as obese without being seen by a doctor? How? Do you fill out height and weight and gender on a form? Just lie. If someone contests it later, say you've been working out. If someone doesn't believe you, then go to a doctor and show them you're not obese (see above, no doctor worth a damn would label Arnold (young Arnold) obese.
Again, show me these "huge numbers" of people being told they are fine when they aren't, or vice versa.
In the end, I believe that BMI charts are accurate for the vast majority of people, and almost anyone who falls outside of that range is already well aware that the chart doesn't apply to them. If they feel persecuted, at worst, they have to lie on a form or find a non-retarded doctor to certify their health and non obesity.
If you allow "his" to be appropriate, you must also allow "her" to be appropriate. There is no logical reason for preference of one over the other.
I have no problem with "his" being used, or being considered the default. I have serious issues with "his/her" and "(s)he" and alternating between the genders.
But let's not fool ourselves. "His" is by no means correct unless the gender is known (and is male).
Right, but they don't follow through all the way with this. If a 21 year-old man and a 21 year-old woman (in the same city, same car, with clean records, etc.) put down that they each drive 10,000 miles per year, the man will be given higher rates. It'll be blamed on the "fact" that men get into more collisions/etc., when in reality, the woman in this case is more likely to have more (and more expensive) claims.
And how has NASA's budget situation been lately? In case you missed it, my comment had 2 meanings:
1: Programmers creating perfectly engineered code are usually only doing so the include in their resume.
2: Programmers creating perfectly engineered code will eventually be fired/let go as that practice is not economically viable. The same holds true for every good or service - cheap wins out over quality every single time.
Slashdot Mirror
Oh, I'm all against the bullshit, fear-mogering pedophile witch hunt.
I was just making a lulz.
Correct.
"6. The Right to possess any information"
thinkofthechildrennowait,you'rethinkingofthemtoomuch
It's a poor carpenter who blames his tools.
Notepad is fine.
In fact, it's awesome.
--Initial comment about superiority of vi/emacs/etc. over notepad.
--Comment vi/emacs/etc. sucking, and emacs/vim/etc. ruling.
--No you.
Carry on, I got you guys started.
I blame all the morans who didn't vote for Ron Paul.
Perhaps he means octarine.
HOLD IT!
Assassin's Creed for the PC uses SafeDisc 4.85.
You need the physical disc to play the game (or sometimes just for the installation).
You claim your original disc "broke".
You claim you needed to reinstall the game.
Something doesn't add up with your testimony.
Would you care to explain how you reinstalled the game with the downloaded copy without cracking the protection?
</Phoenix Wright>
ISPs do not have common carrier status.
Shocking I know, but the internet may have mislead you!
?
Seems to me due process deals with law enforcement, not your precious internet connection.
Is an ISP that connects me to the internet.
Make it reliable. Make it fast. Make it competitively priced.
(Don't give me that "choose 2" bullshit - we're lucky to be able to choose 1!)
Hey Joe, that moron is giving out conflicting info regarding the phone number listed on that page, and the voice sounds different when we call versus when we get a call. Check it out, will you?
K. Beep-boop-beep.
OMFG STOP CALLING ME.
Wait, I'm from Yahoo!. I called from my cellphone, and I get routed to you. When I call from the office phone, I get routed to the other person. HMMMM.
HMMM.
Stop worrying about internet drama.
Then apply it to each section.
Entries that don't get reformed well won't OCR well.
Entries that don't get OCRd well can automatically be flagged for manual inspection.
You can have a confidence interval based on the type of text. A novel will be all text and you would expect it to OCR perfectly. A coffee table book will have mostly photos that won't be OCRd. A text book will have symbols that will fuck up OCR.
Yes, what I suggested was simplistic. But it's not bullshit. It's the base idea on which you can build your tools. My suggestion is almost entirely software based - tripod, camera, book. My idea is cheaper, quicker, and far more distribution-friendly (even when you ignore the patents). The main idea is simple, and implementation can be as complex as you want it to be.
Google models each page in (stereoscopic) 3D.
My idea models a few pages, and intelligently adjusts the others, knowing they'll be nearly identical.
Until Google ramps their tech up to include being able to scan pop-up books, the two methods will achieve nearly identical results.
Do you think all Google is doing is using it's magic infrared cameras? Do you think there is no quality control, or that they don't have a problem? We've already seen pages come out where we see the flipper's hands (with cute pink latex gloves no less).
WOW DUDE.
LEARN TO HISTORY.
Yes, when I say so.
Barring my availability to make a ruling, you may refer to the actual definitions of the words you are using.
Post it everywhere.
Don't hide.
Print that shit out and nail it to the wall, Martin Luther style.
That's the dog's name.
There sure is.
Meet the King of the Internet.
http://www.youtube.com/edarem
Edarem for president!
I know they do - that's why you take a pic of the first and last pages first, and adjust those.
The 3D model for all pages is the same - a sheet of paper of a certain length and width.
The lay of the paper will be between the two extremes of the first and last pages.
Effectively, you can define the lay of the paper as a simple curve in the x/y plane.
The last page will be a flat line, the first page will be the most eccentrically curved.
Page 3(4) lays almost identical to page 1(2).
The curve is just a little flatter.
Page 699(700) lays almost identical to page 701(702).
Just tweak the first and last pages, and maybe some in the middle if you've got fucked up pages, books with different paper stock (maybe you're scanning Independence day and it's got stills from the movie in there on photo stock), etc.
It's cheaper, it's simpler, and the task can be easily distributed to other libraries.
Can't you just calculate the 3D model of the page based on a known stuff?
Make a generic flattener filter that takes in page height and length, as well as page number.
Manually tweak the output a bit for the first and last pages, and then intermediary pages can all be calculated with much more accuracy than you need.
Hell, with this method any book "scanned" (using a camera from overhead) could be processed. Let those college kids who love Google so much run their books through your filters (and do the manual tweaks and verification) for you. They won't need anything but a tripod, a camera, and the book.
CRY SOME MOAR
What huge number of people who are obese are being reassured that they are perfectly healthy?
I'm calling bullshit on this.
How is the definition of obese being used to make laws? There is a proposed tax in NY on games, movies, junk food, etc. There is no proposed tax on obese people.
If your claim is that the law is based on faulty numbers overestimating the problem of obesity, see:
"The government and health industry are declaring huge numbers of people obese who are not, and reassuring a huge number of people who are that they are perfectly healthy."
"Huge number" seems to me to cancel out "huge number". Do you have ACTUAL numbers? No? Okay.
No doctor worth a damn will not label you as obese based solely on BMI if you are actually muscular. A muscular person who is overweight/obese according to BMI will be very obviously muscular.
Are you being labeled as obese without being seen by a doctor? How? Do you fill out height and weight and gender on a form? Just lie. If someone contests it later, say you've been working out. If someone doesn't believe you, then go to a doctor and show them you're not obese (see above, no doctor worth a damn would label Arnold (young Arnold) obese.
Again, show me these "huge numbers" of people being told they are fine when they aren't, or vice versa.
In the end, I believe that BMI charts are accurate for the vast majority of people, and almost anyone who falls outside of that range is already well aware that the chart doesn't apply to them. If they feel persecuted, at worst, they have to lie on a form or find a non-retarded doctor to certify their health and non obesity.
If you allow "his" to be appropriate, you must also allow "her" to be appropriate. There is no logical reason for preference of one over the other.
I have no problem with "his" being used, or being considered the default. I have serious issues with "his/her" and "(s)he" and alternating between the genders.
But let's not fool ourselves. "His" is by no means correct unless the gender is known (and is male).
Right, but they don't follow through all the way with this. If a 21 year-old man and a 21 year-old woman (in the same city, same car, with clean records, etc.) put down that they each drive 10,000 miles per year, the man will be given higher rates.
It'll be blamed on the "fact" that men get into more collisions/etc., when in reality, the woman in this case is more likely to have more (and more expensive) claims.
And how has NASA's budget situation been lately?
In case you missed it, my comment had 2 meanings:
1: Programmers creating perfectly engineered code are usually only doing so the include in their resume.
2: Programmers creating perfectly engineered code will eventually be fired/let go as that practice is not economically viable. The same holds true for every good or service - cheap wins out over quality every single time.