Google Books team open sources a book scanner | Hacker News Reader

Google Books team open sources a book scanner

hackaday.com

151 pointscleverjake19 comments

I thought the design of their book scanner was very creative. It was also really interesting to see how they were basically able to prototype the entire thing in a day using cardboard they found around the office and a vacuum cleaner. I read through their design document and I'm considering trying to model all the various parts in CAD software for others to use.

However, the figure they posted of 45% of books tearing a page was a little troubling. I'm hoping the design can be improved to be gentler on the books. I read someone else mention that if you wanted to use this scanner on a book then you likely don't care about possibly damaging it. If that is the case, then why not just cut off the binding and feed the pages into a sheet fed scanner? If you do care about the books being damaged then you would likely use one of the more expensive automated scanners or a manual scanner.

I think the use of air pressure to pick up individual sheets of paper goes back to card readers (as in http://en.wikipedia.org/wiki/Punched_card_reader). See for example http://bitsavers.trailing-edge.com/pdf/documation/M300_TechM... (page 9):

"The riffle air acts on the first half inch of cards in the input hopper so that they stand apart [...] This prevents the cards from sticking together"

"The picker mechanism utilizes a strong vacuum to grasp the bottom card"

So, this might seem creative, but programmers over the age of 50 or so will find it old-fashioned :-)

Also 45% chance of a page with a teared page to me sounds much and much better than 100% of chance of having the back completely cut away from your book.

It was also 45% chance that a book would end up with a tear, not a 45% chance per page.

A prime example of a difference between Google and Apple both on HN front page at the same time. "Apple patents the page turn": http://bits.blogs.nytimes.com/2012/11/16/apple-now-owns-the-... Then Google open sources a book scanner.

Did you care to read the comments in the associated HN thread?

Not sure I see what you're getting at here. Care to elaborate?

Google open sourced something related to books and Apple patented something ridiculous related to books. Not a direct comparison as both companies have patents, but Google does share with the community a ton more than Apple.

To be clear, 'their book scanner' refers to a '20% project' by an employee - not the actual hardware behind Google Books.

Is there any public knowledge about what the hardware behind Google Books looks like?

There's this snippet showing that humans are involved.

(http://thenextweb.com/google/2012/03/18/google-books-scannin...)

I'd like a comparison between human tearing of books and the machine. 45% seems way too high.

For reference, here is a camera-based DIY automatic book scanner: http://www.youtube.com/watch?v=1PE1Z4wl9fk&feature=relmf...

As far as I can tell, the main advantage of the google device over this one is the ability to keep the book open to less than 90 degrees. Though I think his (close to) 45% damage rate is worse for the books in the aggregate. I'm sure that it could be refined to get the error rate down, but I think that mechanically handling the pages right near the spine (as the google scanner does) may never be that safe.

I hope innovations like this could reduce the price of book scanning. I recently used a book service in San Jose to scan all my books: http://1dollarscan.com. It was good experience, with good support, but a bit pricey with their extras.

Sounds great until you look more closely. Here's the plan:

1) Gather up all your treasured books and send them to us 2) We'll make scans of them 3) We'll then grind your favorite books into compost, so you can Save The Planet! Yay!

$1500 Google design I still have to build myself that only damages 45% of my books? No, thanks.

Scanning service that only charges me 1$/book to destroy my books for me? No, thanks.

I think I'll just look around, thank you.

It depends on the circumstances. This service is brilliant for people outside of the US who are interested in books published in the US. You can send books directly from Amazon,com or from one of the used sellers. Therefore you only pay regular postage and find most of the time used versions of a book (which not always ship international).

When I was in Europe, I would have paid $15 for the only available used version of a book on the local Amazon, with 1dollarscan I payed $1 book + $3 shipping + $2 for the scan.

I think you're being overly critical of the Googler's design. It's only a prototype after all. I think it's more like a proof of concept at this point - that you can have your own book scanner that is much cheaper than professional versions.

That's why I only pay the $1 per 100 page and OCR myself. May take a bit longer, but then I have much more freedom/flexibility.

holy cow those extras could add up quickly. If you OCR'd, how were the results?

So even though for me it was a one-shot deal, I had plenty of time (I was moving) and it made sense for me to get their 'Platinum Membership' which includes many of the extras -- for a few months - which covers 100 sets for free each month. Still it was definitely expensive - more than I'd like. The end result were high quality (but definitely large) pdfs. I could use their "fine tune" services to optimize for different platforms / compress, too, but for me that meant additional storage / time if I kept the originals, so I didn't use it much. In the end I spent a lot of money, but now have copies of my books -- which I didn't have to ship long-distance and which I don't have to physically store now. I'm more-or-less happy-ish.

Reply on news.ycombinator.com