7 ms·

Google's Book Search: A Disaster for Scholars

43 points by jamesbritt 16y ago

waterlesscloud 16y ago
I devote a fair amount of time to indulging my idle curiosities in gbooks, and it takes a lot of work and more than a little luck to find anything useful. It's a total mess. Great resource, but nowhere near what it should be. I'm glad someone is drawing attention to the irresponsible sloppiness.
- cryptoz 16y ago
  I'm pretty sure Google Book Search was never intended for puny humans to read. It's for the future Google AI to learn from.
gojomo 16y ago
The Internet Archive's 'Open Library' is a project to create one reference page per book, created and corrected via the wiki model: http://openlibrary.org/ http://openlibrary.org/ [I work for the Internet Archive, but on our web archiving rather than our book projects.]
- chime 16y ago
  Is there a way to link each of your references to each of Google's Book references? A one-to-one mapping could solve almost all the issues the blog author raised.
  - gojomo 16y ago
    Looks like a feature to formally link to a copy of a book/edition anywhere on the web is coming soon. (Informally linking from description hypertext or a list of related links is already possible.) Whether Google backlinks or otherwise uses the info to improve their own metadata is of course up to them, but Google didn't get where they are by turning down free sources of crawlable quality information.
nhebb 16y ago
I think what Google is doing is fantastic, despite these problems. Some of the sloppiness is a byproduct of the attempts to automate the book scanning and metadata gathering. This has got to be a hugely expensive undertaking, so I can understand the shortcuts to reduce cost. Until this gets cleaned up, what then, is the fallout from the metadata errors? Researchers will have to verify publishing dates and use other means of search for period pieces, but they will have access to historical texts over the internet. That's still pretty awesome.
- ascuttlefish 16y ago
  But without good metadata, accessing and searching those texts will be an exercise in frustration and missed results.
BonoboBoner 16y ago
I often need the very specific tiny bits of information that is trapped between pages of paper stored in some library somewhere. Today I need to fill out forms and afterwards wait days/weeks to get that information delivered to me. The thought about all of the necessary resources for this (paper, transportation, storage, etc) makes you almost feel ashamed, because most of the time the amount of desired information is very small. I would love to abolish this waste and have a service payed for by a flat rate that gives you access to all books, papers, phd-thesises, etc by keystroke. The service should not be completely free in order to maintain the incentive of writing well structured books as way of personal income. Whether it is Google Book Search or something else, I think humanity definitely needs something like a global digital library.
houseabsolute 16y ago
> In any case, none of that should relieve Google of the responsibility of making its collections an adequate resource for scholarly research. I wasn't aware Google was under any such obligation. If it is, from where did the obligation come, when, and by whose authority?
- robryan 16y ago
  I'd imagine in being allowed to scan a large amount of the works they are storing they presented the system in a way that would improve the way scholars did things. It's possible they gave no guarantees but if they did they have the responsibility to live up to them. As mentioned they may have a monopoly on scanning the vast majority of older texts online due to the court settlement and deals with processors. If they are to be the gatekeeper and don't act responsibly then it is really a loss for everyone, regardless if they are bound to or not.
swombat 16y ago
All this is not an argument against Google Book Search, but an argument for improving the metadata of said system. If the scholars are so distraught by this, they should offer to help review and correct Google's digital library. I'm sure they can come to some kind of arrangement.
- nkassis 16y ago
  They did not agree to this in the first place. Most of them were not happy about the deal google was given. I strongly believe it is google's job to ensure that their work is of good quality. This is similar to Wikipedia errors that find their way into newspapers and even research. The difference is that wikipedia isn't claiming to be a reliable source. Google is trying to build an exhaustive catalogue and they are the only one allowed to distribute a large amount of text due to a court settlement. Good metadata should be part of that deal. And like the article mentioned, there already exist much better catalogues (or whatever they are called) such as the library of congress one. They might not be perfect but they sure beat scanning a book cover and picking the first date found. Why not just license it?
  - angstrom 16y ago
    That's what I've often wondered. Why isn't this a government initiative to take the Library of Congress and provide a digital layer on top of it for public access. Then just turn the regular libraries into community centers without the overhead cost of book storage, purchasing, & cataloging. I don't know about anyone else, but the idea of this being in the hands of a single corporation is unnerving.
    - nanairo 16y ago
      I think the EU tried something like that... not sure what's happened to it in the end.
    - CWuestefeld 16y ago
      Why isn't this a government initiative to take the Library of Congress and provide a digital layer... There's nothing in the Constitution that gives the Federal government the authority (let alone the responsibility to do this). the idea of this being in the hands of a single corporation is unnerving. And having it controlled by the government would make you breathe a sigh of relief? The opportunities for curating the collection as in 1984 are scary enough, but consider the accuracy of government data in other areas.
akadruid 16y ago
This article argues seems to be suggesting that Google have a unique responsibility - almost as if they have somehow removed something that already existed. I disagree with their fundamental premise that Google is "the world's last digital library". Google's demonstration that such a thing is actually possible removes one barrier to entry. They have also set a precedent which reduces the negotiation required for future projects. The fact that Google has done a poor job (so far) on the metadata is something that's worth highlighting - but it's not a reason to curse the whole project. This is fixable, possibly even a transient issue. GBS is an entirely new resource, which scholars have not had before; if they find it unsuitable for scholarly research, perhaps they could continue to use their existing tools or work to improve it, rather than slam the effort as "A Disaster For Scholars". "A Useful Tool Which Still Needs Improvement" might not be as snappy but the The Chronicle Of Higher Education doesn't look like the Daily Mail either.
- SkyMarshal 16y ago
  > "I disagree with their fundamental premise that Google is "the world's last digital library". Google's demonstration that such a thing is actually possible removes one barrier to entry." That was my first reaction too, but then I realized the article was implying that the real barrier to entry is the work of scanning billions of books into digital form. After Google has done that, there will be little incentive for anyone else to incur that same expense to create a service that Google already has a 5 year head start on. Alot of startups these days are about finding better ways of doing things in already-crowded domains. Google itself showed the way there. But I'm not sure I disagree with article in its assertion that doing this will be much more difficult, if not virtually impossible, in this domain of creating a digital library from scanned books. There really is a significant moat here.
Tyrannosaurs 16y ago
The idea that the barriers to entry mean that Google will never have competition are very high is true today but won't be true going forward. Once all books are published (or reprinted) in eBook form, the conversion is potentially something which can be automated for all new (or still in "print" books). This will cover the majority of titles of interest. Additional titles can be scanned where they're of significant interest or projects similar to Project Guttenberg may do some of the work. Google has a head start but in actual fact has done a lot of their work when it's high cost / high labour where future competitors won't need to invest this level of effort. You could certainly see publishers providing access to their own works and, having access to the electronic versions of texts even for non-eBooks, they could do so relatively cheaply. Add a standardised interface over the repositories provided by each publisher and you have a decentralised repository, owned by the publishers which you'd imagine could compete with Google.
- _delirium 16y ago
  > Once all books are published (or reprinted) in eBook form, the conversion is potentially something which can be automated for all new (or still in "print" books). I think that's unlikely to ever happen, though. There are still thousands of even relatively prominent movies available on VHS that have never been released on DVD, due to a mixture of disputes over rights, expense of mastering a DVD from the source material, worry about whether they'd be sufficiently marketable this many years past their release, worry about too many back-catalogue releases cannibalizing new releases, etc.

43 points by jamesbritt 16y ago

I devote a fair amount of time to indulging my idle curiosities in gbooks, and it takes a lot of work and more than a little luck to find anything useful. It's a total mess. Great resource, but nowhere near what it should be. I'm glad someone is drawing attention to the irresponsible sloppiness.

cryptoz 16y ago
I'm pretty sure Google Book Search was never intended for puny humans to read. It's for the future Google AI to learn from.

gojomo 16y ago

The Internet Archive's 'Open Library' is a project to create one reference page per book, created and corrected via the wiki model: http://openlibrary.org/ http://openlibrary.org/ [I work for the Internet Archive, but on our web archiving rather than our book projects.]

chime 16y ago
Is there a way to link each of your references to each of Google's Book references? A one-to-one mapping could solve almost all the issues the blog author raised.
- gojomo 16y ago
  Looks like a feature to formally link to a copy of a book/edition anywhere on the web is coming soon. (Informally linking from description hypertext or a list of related links is already possible.) Whether Google backlinks or otherwise uses the info to improve their own metadata is of course up to them, but Google didn't get where they are by turning down free sources of crawlable quality information.

nhebb 16y ago

I think what Google is doing is fantastic, despite these problems. Some of the sloppiness is a byproduct of the attempts to automate the book scanning and metadata gathering. This has got to be a hugely expensive undertaking, so I can understand the shortcuts to reduce cost. Until this gets cleaned up, what then, is the fallout from the metadata errors? Researchers will have to verify publishing dates and use other means of search for period pieces, but they will have access to historical texts over the internet. That's still pretty awesome.

ascuttlefish 16y ago
But without good metadata, accessing and searching those texts will be an exercise in frustration and missed results.

BonoboBoner 16y ago

I often need the very specific tiny bits of information that is trapped between pages of paper stored in some library somewhere. Today I need to fill out forms and afterwards wait days/weeks to get that information delivered to me. The thought about all of the necessary resources for this (paper, transportation, storage, etc) makes you almost feel ashamed, because most of the time the amount of desired information is very small. I would love to abolish this waste and have a service payed for by a flat rate that gives you access to all books, papers, phd-thesises, etc by keystroke. The service should not be completely free in order to maintain the incentive of writing well structured books as way of personal income. Whether it is Google Book Search or something else, I think humanity definitely needs something like a global digital library.

houseabsolute 16y ago

> In any case, none of that should relieve Google of the responsibility of making its collections an adequate resource for scholarly research. I wasn't aware Google was under any such obligation. If it is, from where did the obligation come, when, and by whose authority?

robryan 16y ago
I'd imagine in being allowed to scan a large amount of the works they are storing they presented the system in a way that would improve the way scholars did things. It's possible they gave no guarantees but if they did they have the responsibility to live up to them. As mentioned they may have a monopoly on scanning the vast majority of older texts online due to the court settlement and deals with processors. If they are to be the gatekeeper and don't act responsibly then it is really a loss for everyone, regardless if they are bound to or not.

swombat 16y ago

All this is not an argument against Google Book Search, but an argument for improving the metadata of said system. If the scholars are so distraught by this, they should offer to help review and correct Google's digital library. I'm sure they can come to some kind of arrangement.

nkassis 16y ago
They did not agree to this in the first place. Most of them were not happy about the deal google was given. I strongly believe it is google's job to ensure that their work is of good quality. This is similar to Wikipedia errors that find their way into newspapers and even research. The difference is that wikipedia isn't claiming to be a reliable source. Google is trying to build an exhaustive catalogue and they are the only one allowed to distribute a large amount of text due to a court settlement. Good metadata should be part of that deal. And like the article mentioned, there already exist much better catalogues (or whatever they are called) such as the library of congress one. They might not be perfect but they sure beat scanning a book cover and picking the first date found. Why not just license it?
- angstrom 16y ago
  That's what I've often wondered. Why isn't this a government initiative to take the Library of Congress and provide a digital layer on top of it for public access. Then just turn the regular libraries into community centers without the overhead cost of book storage, purchasing, & cataloging. I don't know about anyone else, but the idea of this being in the hands of a single corporation is unnerving.
  - nanairo 16y ago
    I think the EU tried something like that... not sure what's happened to it in the end.
  - CWuestefeld 16y ago
    Why isn't this a government initiative to take the Library of Congress and provide a digital layer... There's nothing in the Constitution that gives the Federal government the authority (let alone the responsibility to do this). the idea of this being in the hands of a single corporation is unnerving. And having it controlled by the government would make you breathe a sigh of relief? The opportunities for curating the collection as in 1984 are scary enough, but consider the accuracy of government data in other areas.

akadruid 16y ago

This article argues seems to be suggesting that Google have a unique responsibility - almost as if they have somehow removed something that already existed. I disagree with their fundamental premise that Google is "the world's last digital library". Google's demonstration that such a thing is actually possible removes one barrier to entry. They have also set a precedent which reduces the negotiation required for future projects. The fact that Google has done a poor job (so far) on the metadata is something that's worth highlighting - but it's not a reason to curse the whole project. This is fixable, possibly even a transient issue. GBS is an entirely new resource, which scholars have not had before; if they find it unsuitable for scholarly research, perhaps they could continue to use their existing tools or work to improve it, rather than slam the effort as "A Disaster For Scholars". "A Useful Tool Which Still Needs Improvement" might not be as snappy but the The Chronicle Of Higher Education doesn't look like the Daily Mail either.

SkyMarshal 16y ago
> "I disagree with their fundamental premise that Google is "the world's last digital library". Google's demonstration that such a thing is actually possible removes one barrier to entry." That was my first reaction too, but then I realized the article was implying that the real barrier to entry is the work of scanning billions of books into digital form. After Google has done that, there will be little incentive for anyone else to incur that same expense to create a service that Google already has a 5 year head start on. Alot of startups these days are about finding better ways of doing things in already-crowded domains. Google itself showed the way there. But I'm not sure I disagree with article in its assertion that doing this will be much more difficult, if not virtually impossible, in this domain of creating a digital library from scanned books. There really is a significant moat here.

Tyrannosaurs 16y ago

The idea that the barriers to entry mean that Google will never have competition are very high is true today but won't be true going forward. Once all books are published (or reprinted) in eBook form, the conversion is potentially something which can be automated for all new (or still in "print" books). This will cover the majority of titles of interest. Additional titles can be scanned where they're of significant interest or projects similar to Project Guttenberg may do some of the work. Google has a head start but in actual fact has done a lot of their work when it's high cost / high labour where future competitors won't need to invest this level of effort. You could certainly see publishers providing access to their own works and, having access to the electronic versions of texts even for non-eBooks, they could do so relatively cheaply. Add a standardised interface over the repositories provided by each publisher and you have a decentralised repository, owned by the publishers which you'd imagine could compete with Google.

_delirium 16y ago
> Once all books are published (or reprinted) in eBook form, the conversion is potentially something which can be automated for all new (or still in "print" books). I think that's unlikely to ever happen, though. There are still thousands of even relatively prominent movies available on VHS that have never been released on DVD, due to a mixture of disputes over rights, expense of mastering a DVD from the source material, worry about whether they'd be sufficiently marketable this many years past their release, worry about too many back-catalogue releases cannibalizing new releases, etc.