Out-of-print books

If you ever need an example of how unhelpful and badly designed our current publishing system is, the existence (or rather, lack of existence) of out-of-print books is ready-made.

Now there was a time when not publishing a book could make serious economic sense.  Publishers couldn’t afford to publish runs of books below a certain number, and the demand for some books can become so small that there was no way to profitably print them.  It’s a shame but an understandable economic reality.

This is simply no longer the case.  Print-on-demand services (for example, lulu.com) can now print books as people order them for a cost considerably lower than the list price of any math textbook.  All a publisher needs to do is put PDFs of their books on such a website, put a $30 markup on them (or more, considering how much math books cost), and let the money roll in.  If they don’t have PDFs, I bet Google Books would make them for free.  In short, publishers are leaving money they could be making on their back catalogue on the table, and hurting the mathematical community at the same time.  Thanks, guys.

This rant was engendered by a post of Timothy Chow’s at What’s New (a.k.a. Terry Tao) about a new website, where one can express one’s desire for a old math books to be brought back into print.  The website’s a good idea but ultimately getting specific books that are particularly popular back into print is a short-term fix.  The real problem is that publishers’ mindset still hasn’t caught up to the advances in technology. When are they going to enter the 21st century?

[Ed. – last paragraph edited a bit in response to comments]

16 thoughts on “Out-of-print books

  1. In principle I agree with your points, but I don’t think that that the process of making a book available is quite as labor-free as you make it sound. There are plenty of out-of-print books that predate the electronic era. Even scanning the books in as images does take a certain amount of time. And what one really wants is for the text to be converted to electronic form. Optical character recognition software isn’t bad, but it doesn’t fare so well on mathematical texts with a lot of symbols and diagrams. A significant amount of human labor is still needed.

    Sometimes there is also, unfortunately, considerable labor involved in tracking down exactly who owns the copyright to a text, especially if the author is dead and the publisher no longer exists as a legal entity.

    These problems will presumably solve themselves in the long run, but there are some books that I would like to have for myself, not just for my grandchildren. In the real world that we live in today, some prioritizing is necessary. So in addition to thinking about the long-term future, we should think about what we can do today to find the most-in-demand books and make them available ASAP.

  2. The real problem is that the AMS is still wasting their own and referees’ time trying to figure out what part of their back catalogue they should print.

    Assuming that the book Chow was interested in is the first entry at the new site (Welsh, “Matroid Theory”), then it isn’t part of the AMS’s back catalogue but rather Academic Press’s. So there would be more work on the AMS’s part to negotiate copyright issues, etc., which would actually cost money even for the print on demand model you propose.

    But your broader point is clearly correct; AMS, Springer, etc. should have all their back catalogue available on LuLu or similar. It’s sad that copyright, which is supposed to promote the creation and dissemination of creative works by providing a (limited!) monopoly as an incentive, seems so often to retard their spread to no ones benefit.

  3. The AMS does make available its entire back catalogue (except proceedings), according to the Report of the Executive Director in the August 2007 Notices that I was coincidentally reading this morning. (Presumably you can find the same info here, but my link is slow.) The issue was then whether to acquire the given title, as Nathan suggests.

    If it is the book on matroid theory, I am also quite surprised the referees thought that matroid theory was not thought to be active, although I’m certainly no expert. (And I guess I’m not too surprised to hear there’s a large set of mathematicians that look down on it.)

  4. There are plenty of out-of-print books that predate the electronic era. Even scanning the books in as images does take a certain amount of time.

    Google does this for free, and I’m sure would be overjoyed to get a publisher’s out-of-print titles in exchange for giving the publisher excellent scans and whatever their OCR got out of the text (which in my experience tends to be pretty good).

    Sometimes there is also, unfortunately, considerable labor involved in tracking down exactly who owns the copyright to a text, especially if the author is dead and the publisher no longer exists as a legal entity.

    This I’ve got no slick answers for, other than writing your congresscritter. The present state of copyright law is a travesty, and Congress needs to know that people are pissed off about it.

    The AMS does make available its entire back catalogue (except proceedings).

    Great! I stand corrected. Somebody get them a cookie.

  5. To follow up on Dylan’s comment, here’s the relevant part of the 2007 AMS ED’s report

    The Society has more than 3,000 titles in print (and, by the way, all 3,000 are searchable online through
    the Google book program, and soon will be through the comparable Microsoft book program as well). The AMS has this staggering number of titles because it pledges to keep every authored monograph in print—forever. We do not let authored books go out of print (but, of course, we do let proceedings go out of print). This is a policy that serves both our authors and the community well. Until recently, it was a difficult policy to administer because it meant printing small quantities of books that only sold a few copies each year. We now have a full-featured print-on-demand program, however, that allows us to produce one copy of a book, at moderate price and high quality. We will expand this program in the coming years.

  6. The quality of the print-on-demand books I’ve ordered is quite low, to the point where now I usually don’t even bother buying them. These are all from major math publishers. But I could imagine that for scaling reasons, a centralized print-on-demand publisher could raise the quality. On the other hand, maybe the major publishers already contract it out to such a company.

  7. James wrote: “The quality of the print-on-demand books I’ve ordered is quite low, to the point where now I usually don’t even bother buying them. These are all from major math publishers.

    At “http://www.math.byu.edu/~grant/dt/index.htm“, I have an extremely primitive web page complaining about the quality of print-on-demand math books and presenting a few samples as evidence. Cambridge appears to have gone overwhelmingly to print-on-demand for their back catalog, but at least they are fairly open about it, adding the somewhat euphemistic phrase “Transferred to digital printing” to their copyright pages. (Euphemistic because digital printing doesn’t have to be as inherently low-quality as theirs is.) Springer is going a similar route but without the same level of candor. Usually I can find nothing in an advertisement for a Springer book that indicates that a substantial change has been made, so I’ve had a number of bad surprises when shipments have arrived.

    If publishers have high-quality (unscanned) PDFs of out-of-print books, I’d much rather that they make arrangements to sell me an electronic copy than a poorly printed one.

    But I could imagine that for scaling reasons, a centralized print-on-demand publisher could raise the quality.

    As I understand it, a lot of the academic publishers are already using centralized print-on-demand companies like BookSurge and RR Donnelley, and the quality is still lousy.

  8. Chris, I had wondered why a number of math books I’ve purchased recently seemed “blurry”, with print quality inferior to that of a typical laser printer. It looks like a photocopy, basically. Thanks for explaining what’s going on. Even quite recent books are being printed this way, I first noticed this in this book which appeared in 1999 but I purchased in 2006. I’m guessing these “digital reprints” are created by scanning the originals, not printing from the original PS or PDF file, since it’s hard to see why the quality would be so poor otherwise.

  9. I once ordered a book from the AMS that was printed on demand, and had no complaints about the quality. But it sounds like the AMS is unusually good in this regard.

    In any case, based on what I’ve just learned from reading the above comments, I now have a new question. Are most out-of-print academic books already, in fact, available on demand now? If not, and if it really is as profitable and easy as Ben says, then why not? Publishers are not usually slow to seize an opportunity to make a fast buck.

  10. I have scanned for myself a couple of out-of-print books, and I think there is no problem with the quality, as long as one is prepared to make 600 (or bigger) dpi scans. Surely, such a file is big, and scanning at high resolution is slow, but technically it is more than possible.

    I presume it’s publisher’s monopoly position (they own the copyright) that makes them careless here.

    Threads above also mention OCR, but OCR is completely unnecessary for copying (yes, certainly, OCR makes files much smaller, that’s right)

  11. I agree with Timothy Chow that what is needed is some form of grassroots action. Mathematicians are a relatively small community, and we should be able to take mathematical publishing into our own hands continuing what ArXiv is already doing. Paying companies like Springer a one-time lump sum to put certain out-of-print math books which it never plans to make any money off anyway into the public domain, making the text a PDF, printing it with a really nice printer, bookbinding it, and selling it to individuals. Maybe one could even do it with long ArXiv papers- print them on demand. Greg, would there be any legal obstruction to somebody doing that? Printing and binding long ArXiv papers and selling them to individuals? I would pay for such a service- easier and nicer than printing them badly and stapling together 200+ pages…

  12. Daniel,

    FWIW, many of the photocopy shops (at least here in Berkeley) can do a credible job binding loose paper printouts into a book. I’ve got a copy of Kapustin & Witten’s preprint bound this way. It’s not an ideal solution, of course — ordering the books online and having them delivered to your mailbox would be better. But it only costs about $10 and a few minutes, so it’s pretty well worthwhile for things you intend to spend a lot of time with.

  13. From the other side of niche publishing: my novel, which was published by a “big small press,” is, as far as I understand, to be in print perpetually, even though, five years out, it surely sells only a handful of copies a year. And I don’t think this is strange in the literary world.

  14. It sounds like you have amazing resources in Berkeley… I wish I could have Kapustin-Witten bound for 10 dollars!
    The more I think about it, the more I think that the dream solution would be to have an “order hardcopy” option on ArXiv which would send you a bound copy nicely printed to anywhere in the world for a decent price. Expanding ArXiv sufficiently would also solve the “out of print books” problem… imagine if they had ArXiv codes and you could order them this way! And if you ordered, they would make the book out of high-quality scans like Dima suggests- but the ArXiv version would be a PDF. Of course, there are many intermediate steps… I wonder whether ArXiv has considered giving codes to all the journal papers available online, and making them ArXiv indexed and searchable (even if the .pdf/.ps/whatever links went to the journal site).
    One problem with this solution is that ArXiv isn’t a company, and the customer service could be much better. Maybe having a print-on-demand section for papers would also give them a financial incentive to improve customer service…

  15. It sounds like you have amazing resources in Berkeley… I wish I could have Kapustin-Witten bound for 10 dollars!

    You may well already be able to. I haven’t checked prices, but all large copy center chains (such as Kinko’s) will do it. In fact, you can often upload your file online and just drop by to pick it up.

    Maybe having a print-on-demand section for papers would also give them a financial incentive to improve customer service…

    Having any financial incentive would seriously jeopardize the future of the arXiv. It’s much easier to get legal permission to put your paper on a service that makes it available for free, and publishers are more tolerant in the cases of authors who don’t actually get permission. If the arXiv started charging money for anything, even if it was only to recover costs, then publishers would throw a fit.

    I wonder whether ArXiv has considered giving codes to all the journal papers available online, and making them ArXiv indexed and searchable (even if the .pdf/.ps/whatever links went to the journal site).

    That would be contrary to the spirit of the arXiv. The purpose isn’t just to provide access to papers but rather to make them freely available forever. Linking to journal web sites wouldn’t help, and it might actually hurt. (Lazy authors would say “Well, everyone I care about is at a university with a subscription, so they can already access my paper through the arXiv. Why then should I submit the paper directly?”)

Comments are closed.