viaLibri adds ISBN searching. Please ignore.

You may have noticed that a new feature has been introduced with viaLibri’s latest update. It is something many people have asked for. Most thought it should have been included a long time ago.  As in, from the beginning.  I resisted for many years, but have finally capitulated.  You are now able to search for books on viaLibri using ISBN.

Please don’t.

The reason is simple.  ISBN numbers are a terrible way to search for books.

I will certainly grant the fact that they serve an important purpose for the activities of publishers, distributors and new book stores.  I’m sure they are useful in other contexts as well, especially for those who are only interested in new books. If you inhabit a world where data is always orderly and you like the idea that books are generic objects suitable to the algorithmic demands of data processing and purchaser profiling, then ISBN is most definitely for you.  Happily, viaLibri does not yet live in that world, and I feel confident that most of its users do not want to live there either.  And they do not have to.  They do not need ISBN numbers, and are cordially invited to ignore them.

Because, as I said, ISBN numbers are a terrible way to search for books.  You will quickly discover this the first time you attempt to search online for an out-of-print book using its ISBN number and then repeat the search the old-fashioned way using author and title.  Author/title searches nearly always yield more and better results than searches based on ISBN.

The reasons for this are simple:  many of the booksellers who deal in older books do not bother with ISBNs, so the listings they put on the internet do not include them.   To a collector the information is meaningless, and the booksellers who focus on serving collectors generally share that attitude, even when they are also selling books to the general public.

But that is not the only reason why a second-hand book might be catalogued without its ISBN number.  Often a book will have a number, but it does not actually appear inside of it.  This is especially likely in the case of reprinted works that were originally published before ISBNs were firmly established. There are also many cases where the publisher didn’t obtain the ISBN until after the book was printed, or just didn’t think it was worth including as part of the text.  In all of these cases the book is very likely to be catalogued without its ISBN, and if you search for it using that ISBN there will be many available copies that you will not find .

A few examples pulled from my personal reference shelf will demonstrate.

You might, for instance, want to buy a copy of BOOKBINDING IN AMERICA 1680 – 1910. FROM THE COLLECTION OF FREDERICK E. MASER, published in 1983.  The ISBN number for this book is 0813910137, although it is nowhere to be found within the book itself.  But if you don’t have the number already you will have no trouble finding it by looking in WorldCat or an ISBN database.   If you use that number to search for your copy on viaLibri you will get 12 listings.  Only two copies are available for less than $25, both of them from Amazon.  However, if you try your search again, while ignoring the ISBN, and search instead for: title = “BOOKBINDING IN AMERICA MASER COLLECTION”,  you will receive 39 matches, including 3 additional copies that are priced for less than $25.  This is a significant difference in results.

Or, suppose you stumbled upon a reference to the 4 volume set of ARTS IN AMERICA, A BIBLIOGRAPHY, edited by Bernard Karpel and published by the Smithsonian Institution Press in 1979.  Suppose you could not resist the impulse to buy a set of your own.   If your reference did not give you the ISBN number (0874745780)  WorldCat will, as will many other online sources.   It is also printed in the book.  The 10 digits seem so precise and unambiguous. It is easy to think that they would be the logical way to find your copy.  Please do not be fooled.    If you use those numbers for your search parameter you will find only 66 matches (including many odd volumes and duplicates)  and there will be no complete sets available in North America for less than $45.  If, on the other hand, you try your search using author and title you will, instead, be rewarded with a total of 104 matches, including five complete sets in North America selling for $40 or less.  The ISBN matches will still be there, but so will many others that would have otherwise been missed.

These are not the only good reasons for ignoring ISBNs.  For me, the most compelling reason is the potential for discovery.  You can’t always know whether the ISBN you are using will correspond with the best possible version of the book you are interested in.  What if there is a later enlarged edition that has a new ISBN?  You would not find out about the updated version if you did your searching with the ISBN of the earlier edition.   The author/title search would quickly let you know.

Sometimes, when you use author and title to search for one book the results you receive will also show you another, different work by the same author that could also be of interest.   With ISBNs you rarely discover anything you are not specifically looking for.  With names and words you may find something unexpected that is even more interesting than the book you thought you wanted.

I would also mention the problem of typos, a problem that comes from both buyer and seller.  These, of course, can happen anywhere, but they are much harder to notice and correct when it is only a string of numbers that have been mistyped.

Are there circumstances where only searching  by ISBN is worthwhile?  Very few.

It might sometimes be useful to check for strays after the old-fashioned author/title search had been tried.  This might find a copy of a book with a typo or other cataloguing error that might otherwise be missed.  Anything is possible.

Sometimes students are assigned text-books that are being continually “updated” by their publishers with new ISBNs.  In this case the student will only want a copy with the correct ISBN.  Used copies that are listed without this information will not be satisfactory, so searching by number would not exclude anything the searcher would want to buy.

Lastly, I have been told that there are online listings of books entered using non-Roman alphabets and that, unless you have a special keyboard, these books can only easily be found using ISBN numbers.  Having never encountered such a book during my own extensive burrowing through online data I am a bit sceptical that such listings actually exist. But I do not rule it out.

It is with these special circumstances in mind that the latest change was made.  I hope it will be regarded as an improvement.  But I still worry that people will actually use it for a purpose it does not serve.

At least I can tell myself that you, patient reader, have been warned.

 

Two Hundred Years and Still Searching

I received an email the other day from one of my favorite librarians at one of my favorite libraries.  The original cause for writing is unimportant, but on a cold gray day I got a big boost out of something that was mentioned at the end.

The library in question, the Redwood Library in Newport, Rhode Island, is one of the oldest in North America.  Its original collection consisted of 751 titles shipped  from London in 1749, plus 126 additional early donations “by Several Gentlemen”.    To modern collecting tastes these are not particularly exciting books, but that is also unimportant. They are of interest to me, however, as a demonstration of the fact that books, even run-of-the-mill reprints,  are so much more vulnerable and hard to replace than the buildings that shelter and attempt to protect them; because in this case, while the library itself still stands, the collection it originally housed was stolen, destroyed or dispersed within a few decades of its original formation.

The loss, I should add, was quickly perceived.  For over two centuries now the successive librarians in charge have been working hard to replace the lost volumes and recreate the collection they started with over 250 years ago.   The list of missing volumes has been widely distributed and no sale list or catalogue of 18th century books arrives at the library without close scrutiny.  Acquisition funds have been available.  Scouts are on the hunt. Two hundred years is a long time to look for a book, and yet over 90 items (out of 877) still elude the empty shelf space that is waiting for them.

Libribot wants a shot at that list.  And it is going to get it.

I am curious to see how hard to find those books are actually going to be. I’ll let you know.