When databases walk the earth

Alert readers will recall that I live in Maine. In fact, that I live in that geographically squishy area known as Central Maine, about 3 hours from the Massachusetts state line and 2.5 hours from the Canadian border, if you’re feeling sanguine with regard to Coburn Gore.

Now, not only do I live in Maine, but I am a veritable giantess among Maine women — six foot tall in my striped sockfeet. This means that I need to either (1) make my own clothes, which I used to do when I was young and ambitious, but have not done for more than twenty-five years, (2) wear mens clothes, which I do pretty often, or (3) buy girl clothes from tall shops on the internet, which is what I do somewhat less often than (2).

One of my favorite vendors of tall women’s clothes is Long Tall Sally, a British chain with stores/distribution points in Massachusetts and in Canada.  Mind you, I order from the internet, and make no secret of my US address.

Which is of course why two out of three orders that I make with Sally are fulfilled by the Canadian distributor. I wouldn’t mind this so much, except that the Canadian facility runs the credit card and I get whacked with a currency conversion fee.

And, also, on the rare occasions when I need to return something, the Sally folk in charge of issuing RMA numbers pretty much always insist that my original order of course did not come from Canada; that would be. . .silly.


I suppose there’s a database somewhere in Sally’s kingdom that figures out which shipping facility is closest to what customer and shuffles the orders that way, with a fine — let us even allow, a joyous — disregard for such human vanities as the borders between countries.

* * *

Speaking of databases — this post is about databases — I wonder if someone who is more savvy than I am can explain why it is that bookstores can’t seem to handle co-authors in their databases. It seems universal, from the Big River on down to Ma & Pa’s Bookstore and Pizza Emporium.

Steve, as the second author listed on our collaborative work, is constantly dropped from the database record. This is not only unfair and untrue, but it means that readers who may only recall that “Steve Miller” is one of the authors of those space opera books they like so much can’t find what they want to read.

Which seems a disservice to readers, authors, bookstores, and publishers.

In other words — it’s lose-lose.

And yet the error persists.

Is it Just Too Hard to build a database that will accommodate the reality of co-authors? Or do the builders of databases, being, perhaps techies more than readers, not understand (and therefore don’t care about) the issue?

* * *

I have met folks who haven’t believed that co-authors do equal work on the projects that bear both of their names.

My co-workers on the newspaper many years ago for instance widely believed that my husband “let me put my name” on his books. To keep peace in the family, one assumes.

More recently, I submitted a novel by Sharon Lee and Steve Miller to the Maine Arts Commission, in application for a grant. After they had the application in hand, and despite having asked beforehand and told to go ahead, I was told that co-authored works were not acceptable. The official’s suggested solution was that I simply remove my co-author’s name from the application.

I don’t believe that, to this day, she understands why this was wrong, though, be assured, I did my Very Best to tell her.

So, it’s not at all outside the realm of possibility that those who are charged with building database may be. . .misinformed regarding the necessity of listing all the authors of a particular work.

What I’d like to know, I guess, is how to get to these folks in order to educate them.


4 thoughts on “When databases walk the earth”

  1. Some years ago, more than I care to count, I worked at B&N in Rochester MN for a while, I quickly learned that their database simply sucked.

    I would frequently look up books on amazon at home, this was before mobile web, to get ISBN and use that to find the book on the B&N system so I could order them. Sometimes did this for customers because there was no other way to find them.

    IMHO brick & mortar bookstores are not designed to look up books, they carry what they carry and sell that. Lot of sales are impulse sales they don’t really cater to hardcore readers.

    Amazon has forced them to try to adapt, but their whole system (infrastructure) isn’t really designed for it.

    I don’t think it is something that will be changed at brick & mortar stores. Simply isn’t cost effective.

    I know when I worked at B&N we did a lot of special orders, because of the Mayo clinic we had as much sometimes more orders that shipped over seas as the east coast B&N stores.

    Pretty much all the fulltime staff there, myself included, had one to three foreign customers that would call in to order books from “their” bookseller. If that person wasn’t working when they called in they would find out next shift their person would be working and call back to order with that person even with the cost of the phone call!

    We couldn’t even get typos in the database corrected with all that business. I still have trouble spelling Modesitt’s name correct because several of his books were under typo version of his name.

    Don’t know if that helps, this was around the time “Plan B” came out, I know that was when I got that book. Think I had a copy on the shelf of the store for a while. I know I personally introduced several customers to your books while working there.

    Sorry for the rambling post, just a long time fan, hope things with Baen are going better than with your previous publishers.

    As a reader Baen is my hands down favorite publisher, I have bought most if not all of your books, since I realized you were there, as ARC (Advance Reader Copies). Hope that doesn’t hurt your commissions, but I don’t have to wait as long!

  2. I’m both a techie and a reader. 🙂

    I build databases for a living and can easily see why this happened (and while not exactly the situation, I’ve had this problem).

    Data modeling is a balance of many factors including development time, run-time performance, data storage – combined with understanding customer need.

    I think you have mentioned in the past that you use excel. You could build a database of books in Excel with lots of columns – which database people would call a very ‘flat table’. You might have a single column called Author and you could put anything into it. Each book would be a row and in your case this would likely have “Sharon Lee, Steve Miller”.

    In databases though, this would be space inefficient. What they would do is create a table that would have one row per author. Than in the books table, there is a pointer to the row in the author table. The key in the book table takes up little space. It also makes it very fast to find all books by Sharon Lee.

    Because we are anal, we likely designed the Author table to have separate fields for names. First = Sharon ; Last = Lee.

    As you commented, this is obviously wrong. It was fast and easy for the modeler to setup (and for the programers to use), but it has flaws.

    There are several options
    A) Have a co-author key in the book table (optional) and have it point to authors table. That way Steve Miller could get associated.
    Con: Doesn’t support multiple co-authors like we see with Mercedes Lackey, Eric Flint, and Dave Frier (although if you go to Amazon, you will see they just list the first two in the header).

    B) Have the author actually be the collaboration
    Single key in book table, but the author would be displayed as ‘Sharon Lee and Steve Miller’. The collaboration is really a new entity.

    Con: Makes it harder to find just novels of Sharon Lee (like Barnburner) and if you modeled first, last names as seperate fields…how do you type Sharon Lee and Steve Miller. 🙂

    C) Have a table in the middle between books and authors
    Books no longer has an author key. Instead there is a table in the middle called BookAuthors which has a key pointing back to Author and another one pointing to Books.
    This allows multiple authors per. It also probably needs an order number (to know to put Mercedes before Eric before David) and possible a type if you need to find primary author (aka first one).

    Con: This is the right way to model, but it is possible to now have a book without an author (in the data model at least). We would rely on the developers to enforce it…and problems have occurred in past with bugs.
    It is also harder and more complex.

    So – C would be the way to go to ‘Do it right’. But due to its complexity and desire to go fast, it didn’t happen. Changing it now is much harder since applications have been built around the original model.

    (note, this is just guesses, but as mentioned, I’ve been in this situation before where I’ve had to explain to the stakeholder on why we are in the situation we currently are in).

  3. Wow. Thank you for the clear explanation of the process!

    And this? Is why I write books, not databases.

  4. Have you considered contacting the customer service department at the vendor and explaining the problem and asking them that your account be marked to always ship from their U.S. warehouses? It is hip to anonymize sellers and pretend that they are faceless and conscienceless, but if you actually investigate the situation you are often pleasantly surprised.


Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.