Let's start with the numbers. Commercial publishing houses in America produced 5 percent more titles in 2010 than the year before, according to the May 18, 2011, issue of Publishers Weekly. The count: 316,480 titles.

But non-traditional publishing — mostly print-on-demand public domain and self-published titles — rose to a staggering 2,766,260.

Since 2002, production of traditional books rose by 47 percent. Non-traditional production rose by 8,460 percent. As of last year, commercial publishing is a little more than 11 percent of the total.

The job of the library is to gather, organize and publicly present the intellectual content of our culture. But the mechanisms used by libraries to do that are a poor match for this explosion of literature.

The truth is, most public library processes are about gatekeeping: we try to assure, in advance, the quality or at least the popular demand for titles we purchase for our collection.

How? Well, commercial publishers let us know six months in advance what's coming. We read lots of review magazines, where people thumb through those early copies to let us know if they're worth the money. We track bestseller lists.

But self-publishing doesn't work like that. Independent publishers and authors may or may not have catalogs of upcoming works. Traditionally, they've had a hard time getting picked up by review magazines, too. Often, the independents didn't produce enough copies for national distribution.

But add in e-publishing, and "distribution" gets much simpler. Publishers and authors just have to upload one file to one server.

So I have an idea. To enable our public to try to sample this rich, untapped world of new writing, all we have to do is flip our processes upside down. We'll let you decide what our community should buy.

Since the end of 2010, the Douglas County Libraries has set up a powerful new infrastructure, one of the first in the nation. We can receive, catalog and manage electronic books ourselves.

Suppose we let authors and publishers upload their books to our catalog. Any author. Any publisher. It's just a simple online form. They would create a record that might not be up to our usual cataloging standards, but would suffice to help people find it.

Our new system will let our patrons, at their sole discretion, remember what books they read, and start recommending other books to them on that basis. A combination of that recommendation engine, plus virtual displays of e-content, plus mobile apps to put all of that in the palm of your hand, would make browsing our collection a lot of fun.

And patrons can rate everything they read, and even leave comments. Which other people can respond to.

Here's the twist. Somewhere down the line — six months, for instance — we consult our statistics. If a title hasn't been checked out at least three times, and if our readers haven't rated it at least 3 stars out of 5, we delete the file. Our community doesn't like it.

But if it did get used, and it did get rated well, then we buy it. In other words, we "crowd-source" our collection development to the people who pay for it: Douglas County taxpayers. We provide the technical system to sample and present the writing; you tell us what's worth keeping.

Now let me be clear about something: this process is going to mean we'll see a lot of wild stuff. Some of it will be very poorly written or edited. Some of it, unfiltered by the commercial presses, will truly be out there on the fringe.

And some of it will be wonderful.

I pitched this idea to a gathering of the Colorado Independent Publishers Association recently. One author said he really liked it: it's the Wikipedia model. Put the work out there, let everybody weigh in, and you wind up with something that's actually of far higher quality than you might have predicted.

This experiment would be something new in libraries. I'm ready to try it. Is Douglas County?

Jamie LaRue is director of Douglas County Libraries. LaRue's Views are his own.