You can't really plan for a launch. Things always happen that you don't expect. It helps to have had a good night's rest, but other than that...
Our first unexpected event was that Library Journal ran a piece about our "soft launch" on their Digital Shift website, while we were in the process of deploying the website to production. They didn't link to us, but a few impatient readers typed in the website name and started exercising the site before we were finished testing the deployment. Nothing awful happened. Thanks, dave and gsf! Then Google spidered the site, exposing one or two errors. Thanks, googlebot!
We wanted our mailing list subscribers to be the first to see our work, and we finally sent out the email on Thursday. List readers discovered that our "popular" and "unglued" views were running very very slowly, loading down the site. Raymond studied the problem, and, as seems to happen so often with Django, found the answer hidden in plain sight (the documentation). After moving some nested queries, the pages returned 100x faster. The miracles of EC2 allowed us to spin up a bigger server to help with load. And the high load from the glacial queries helped expose some concurrency problems that we never would have found in a million years of normal operation. Or so says the errant coder- me.
We wanted to open the website when we did so that we could show our work to our many friends at the American Library Association Midwinter meeting in Dallas. So on Friday, I got up at 5AM (after bugfixing till 1AM) to catch a flight. I decided not to have any coffee so I could sleep on the plane. When I arrived at my Dallas hotel, I discovered another unexpected occurrence: I had left my laptop on the plane.
Before reading the next paragraph, check your laptop, your iPad, your kindle, your nook, or whatever. It probably looks plain, like mine (left). If there is no identification on it, go get one of those free address labels you got from the Awful Disease Foundation, and stick it on. Also some stickers from your favorite organizations. When you are done, it should look like @vmbrasseur's (right). Are you done?
Here's what I learned about lost MacBook Pros and airlines. Once the battery runs out, you can't even find a serial number. The typical baggage claim operation does not have geek squad backup. They don't have spare power cords or batteries to help them ID lost laptops. What they DO have is a safe, and that's where errant laptops go to die. If you ever find yourself in my position, go to the airport and ask the friendly lost-luggage attendant to go look in the safe. Otherwise, you will never see your laptop again, even if you have entered its serial number into the web form that has replaced the lost and found phone number that no one helpful ever answers.
In contrast, Jeanette, the DFW Continental Airlines baggage claim professional that I talked to in person, was very helpful. She called back to the guy with the safe, and we hardly had time to joke about the huge bag of dried fish from Africa that was smelling up the lost baggage area before safe-guy came out with MY LAPTOP. Yay!
Raymond had emailed me with a status report from the virtual home office, reproduced here in its entirety:
http://idioms.thefreedictionary.com/All+systems+goSo I was feeling pretty good. I got to the Convention Center and found Andromeda doing an in-person demo of unglue.it. The in-person demos are an invaluable complement to submitted feedback reports because they let you see expectation mismatch as well as outright failures. Andromeda seemed to have the demo drill down to a science. I am thankful for the generosity of our in-person testers, whose insights will soon be incorporated into the site.
And then we had dinner at Wild Salsa. (I'm skipping a few things here and there.)
On Sunday, Beth Kephart's article on Unglue.it went live at Publishing Perspectives. Beth writes so beautifully that it hurts. Her first book, A Slant of Sun is on my Unglue.it wish list. Her article introduced the Unglue.it concept to hundreds of new book lovers, more than a few rights holders, and generated a bunch of traffic. We've always expected that we'd need some publicity to find significant numbers of rights holders willing to take the plunge for a completely new business model, and the Publishing Perspectives article was a great start. Ed Nowotka's more cautious commentary is spot on, as well.
In the first week, the preview site has signed up 133 users (a conversion rate of about 10%) and we've received numerous suggestions for improvement. Our intrepid ungluing pioneers have added over 7500 works to our database. The most frequent comment is that we need better ways to indicate works that are already "unglued", either by virtue of being in the public domain, or by being already available under creative commons licenses. Raymond is currently working on loading Project Gutenberg titles; there will be more "unglued" books added as we go on, as well as ways of adding them directly. Coming in second were requests to have more selective imports from GoodReads and LibraryThing.
I should mention that we've had some great talent helping the core unglue.it team. Most prominent is the design work of Stefan from Design Anthem. We've had part-time help on systems and software from Ed Summers and Jason Kace. And it's hard to overlook the contribution of the countless developers who contributed to the open source Python and Django projects.
If you haven't tried the site yet, please give it a spin and tell us what you like (or dislike). The more people that sign up, the less skeptical rights holders with interesting books will be about the concept. If you're a rights holder or a rights manager of any kind, please contact Amanda at email@example.com with your ideas and questions. Follow @unglueit on Twitter, like unglueit on Facebook.