My most in-depth discussion yet about SpokenWord.org has just been published on Technometria with Phil Windley and Scott Lemon. As the show’s description below suggests, we also got into a discussion of many of the programming challenges of this project.

Doug Kaye joins Phil and Scott to discuss the recently launched SpokenWord.org, a free service that helps you find, manage and share audio and video spoken-word recordings. In addition to giving a basic description of the site, Doug also discusses the technical aspects of the project, including how it was developed and what kind of challenges he is facing now that it is operational.

Visualization courtesy wordle.net.

Chuck Joiner just published an interview with me on his MacVoices podcast about SpokenWord.org.

The day is finally here. No more excuses. No more alpha or beta. It’s time to open the doors. SpokenWord.org is ready for prime time and ready for you.

If you’re a regular reader of Blogarithms, you’re probably tired of hearing about SpokenWord.org, but if you’re a newcomer, here’s a portion of the press release:

There are perhaps millions of audio and video spoken-word recordings on the Internet. Think of all those lectures, interviews, speeches, conferences, meetings, radio and TV programs and podcasts. No matter how obscure the topic, someone has recorded and published it on line.

But how do you find it?

SpokenWord.org is a new free on-line service that helps you find, manage and share audio and video spoken-word recordings, regardless of who produced them or where they’re published. All of the recordings in the SpokenWord.org database are discovered on the Internet and submitted to our database by members like you.

SpokenWord.org has been ten months in the making, and like any such undertaking there are many people who contributed to its successful launch. I’d like to use this opportunity to thank just some of them and tell a bit of the story behind SpokenWord.org.

In April 2008 we held meetings of our Board of Advisors and Board of Directors in Cambridge, Massachusetts. It was approaching the fifth anniversary of IT Conversations, and hence a good opportunity to review our mission. In the early days (more than a year before the birth of podcasting) we had no choice but to do everything ourselves — recording, post-production and publishing — and we were virtually alone on the Web. But today there are tens of thousands of podcasts, and it’s now de rigueur for conferences to post their audio and video recordings on line. Add to that the podcasts from public radio and universities, and it’s clear that anyone can now be a publisher. Looking at the big picture of spoken-word content on the Internet, the greatest need (and hence the potential for the greatest public benefit) has shifted from production and publishing to helping people find, organize and share the programs published by others. That is why we created SpokenWord.org.

Present at that seminal meeting were directors Jake Shapiro, Jon Udell, David Weinberger and advisors Dan Bricklin and Bob Lyons. The big Aha! came from Jon, who has continued to be an incredible inspiration to the project. Jon also introduced me to Lucas Gonze and Hugh McGuire, both of whom have graciously given me the benefit of their been-there/done-that experience.

Finally, I want to thank the active alpha testers who not only succeeded in breaking everything I wrote, but were also kind enough to provide the constructive criticism that got us to the version 1.0 release. Most notable among the nearly 100 active testers were David Marks, Bruce Sharpe, Steve Williams, Ken Kennedy, Joel Tscherne, Rashmi Sinha and Thilo Planz. (I feel like I’m delivering an Oscar acceptance speech and they’re trying to get me off the stage.)

Working with this awesome team of advisors has made this project one of tremendous personal satisfaction. I hope you enjoy the results of our efforts.

Getting a logo for SpokenWord.org via a contest on 99designs.com was quite an interesting experience. We offered a prize of $275 and received well over 100 proposed designs from perhaps 20-30 designers. The quality was better than I expected, and the choice was difficult. To make my life easier (or so I thought!) I posted a short questionnaire on SurveyMonkey.com and invited members of our team as well as all of The Conversations Network’s registered members for their opinions. It was on short (24hrs) notice, but we got more than 100 survey responses.

My first surprise was that the opinions were all over map. There was virtually no consistency. Maybe one or two designs were disliked by all, but the respondents’ first, second and third choices were evenly distributed across the others.

The second surprise was the number of angry messages (some *very* angry) about the whole idea of holding a contest at all. Little did I know what a can of worms I had opened. The gist of the complaints is that designers should not be asked to submit designs on spec; that any good designer should be paid for their work. One email referred to a web site that specifically trashed 99designs.com. One one hand, I see their point. OTOH, there are many professions in which ‘spec’ creations are the norm. Consider musicians and fine artists, for example. Or real estate developers. Hey — entrepreneurship and the entire venture-backed world is all about speculation and taking risks. Yes, there are commissioned-work oppotunities for artists and some houses are built-to-order, but they’re the exception in those businesses. I could have tracked down a designer and reviewed their portfolios. In fact, most of the previous logos for The Conversations Network were created by Dorothy Yamamoto, who is now retired. But I decided to give the 99designs.com site a try this time, and I’m glad I did.

The winning designer (‘mithrill’ — I don’t know if he wants his real identity to be published) got in touch immediately after the contest. We communicated via IM and we collaborated on a few minor adjustments. Within two hours after the contest, he had sent me the files, palettes, etc. Very professional. Next step is to integrate his logo and the color palette into the SpokenWord.org alpha site. Then onto a complete site design for the public beta launch for which I intend to return to 99designs.com. My flak jacket straps are cinched. I’m ready for the attack.

In about an hour from now the SpokenWord.org servers will have ingested 37,897 new programs from 2,011 RSS feeds. They’re all from LibriVox.org, an awesome and fast-growing collection of volunteer-read public-domain books. More than 2,000 of them! Special thanks to Huch McGuire and Chris Goringe for their help with this wholesale addition to our database and for creating and operating LibriVox.org.

Importing LibriVox.org is a great pre-beta test of the scalability of many components of SpokenWord.org. There’s one peculiar MySQL oddity that has already been triggered by this process. I used MySQL’s fulltext search to search programs by iTunes category. All of the LibriVox.org programs are in the category Arts:Literature. As of fifteen minutes ago — the ingestion is still running — the LibrVox.org programs exceeded 50% of all programs in our database. The quirk is that MySQL can’t find any words if more than half of the rows in the table contain that word. That’s why the top-level category Arts currently shows (0) programs. Ih fact, there are nearly 40,000. I’ve got a kludgy workaround in mind that I may implement until such time as LibriVox.org once again accounts for less than half of the programs in the system. That’s going to take a while since there are new chapters being added to LibriVox.org every day.

Just two days left to go in the contest on 99designs.com. for a new logo and color palette for SpokenWord.org.  (Here’s a version without the eliminated submissions.) We’d love your feedback on these designs. You can add your comments here or on the 99designs.com site.

There are four designers still in the running with a number of good designs.

• ‘mithrill’ has submitted many good designs. #2 is the original. I steered him away from the Halloween-ish colors and the reflection, but I wonder if the reflection shouldn’t be brought back. It looks cool. I was just worried about the clutter. #36 is his first with the .org turned on its side, which I like. #36 though #43 are primarily just different palettes. #51 has a new palette and shows three different alignments of the elements. I like the square one in the upper left, but I also like the one in bottom center for situations where we need something more horizontal. The upper-right version doesn’t have the abstract ‘W’ so it loses that concept. Same for #52: looks nice but I miss the second set of bars. #41 and #43 are still on the page just for the sake of different color choices. I don’t like the bars slanting towards the top right as much as the original design. In summary, I think the design if #36, #37 and #38 may be mithrill’s best (but without the dashed line). All he needs is to work on colors, and we have to help him.
• ‘A d’ has submitted #3, which remains one of the most popular according to email I’ve received from the team. My primary concern is that when it’s small, it’s very hard to see/read SpokenWord. Also, the ‘.org’ really does look tacked on, to me. This was the first of the designs with a bubble. Others have copied it, but not as well.
• ‘customlogo’ has two left. I don’t think #19 is a serious contender, but I do like its clean look. #21 is also very clean and I like the use of color. The pattern of dots makes me think of both ‘digital/bits’ and of a VU meter. Elegant.
• ‘yavan’ submitted #57. Also probably not a serious contender, but again it’s simple. The bubble concept is there, but it’s the container for the whole thing. Very nice.

Alpha 0.6 of SpokenWord.org now seems fairly stable. The biggest change since 0.5 is the deployment of Categories as I blogged about a few weeks ago. I decided to copy the iTunes taxonomy after all for a few reasons.

1. Although the sub-categorization is poor, the top-level categories are reasonable.
2. All taxonomies are flawed.
3. You can spend the rest of your life trying to tweak a taxonomy.
4. No matter what you do, a lot of people won’t like your sense of organization.
5. I can pass the blame onto Apple.
6. Since most programs are submitted to SpokenWord.org via RSS and with <itunes:category> elements, we get automatic categorization of the majority of programs.
7. When we generate RSS output, there’s a decent chance we can automatically include an <itunes:category> element that iTunes will accept.