Marshall Harrison - "the gotspeech guy"

A large app built throughout would be awesome.  I find i get alot more out of books using this method that little examples that really don't get detailed enough or leave out parts

1. There's a lot to cover in MSS 2007 (Managed Code, Workflow Designer, SALT, VoiceXML, Analytics & Tuning Studio, Simulation/Debugging, OA&M, etc.).  My suggestion would be to do a deep dive in one or two areas, as the Microsoft documentation typically is a great reference guide.  My preference would be VoiceXML, but I think MSS support in this area is still too immature to warrant an entire book.  Rather than a "how-to" guide, I'd like to see a book that discusses best practices and design patterns (with implantation examples) for build speech applications.

2. A good programming book should add value to the product documentation.  I don't like buying a book that just adds extra words and pretty pictures to the documentation that you can get for free.  Also, the book has to objective, and focus just as much, if not more, on the pitfalls of the technology than on the pie-in-the-sky stuff you can read in the marketing materials.  Of course, real-word examples of your failures and successes always make the material easier to apply.

3. In almost all cases, a book should use small code samples that can stand on their own.  Very few people read a programming book cover-to-cover, or have the time to understand a large code example.  I do read most of my tech books from cover-to-cover, but the books I like best are the ones that I can go back to and lookup a single topic that can stand on it's own.

4. This relates to #1, and what topics you decide to cover in the book.  If you focus on a dynamic markup approach (VoiceXML or SALT), I'd like to see basic architectural decisions like "client side vs. server side coding" -- discuss the pros and cons of both, and when you should use each.  

For tuning, I'd want to see a list of common things you should look for when deploying a speech app, and where/how you can look for them in the MSS tuning tools.  A common problem on all speech enabled IVR platforms is speech detection.  A chapter on how to troubleshoot/tune the speech detector would be great (processing your prompts, bargein types, speech detector configuration, UI techniques, etc.).

With the Managed Code API, I’d like to see a discussion on the benefits of this approach vs. others (markup, Workflow Designer).  For example, how does the strong typing of a managed code API compare with a loosely coupled markup approach?  And, what tools/utilities/techniques can you use to overcome any drawbacks associated with each approach?

And in general, I’d like to see a chapter dedicated to building open, standards based applications on MSS.  For example, which technologies/tools are best suited to taking advantage of the emerging speech standards (SISR, CCXML,  NLSML, etc.)?

One thing that I haven’t heard a lot of discussion around, or MSS documentation is on CTI integration.  A chapter on how to integrate CTI into your MSS apps would be great.  I’d like to see how each of the programming models (Designer, SALT, Voice XML, API) support CTI, and which vendors each of them support.  At a minimum, I’d expect to see specifics about Genesys and Cisco integrations.

Marshall, congrats and thanks on this huge endeavor!  I’ve never written a book, but from what I hear it can become a full time job.  And, in with tech books, there’s not a lot of fame or money.  Your passion towards improving MSS is greatly appreciated!

This is great (although I already knew).  One of the problems I'm having is b/c a lot more stuff is now public that wasn't when I started writing. But it's changed so much that it 'ought' to be covered. Back in the day when it was NDA it wasn't an issue, but each time it's made public, it needs to be covered.

You and Sara are both brilliant and awesome developers and I look forward to seeing your book ;-)

How is the book going? Can it be ordered soon?