I just finished listening to  Bloggercon III Podcasting session posted to IT Conversations.

There is a lot of things going on in my mind after listening to this twice – twice because of using the Muvo and my iPAQ while trying to drive, and do fast forwards was extremely hazardous to my health.  It was just better to listen again then be distracted while commuting to/from work.

One thing that has begun to resonate with me even before listening was that my audio.xml file is "my station" and one of the shows that I produce, edit, and author is "Tech Rag Tear Outs" that includes content that probably should be considered it’s own show on "my station":

• Technical Items
• Podcasting Thoughts
• Music
• Personal Insights (Faith, Family, Politics, etc.)

Since I don’t want to waste anyone’s time, I think that posting timecodes is one solution.  But there is also the option of just breaking out all the pieces that "my station" offers into individual productions, and subscribers can sort it out on their end to what one wants to consume.  This also shortens the production cycle on my end for what I consider my bread and butter – insight on technical items.  But it still means that bandwidth and storage will be used that may not be as efficient as possible.

Getting a handle on the structure of the METADATA needed for fast, effective, and useful podcasting is really a driver for this.  Key for me is to be able to offer one audio.xml file that users can pick and choose from all the content categories I am offering before downloading. 

To do this second idea, I think the client applications need to change along with the posted xml or I need to have seperate xmi files for each show?