I would like to add some custom data to JPG images we are generating in our app. After much research I think the best approach is to include this information in the XMP packet and save it to the image, thus eliminating the need for side-car file.
Are these assumptions and ideas way-off base or am I close to being on the correct track for this?
You can't store XMP as pure XML. You need to use the XMP data model. Please read the XMP specification for more details:
The steps you described are correct. You will use XMPFiles from the SDK to open/close the files and get/put the data from/to the file. And you will use XMPCore as the datamodel to add custom properties before putting it back into the file.
Hope this helps,
Adobe XMP Team
Excellent, do you know the best way to contract someone to do this type of development. I would assume vWorker or Elance or similar but since this SDK is unique I thought I would ask.