Sword as a standard deposit mechanism

I'm still not coming off the fence as regards writing our own repository management solution, based around a triple store such as SESAME, but I'd like to throw SWORD as a deposit mechanism into the use case mix.

SWORD is the standard deposit api - the following is a link to a blog page by the lead developer where he's put a whole pile of slides online:

http://blog.stuartlewis.com/2010/07/25/the-sword-course-slides-online/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+stuartlewis+%28Stuart+Lewis%27+Blog%29&utm_content=Google+Reader

Supporting SWORD would

1. deposit content directly from inside of applications
2. allow the automated deposit of data and images etc produced out of a data processing workflow, including data capture from instruments
3. provide a standard interface for deposit
4. allow the generation of workflows to ingest data harvested from other datasources (ie if we decide we want the data as well as the metadata)

See also If SWORD is the Answer, What is the Question a research/discussion paper on SWORD