Each CSV entry is between 50 and 67 chars long
To keep track I was storing the address that the next entry will start at, at address 0. That way when returning from reset the queue was always in a known state
Any flaws in or thoughts about this scheme?
And yes as EEPROMS are written in 32bytes pages (or multiple) you have to reuse pages The trick is to create multiple index slots, e.g. use 16 pages to write the start and end address of the new data.
This description indicates I'd be using a page to store a single index value (ie 32 rather than 2 bytes).
http://www.atmel.com/dyn/resources/prod_documents/doc2526.pdfHaven't read it, not sure if there's anything in it that hasn't already been discussed here, so FWIW.