Not a bad summary, but the idea of Modularity is just not achieveable due to practical aspects of how the player actually works in normal operation. The guys have looked at this (remember the big discussions about the "plug in decoder" architecture?) and it's just not on, otherwise they'd have done it! They are pretty smart cookies - if it's not been done so far, there's a good reason for it.

No, we are better off focussing in stability and correction of dysfunction. A final release version of V3 with all the minor playlist issues, caching, graphics, memory allocation, radio tuning and RDS to get a good meat-and-potato release of the quality of V2 Final, along with a revised disk builder seems to me to be the way to go. It is work based on established material, and has a better chance of success.

Why not an effort to work on the open source players that have been floating around for a while? Well, indeed - but why hasn't it been done by now? As much as I feel this is an alternative (Look at Mike's jEmplode), it has not happened. I am certain if I worked on material like this for 6 months, 40 hours a week, I could produce some excellent results. But I can't survive another 6 months without something to pay my mortgage and bills - I still would need funds. And why pay for an open source one when I could (potentially) work on the real thing?