fetchmail is quite able to write files to disc - examine the "sendmail" interface. And obviously, you'd never delete mail from the server until you had insured that it was safe on disc anyway. I was just planning on launching multiple instances of fetchmail for concurrent fetching. However, if it turns out it can do it itself, then that's great to. Fetchmail has been developed over many years, and has learnt to use usual stratigies to do things that sound simple, but are suddenly a bunch of no fun when people's varying implementations of POP servers come into play. Writing an interface that's sufficent for the vast majority of users to fetchmail wouldn't be anywhere near as complex as yours sounds. Fetchmail's errors are quite details and constructive to the solution. You forget that the vast majority of the logic of using a fetchmail-based solution has probably already had *hundreds of years* of man hour testing done to it. Are you suggesting that free software is buy default not as well tested as commercial software? That's the type of FUD that only Microsoft seems able to produce. If most people don't know about mail fetching past the fetch and send buttons, why is your clients configuration so incredibly complex?
Anybody who automatically thinks that any free solution to a problem will automatically be inferior to the expensive one is quite, quite deluded. Try again.