Tavin Cole <tavin at mailandnews.com> writes: > Sounds like what you're essentially proposing is eliminating the > multi-part nature of the metadata standard entirely and just making > it one long FNP message per DocName where the distinct parts of the > old way become the top levels of sub-fieldsets.
That's what I'm proposing. > IMHO it would probably be best just to keep it multi-part and use > the DocumentName to associate the parts with each other. Whereas IMHO it wouldn't. > I just see this getting out of hand.. and actually being harder to > parse and react to programmatically. I don't see it getting out of hand -- it's really easy to parse Name.Subname=value. > It seems like a spurious concatenation It's not spurious -- we're semantically grouping that information together, so let's do it syntactically as well. > that's inconsistent with the general idea that metadata can be > sourced from more than one place and correlated by DocName (e.g., > metadata across two freenet keys or embedded in a key URI).
