It’s hard to say which way to interpret that statement in a never-deployed (well, deployed but never approved for deployment) spec, but I would say that actually says that only one is allowed. Clear or not in 0.92 and 2.0, that unambiguously says that 0.92 only allowed one. 0.93 thought about allowing more but never left the gate, and 2.0 didn’t change anything about 0.92 that would alter it only allowing one, if it only allowed one, which it apparently did.
It’s hard to say which way to interpret that statement in a never-deployed (well, deployed but never approved for deployment) spec, but I would say that actually says that only one is allowed. Clear or not in 0.92 and 2.0, that unambiguously says that 0.92 only allowed one. 0.93 thought about allowing more but never left the gate, and 2.0 didn’t change anything about 0.92 that would alter it only allowing one, if it only allowed one, which it apparently did.
More important than what the spec (ambiguosly in the case of 2.0) states is what it should state. Only allowing one enclosure renders RSS 2.0 Enclosures effectively useless for representing blog entries.