Peter and I had a great IRC chat about this topic. This is more than a formating issue. If the content of the entires isn't media that is usable in itunes (audio, video, images) then the construction of a media feed for the entire repository and even committees or collections is a costly iteration over content to derive the appropriate media entires.
I suppose a lucene search or Browse index could be constructed to optimize this. But, I tend to want to promote using Discovery/Solr to produce a truly powerful and open search engine for DSpace and that includes serialization of the users queries into various formats including those supported already by Solr (rss, atom, json,...) and those that our community may need, (opensearch, media and podcasting feeds, SRW, OAI-PMH, etc)
It is my opinion that using Solr for ArtifactBrowse/Search will allow us to replace a very large portion of the DSpace user interface with a very scalable and optamizable query solution that will seriously reduce need to interact with the database to construct views on DSpace content.
I'm attaching our transcript...