Web Architecture trumps repositories

If research was posted on the web, using the fecking web architecture design that was laid down *years* ago, then repositories would never have happened. There is absolutely no need for them, apart from to make middle managers feel warm and fuzzy and occassionally, a little moist when the numbers of deposits pass some inconsequential milestone - 'omg, squee, I have 1000 items in my repository. That means that this university is doing some research!' Yeah, great. Whoopdee doo. There are some unreadable PDFs somewhere in a system. Oh, btw, don't lose the database if you are using DSpace, bye bye all data. Oh, and using Fedora? Sure, it's easy to get stuff, you know how to do RDF queries, right? Oh, sorry, it's not SPARQL, it's its own weird query language that does its own damn thing. Frontend for Fedora? Why would you want that? And EPrints? Everything you want to store is a document, right? Oh, and keep that db too!