primarily a problem if people are trying to directly access files in S3 using copy-and-pasted links from a site that doesn't maintain case on the link text (kchung's sites, and any copycat tumblr subdomains, currently force lowercase, for example). if case / spelling / punctuation are not exact, raw xml is returned showing "access denied" - could we try returning fuzzy matches (allowing case-insensitive results, for example)? or maybe a custom error result that gives better / more friendly options for troubleshooting?