On Sunday, Dave Winer wrote about having to switch the links for his blog images to HTTPS so they would show correctly in Feedland.
I found myself nodding along when he said "They almost certainly have been showing up broken in other feed reading software, for years."
When I built /reader I added a feature especially for this – it parses the post content for images addressed as HTTP, saves a temporary copy to my server then replaces it with that new one but over HTTPS. Although it was scripting.com that prompted me to add this, I've found it coming in useful for other sites that (presumably) accidentally serve the odd image over HTTP.
Dave may have updated his image links but it's still a nice feature to keep for those "just in case" moments.
