Given this URL:

http://www.aquarionics.com/gallery/Gid_[and]_Suzi’s_New_Year_2003

IE does the following:

http://www.aquarionics.com/gallery/Gid_%5Band%5D_Suzi/’s_New_Year_2003

Now, I realise the escaping error in the generated URL was my own stupid fault, but the fact that IE automatically reverses any backslashes in a URL – to retain compatibility with Windows’ broken directory seperator – is interesting. It means, for example, we can do this:

@import url(”/assets/cssspecial-ie-stylesheet.css”);

and IE will load it (It will try to “fix” the broken backslash) where Gecko/KHTML will attempt to load a file called “cssspecial-ie-stylesheet.css” in the assets directory is interesting. New browser-hack?

This isn’t news, really. When the first version of the new, all accessible RNIB site went live (And I ranted about it) some of the links contained backslashes, and thus broke for Mozilla, and it’s still annoying, but it might be useful.

What would be really interesting would be combining this with an IIS server. Does the server resolve it as the right path on the system (The Windows one) or as the RFC 2068 compliant one?

My solution, by the way, was to rename the album to “New Years 2003” and leave the escaping problem until I’ve got time to fix it properly.