Has anyone tried this?
I wrote this up in Powershell real quick: http://pastebin.com/KTXP08J0
I did some googling (Not binging) and they don't have a easy way to access the archive. You would have to do some .Net programming and create a IE object and manipulate IE. I think if you set this script to run daily you would build a nice collection pretty fast.
Also look on line #9 you will see 'idx=1' you can increment this up to 15 to go back that many days before they remove it from this method of accessing it.
This helped alot and provided the JSON, XML, and RSS urls for accessing the data.
https://stackoverflow.com/questions/10639914/is-there-a-way-to-get-bings-photo-of-the-day
Hope this helps.
After a couple hours deobfuscating the JavaScript source, and taking a few shortcuts using the Chromium "Network" tools pane... I've got a minimally working bash script.
https://gist.github.com/smaeul/ab49dec49b0fa5e5eb45
It requires python for the math algorithm, but I'd already written the rest in bash, so I left it. Replace the "head -n4" on line 36 with "cat" to download all of them, instead of the first four. They should be in reverse chronological order. Note that some of the ID's don't have file names associated with them. This script downloads the full resolution images -- as if you had clicked the "download this image" link in the gallery.
Comments, complaints, improvements are all welcome.
Edit: I downloaded the first 200 ID's. 176 of them pointed to actual images. They take up 120MiB total space.