Publishing mutually hyperlinked documents

Suppose Alice wants to anonymously publish HTML files A and B. Assume that file A contains a hyperlink to file B. Alice would like the anonymously published file A to retain its hyperlink to the anonymously published file B. To accomplish this, Alice first publishes file B. This action generates a Publius URL for file B, B_url. Alice records B_url in the appropriate location in file A. Now Alice publishes file A. Her task is complete. Alice now wishes to anonymously publish HTML files C and D. File C has a hyperlink to file D and file D has a hyperlink to file C. Alice now faces the dilemma of having to decide which file to publish first. If Alice publishes file C first then she can change D's hyperlink to C but she cannot change C's hyperlink to D because C has already been published. A similar problem occurs if Alice first publishes file D. The problem for Alice is that the content of a file is cryptographically tied to its Publius URL - changing the file in any way changes its Publius URL. This coupled with the fact that file C and file D contain hyperlinks to each other generates a circular dependency - each file's Publius URL depends on the other's Publius URL. What is needed to overcome this problem is a way to break the dependency of the Publius URL on the file's content. This can be accomplished using the Publius Update mechanism described in Section 3.5. Using the update mechanism Alice can easily solve the problem of mutually hyperlinked files. First Alice publishes files C and D in any order. This generates Publius URL C_url for file C and Publius URL D_url for file D. Alice now edits file C and changes the address of the D hyperlink to D_url. She does the same for file D - she changes the address of the C hyperlink to C_url. Now she performs the Publius Update operation on C_url and the newly modified file C. The same is done for D_url and the newly updated file D. This generates Publius URL C_url2 for for file C and Publius URL D_url2 for file D. The problem is solved. Suppose Bob attempts to retrieve file C with C_url. Bob's proxy notices the file has been updated and retrieves the file from C_url2. Some time later, Bob clicks on the D hyperlink. Bob's proxy requests the document at D_url and is redirected to D_url2. The update mechanism ensures that Bob reads the latest version of each document.