Brave Tech World

November 19, 2006

SUN 19 NOV 2006

RSS is broken, and it is going to hurt you. By Marcelo

 

    I just wrote about the traps of writing a blog post and thinking that you can delete it later (short answer: no you cannot erase from the Internet).

 

    The problem with RSS is that is not a fully implemented protocol to notify content changes in a blog. If you think about a change notification system, you usually have three types of notifications: Add, Modify and Delete. RSS only supports the first two.

 

    To be correct, RSS only supports "Add", but with the "dc:dateModified" extension it also supports "Modify" notifications.

 

    Since RSS doesn't have "Delete", there is no way for you to control what Feed Readers or Search Engines are displaying to users.

 

Page vs. Blog Post

 

    If you have a regular page on your website called "mypage.htm" you have many way to control Search Engine behavior. You have robots.txt, you have META tags to prevent search engine from Caching it (or even indexing it), and you can always delete that page and the next time a search engine crawler comes looking for it the page will be gone and it will be removed from the index.

 

    With blog entries, you don't have that. Bloglines, admirably, is working on ways to improve that but allowing users to control how search engines will surface blog content, but that is about it.

 

 

What is missing..

 

    RSS needs to improve by providing the following:

• Support for "deleted" post;
• Support for Search Engine control (nocache, noindex, nofollow) on a per-post basis that can also be used to change the settings of old blog posts (not only the last X posts).
• Some type of authentication so blogs can control what is served. At least an "anonymous" vs. "authenticated" granularity.
• Support for Time-To-Live and real support for "Don't Publish Before [Date]".

    I'm not a top expert on RSS/Atom, but I'm implementing a service that makes vast use of RSS and we need a better way for users to control their content.

 

    I want the users of Sampa to be able to control if a blog post is for "family" only, and any member of the family can be using any RSS Reader and just enter their authentication info for that blog.