|
Refback |
Trackback |
Pingback |
Webmention |
Trigger mechanism |
Visitor to linking site clicks on the link, and their browser takes them to the linked site |
Code on linking server examines added or updated documents, extracts links, and sends notification to linked server for each link found |
Code on linking server examines added or updated documents, extracts links, and sends notification to linked server for each link found |
Code on linking server examines added or updated documents, extracts links, and sends notification to linked server for each link found |
Notification medium |
HTTP referrer value |
HTTP POST[1] |
XML-RPC call |
HTTP POST with source and target parameters[2] |
Capture mechanism |
Examination of incoming HTTP referrer values |
Trackback capture script |
XML-RPC function |
Webmention capture script |
Information sent by linking server |
None |
- Linking site name (Optional)
- Linking post title (Optional)
- Linking post excerpt (Optional)
- Linking post URL
|
- Linked post URL
- Linking post URL
|
- Linked post URL (target)
- Linking post URL (source)
|
Additional information presented to linked server |
HTTP referrer sent by a visitor's browser upon clicking the link |
IP address of linking server |
IP address of linking server |
IP address of linking server |
Autodiscovery mechanism (how the linking server finds out how and where to send the notification) |
None |
LINK tag in the header of the linked page or trackback RDF documents |
Special HTTP header or LINK tag on the linked page |
HTTP Link header or link element on the linked page |
Action required when notification is received |
- Extract referrer value from incoming HTTP headers
- Retrieve referring page
- Parse retrieved page for desired information
|
- Gather desired information from
- Given parameters
- or retrieving and parsing the given URL
|
- Retrieve page at "linking post URL"
- Parse retrieved page for desired information
|
Verifying that linking page does indeed link to linked page is recommended, not explicitly required |
Advantages |
Requires no special code on linking server (the link itself becomes the notification when someone clicks on it) |
All the information desired by the linked server (Linking site name, post title, excerpt) is present in the notification itself |
- Notification mechanism has a complete technical specification
- Less susceptible to spamming
|
- Uses well-known parts of HTTP wherever possible (autodiscovery, encoding of data, response status)
- Reuses pingback's existing semantics
- Minimum amount of data transferred on-the-wire
|
Disadvantages |
- No notification unless someone actually clicks on the link
- Relies upon visitors' browsers sending proper HTTP referrer information
- Linked site must retrieve and parse linking site's page to extract the information it wants
|
- Notification requires positive action by linking server
- Notification mechanism has only a partial technical specification
- Autodiscovery information may prevent XHTML validation
|
- Notification requires positive action by linking server
- Linked site must retrieve and parse linking site's page to extract the information it wants
- Can be abused for DDOS attacks[3]
|
- Notification requires positive action by linking server
- Linked site must retrieve and parse linking site's page to extract the information it wants
- Relatively new, so less widely implemented
|