Pingback & trackback Anders Markstedt, Samuel Bengtson, Samuel - - PowerPoint PPT Presentation

Pingback & trackback

Anders Markstedt, Samuel Bengtson, Samuel Sjöberg

Background

๏ The blogging community ๏ The need to refer to others ๏ The need to make the referring site visible ๏ The emergence of trackback

The idea of the concept

Blog A Blog B www.a news

• 1. Blog A posts

some news

• 2. Blog B links

to A's news

• 3. B sends a

ping to A

• 4. The ping provides an

explicit link back to the referrer

Demo

Trackback

๏ Developed by Six Apart, the creators of Movable Type ๏ Explicit declaration of a trackback URL ๏ Client pings the server and receives a response in XML format ๏ The content type header ๏ Auto discovery of trackback URLs

Pingback

๏ Developed by Stuart Langridge with help of others in the blogging community ๏ Pingback was a reaction against the shortcomings of Trackback ๏ A transparent way of sending trackbacks, using standard techniques

Autodiscovery

๏ First, all links in the post is extracted. Then a connection is opened to each link, trying to find the pingback address ๏ Autodiscovery is done in one of two ways

๏ The X-pingback header ๏ Link element in the HTML header

X-pingback header

HTTP/1.1 200 OK Date: Sun, 1 May 2005 15:05:37 GMT Server: Apache/1.3.26 (Unix) Last-Modified: Sun, 1 May 2005 03:18:26 GMT ETag: "65044-15b9c-3a4ab102" Accept-Ranges: bytes Content-Length: 88988 Connection: close Content-Type: text/html X-Pingback: http://bob.org/xmlrpc

Link in the HTML header

๏ Scan the HTML header with the following regular expression:

๏ <link rel=”pingback” href=”([^“]+)” ?/?>

๏ If a link is found, the search stops ๏ If no link is found before </head> is reached pingback is not supported

Pingback scenario I

• 1. Alice posts a message to her blog. In the post, a link to Bob is included. The address

to Alice’s post is http://www.alice.net/bob and the URL she links to is http://www.bob.org/year.

• 2. Alice’s blogging software parses the post and finds the link to Bob’s site.
• 3. Now, a socket connection to http://www.bob.org/year is opened.
• 4. The software looks for an x-pingback header, but fails to find one.
• 5. It starts to scan the page in search of the phrase

<link rel=”pingback” href=”http://www.bob.org/xmlrpc” />

• 6. If the tag is found, pingback is supported and the socket is closed. If the tag is not

found before the page header ends, pingback is not supported.

• 7. If a pingback-server was found, the following XML-RPC call is invoked by the

software: pingback.ping(‘http://www.alice.net/bob’, ‘http://www.bob.org/year’)

• 8. If more links were found in step 2, the process is repeated for every link.

Pingback scenario II

1. Bob’s pingback server receives the XML-RPC call from Alice’s software, with information about which page the link comes from (Alice’s site) and to where the links is directed (Bob’s site). 2. Bob’s server now confirms that the address http://www.bob.org/year is found on Bob’s blog. 3. The server then connects to Alice’s post and confirms that is a post containing a link to his site. This is a security check to avoid false or erroneous pingbacks. 4. Optionally some content from Alice’s post is saved, e.g. the title and an excerpt of the text surrounding the link to Bob’s site. 5. If the site is verified in step 2, the information about the pingback is saved in Bob’s database.

Trackback vs. Pingback

Pingback Trackback

Embedded link description No Yes Autodiscovery Yes Yes * Automatic discovery of links Yes No * Supported when explicitly stated

How to implement?

๏ Both Trackback and Pingback can be implemented on the server-side with PHP , ASP or PERL etc. ๏ No matter which language, the solutions are compatible since standard protocols are being used

Blog B Blog A www.b Blog A Blog B www.b

*

Blog A Blog B www.b www.a

The WWW today

Links are unidirectional and there is no easy way of telling which sites are linking your site.

The pingback/trackback approach

Site A reports to site B, that a link has been created. Site B then creates a link back to A, completing the circle

Bidirectional utopia

The ideal situation where a link can be followed from both ends. The link metaphor is actually a misnomer since it indicates a two-way functionality, which does not apply to the WWW today Bidirectional links could for example be used to give an overview of the WWW topology

Type of links