Feed Sign in with OpenID OpenID

Simon Willison’s Weblog

More on Pingback

More on PingBack. PingBack can be seen in effect on this blog and in this entry on kryogenix.org. If you wish to ping my blog you can do so using the following XML-RPC details:

Server: www.bath.ac.uk
Path: /~cs1spw/blog/pingback/server.php
Method: pingback.ping(pageLinkedFrom, pageLinkedTo)

pageLinkedFrom must be the URL to a permalink archive page on my blog (complete with #fragment) and pageLinkedTo must be the URL of a page that contains a link to pageLinkedFrom. My server is a bit flaky at the moment (hosting problems again) so you may need to repeat the ping a few times before it gets accepted. Stuart’s PingBack server can be found at http://www.kryogenix.org/mt/mt-xmlrpc.cgi. Both of our blogs now include a link element pointing to our PingBack servers, with the rel attribute set to “pingback” and the href attribute indicating the server URL. The link elements should enable auto-discovery of PingBack servers in the future (an idea based on Mark Pilgrim’s RSS auto-discovery).

This is More on Pingback by Simon Willison, posted on 2nd September 2002.

Tagged

View blog reactions

Next: Useful XML-RPC links

Previous: Pingback implemented

9 comments

  1. Have you considered using an ordinary HTML header for the pingback purpose (something like 'X-Pingback-URL: <pingback url>') ?

    The neat thing of that system is that you can issue a HTTP HEAD request instead of a GET request, which gives you only the headers, bypassing the double hits problem.

    But good work ! looking forward to the code ;)

    Breyten - 2nd September 2002 14:53 - #

  2. Ah, but to do that you have to have control over the server, which a lot of people haven't got. HEAD is also not all that well supported across servers, which is a bit of a flaw. Moreover, the Pingback server checks the text of the referring page to ensure that it really *does* contain a link to the referred-to page, but you wouldn't know that because we haven't published how the protocol works yet ;-)

    sil - 2nd September 2002 15:24 - #

  3. I just tested pingback here and at kryogenics. Works great! :-) An odd difference between your weblog and stuart's, "pagelinkedto" and "pagelinkedfrom" have to be exchanged to work. Could you check it? Anyway, thank you for this nice piece of "remote" candy.

    mort - 3rd September 2002 13:33 - #

  4. I'll investigate the pageLinkedTo / pageLinkedFrom problem immediately - thanks for pointing it out :)

    Simon - 3rd September 2002 13:51 - #

  5. Issues:

    1. This blog's pingback always seems to say "only POST accepted here", whatever I send.
    2. The spec doesn't say what the corrent _response_ is! (Implementations seem to just send back an aribitrary string.)
    3. I assume implementations are naiive, and so would get confused by a commented out <link> element. Is that ok? Would people notice links like:
         <link title="pingback server" 
               href="relative/url/to/server"
               rel="talkback pingback"
               rev="subject" hreflang="en"/>
      
      ...? Because that's a completely valid link to a pingback server in an XHTML document, right.
    4. What's the official order for the arguments?

    Ian Hickson - 6th September 2002 05:19 - #

    1. I'm stumped on that one. I've dropped you an email about it.
    2. Good point - we need to sort that out. I don't think we're even throwing XML-RPC errors when a pingback fails (which would definitely be a god idea, we should standardise the error number as well).
    3. That's completely up to the client implementation. We're talking about changing the link element to point to an XML file describing the PingBack server rathjer than the server itself, but we'll bare those points in mind whatever we decide.
    4. pageLinkedFrom, pageLinkedTo (at least we've managed to sort that one out :) )

    Simon - 6th September 2002 08:59 - #

  6. I've tried to ping this page with a siple XML-RPC client. I've got this error: XML-RPC Fault #2: Invalid return payload: enabling debugging to examine incoming payload and I don't know what it mean. Anyone has an idea?

    Daniel - 24th August 2004 16:00 - #

  7. 7

    7 - 18th March 2005 12:18 - #

  8. http://conversion.accenture.com/global/research_an d_insights/accenture_blogs/Kelly_Dempski/AT_BlogPo st_Pass1.htm

    Manong - 21st February 2006 11:39 - #

Comments are closed.

Previously hosted at http://simon.incutio.com/archive/2002/09/02/moreOnPingBack

A django site