Re: [blogite] Before we update the spec...

Date view Thread view Subject view Author view Attachment view

From: Ian Hickson (ian@hixie.ch)
Date: Sun Sep 08 2002 - 01:39:38 BST


On Sat, 7 Sep 2002, Simon Willison wrote:
>
> This is currently achieved using XML-RPC, but should we open up the spec to
> alternative methods? This would make it a lot easier for people to
> implement things - for example:

I think we should keep the spec as simple as possible.

In order to make it possible for non-XML-RPC-enabled people to use
pingback, I have implemented two proxies: one which turns GET pingback
requests into XML-RPC requests, and one which turns XML-RPC pingback
requests into GET requests.

They are, effectively, pingback gateways for clients and for servers
respectively.

You can find out more information at:
   http://software.hixie.ch/utilities/cgi/pingback-proxy/

Feel free to use these scripts either directly through my server or by
downloading the source and using them locally.

> Or even a form on your blog:
>
> """
> If you link to this site and do not have a PingBack enabled blogging tool,
> please use this form to inform us of the link:
>
> Page Linked From: _________________________ (URL, including the http://)
> Page Linked To: _________________________ (URL, including the http://)
>
> """

Someone should just write a page to do that. This is just front end -- the
back end could either just use XML-RPC or go through my scripts. In fact
it would be trivial to write a single HTML page which had my client proxy
as the action URI.

> Both of these methods would have the same end effect. Own up, how many
> times have you sent your own PingBack server an XML-RPC ping telling
> it about a link on another site that probably isn't even aware
> PingBack exists?

Never. :-) All my manual pings have been me trying to fix bugs in my code.

> I'm afraid I'm still concerned with the overhead required for an
> auto-discovery client.

Having implemented it, I don't see that there is much overhead. (And I
haven't even bothered with optimisations -- I just slurp the whole target
file in one go and perform a regexp search on it to find a link tag.)

> If I post a blog entry with 20 lists I'm not too keen on my blogging
> tool having to make 20 HTTP requests checking for <link> elements.

There is no way a centralised server for pingback can scale. Just look at
how weblogs.com is doing. And they only have to worry about a single ping
per post.

Anyway, decentralised systems are always more robust than centralised
ones, for things like this especially.

-- 
Ian Hickson                                      )\._.,--....,'``.    fL
"meow"                                          /,   _.. \   _\  ;`._ ,.
http://index.hixie.ch/                         `._.-(,_..'--(,_..'`-.;.'
Message sent over the Blogite mailing list.
Archives:     http://www.aquarionics.com/misc/archives/blogite/
Instructions: http://www.aquarionics.com/misc/blogite/

Date view Thread view Subject view Author view Attachment view

This archive was generated by hypermail 2.1.5 : Mon Sep 09 2002 - 04:05:01 BST