atom feed35 messages in org.foaf-project.lists.foaf-protocolsRe: [foaf-protocols] fp:ping
FromSent OnAttachments
Story HenryApr 14, 2010 2:00 am 
Toby InksterApr 15, 2010 12:41 am 
Story HenryApr 15, 2010 2:09 am 
tyler gilliesApr 15, 2010 2:17 am 
tyler gilliesApr 15, 2010 2:19 am 
Mischa TuffieldApr 15, 2010 4:05 am 
Toby InksterApr 15, 2010 7:13 am 
Kingsley IdehenApr 15, 2010 7:20 am 
Kingsley IdehenApr 15, 2010 7:27 am 
Jiří ProcházkaApr 15, 2010 7:41 am 
Peter WilliamsApr 15, 2010 9:41 am 
Toby InksterApr 15, 2010 10:38 am 
Kingsley IdehenApr 15, 2010 11:13 am 
Story HenryApr 15, 2010 11:22 am 
Peter WilliamsApr 15, 2010 11:54 am 
Story HenryApr 15, 2010 11:55 am 
Kingsley IdehenApr 15, 2010 12:05 pm 
Toby InksterApr 16, 2010 4:28 am 
NathanApr 16, 2010 5:12 am 
Toby InksterApr 16, 2010 6:33 am 
NathanApr 16, 2010 6:44 am 
Story HenryApr 16, 2010 7:09 am 
Story HenryApr 16, 2010 7:40 am 
Toby InksterApr 16, 2010 8:58 am 
Danny AyersApr 16, 2010 9:26 am 
Story HenryApr 16, 2010 11:42 am 
Mischa TuffieldApr 16, 2010 12:10 pm 
Story HenryApr 16, 2010 12:14 pm 
Kingsley IdehenApr 16, 2010 12:29 pm 
Story HenryApr 16, 2010 12:58 pm 
NathanApr 16, 2010 1:02 pm 
Story HenryApr 16, 2010 1:16 pm 
NathanApr 16, 2010 1:33 pm 
Sebastian TrampApr 17, 2010 2:38 am 
Story HenryApr 17, 2010 5:27 am 
Subject:Re: [foaf-protocols] fp:ping
From:Toby Inkster (ma@tobyinkster.co.uk)
Date:Apr 16, 2010 4:28:20 am
List:org.foaf-project.lists.foaf-protocols

On Thu, 2010-04-15 at 19:55 +0100, Story Henry wrote:

I think the only thing I am really against is the use of the xml/rpc. Perhaps I can convince the semantic ping back people to enhance the ping back service to enable application/x-www-form-urlencoded results.

Hmmm... makes me think of ConNeg in reverse. It seems to me that HTTP 1.1 offers all the syntax needed to allow the *server* to negotiate a request entity media type from the *client*.

Client says:

POST /foo HTTP/1.1 Content-Type: something/the-server-does-not-accept

Server replies:

HTTP/1.1 415 Unsupported Media Type Accept: something/the-server-accepts

Client says:

POST /foo HTTP/1.1 Content-Type: something/the-server-accepts

Technically speaking, the Accept header isn't supposed to be used in responses, but it seems like a nice place to put a hint to the client.

Provided that Pingback's XMLRPC and Henry's form post idea use different media types (which they do IIRC), then this pattern ought to work, allowing them to each use the same endpoint URI, and same discovery headers.