Yeah, my weblog items are done with Movable Type, but I don't do the whole site with it. As I've mentioned, I publish the rest of the site using POT.py. However, I do expect that people link to those pages--why not have them link back?
So last night I wrote a nonweblog Trackback server in Python. Updated 12:21am: Only it doesn't work. At least, it didn't ping the pages I named here--so I need to figure why it didn't see the RDF, I think.
Comments
comment
Well, don’t know about why it didn’t want to do autodiscovery from your post, but it worked fine for me pinging a single page from a bookmarklet post.
The more Python I look at, the more tempted I get: that’s a whole lot of stuff being done in not much code.
comment
Because the code can’t get a title, description, etc for an arbitrary web page like a weblog tool can for an entry, the RDF is missing some fields. My guess is MT doesn’t consider them optional, but I didn’t find where to look in MT to find out if that’s so.
comment
I got half way through typing up a long explanation of where it calls from in CMS.pm, and what gets called in Util.pm, and what it looks for with what regex, when I finally noticed: you don’t have a dc:identifier, you have a dc:indentifer.
comment
That might be the trouble then! I did find the right place in MT when I looked today, by grepping for “discover”, but I didn’t see where something was demanded that I didn’t have. Was about to look in my code to make sure I was giving a dc:identifier at all, but got distracted.
So, thanks. ;)