Anne
van Kesteren: Sam catches that referrer, does some magic,
and displays an excerpt of the post in his comments, along with a
link and the title of the post that linked him. I believe he uses
the feed that is linked from the referrer’s post for that.
(Yes, I’m jealous and like this feature.)
Anne, If you have any questions, feel free to leave a comment
here, and I’ll try to answer them. As to the referers
“magic”, here’s an outline:
Twice an hour, I scan my referers since the last scan. I
also use the
Technorati
API to identify additional referers.
For each unique referer, I fetch the page and look for
autodiscovery links (max one per page)
For each feed, I scan for entries with a content that links to
one of my entries
For each such entry, I grab the summary, link, and title.
If no such link already exists from my referenced weblog entry
to that refering link, I sanitize and truncate the summary and post
it as a comment.
I do support a number of different versions of RSS, which
essentially amounts to a number of different synonyms for the xml
elements (example: description, content:encoded, xhtml:body).
Those with only one such element gets that element treated as if it
were both the summary and content.
People without autodiscovery don’t get linkbacks.
People with not-well-formed feeds don’t get linkbacks.
People with summary feeds or who strip out all html don’t get
linkbacks.
The code is
here.
I also support
pingback,
and trackback is integrated with my
comments.
On my todo list is revisiting my
email
logic to reject more of the lame spammers.
I was wondering how you did that. Especially when I saw you trackback my linkblog that had to have been tracked via Technorati (which is how I discovered that you had done the trackback in the first place...) Talk about your possible recursion loops!
* People who have a per-entry comment feed as their first autodiscovery <link> cause recursion on Trackbacks
If you ever feel like stopping that, and I can improve my feeds somehow to help, let me know: that feed template is probably long overdue for a refactoring (dunno, I haven’t looked at it in forever).
Phil - if there is any way I can distinguish between a trackback and a post or a comment, I would be glad to add a line of code to my logic; even if it is specific to your weblog. I have precedent where I have done this for others.
As it stands now, it is pointless for me to delete the autogenerated excerpt as it will likely reappear in the next cycle. ;-)
Hrm. I could make the permalink within my comments a <guid>, and the link of the pinging post the <link>, but if you ignore those then you don’t pick up linklog posts. Is there anything Atomic that would help us?
Sam Ruby reveals something that’s been puzzling me for a while: how he tracks referring pages without needing the likes of “Trackback"="http://www.movabletype.org/trackback/” or “Pingback"="http://www.hixie.ch/specs/pingback/pingback”. It turns out...
Heh. “But I can’t put an rss:author in the feed, because that’s an email address, and I don’t know the email address for a ping!”
Okay, if I didn’t screw it up too badly (and thanks to you and Mark and Joseph for making the validator that kept me from having to think about both what it should say and how), the number one feed is an Atom feed with an author Person construct for the pinging bit-person’s name and url. For now, pending criticism, it’s a summary with the alternate link going to the pinger, since I couldn’t see why anyone would really want to see it in my comments (where the summary would, oddly, be content rather than summary).
Okay, so now I’m interested into how we can find the full-post feed alternative to a page, when there are also comment feeds, category feeds and all that. Time for another microformat?
Mark: I have a overall comment feed and a overall comment page. I have separate comment feeds for each html page. I don’t do categories, but Rael does: computers/internet/weblogs/blosxom html and rss
Phil: excellent. I already had logic in place to handle such feeds, and so this morning I deleted the trackback echo from here and it hasn’t reappeared.
Meanwhile, I entries such as these will still be able to link back to comments found in your weblog.
Sam, I realize now that the problem is pages where the discovered feeds are general feeds, instead of true alternative feeds. If I recall correctly Atom fixed that, right?
Thanks for the write-up. It’s quite clear how it works now. I have to find some way to run a cronjob though or every time a user hits my site execute something from the command line (if possible) when it looks like a good referrer...
Sam Ruby: Sincerest Form Of Flattery Sam describes his linkback system. (tags: blogs dev hacks) Suggested settings for the Canon EOS 20D From the pros at Sports Illustrated (tags: dslr Canon 20D config)......
[more]
Sam Ruby: Sincerest Form Of Flattery Sam describes his linkback system. (tags: blogs dev hacks) Suggested settings for the Canon EOS 20D From the pros at Sports Illustrated (tags: dslr Canon 20D config)......
Sam Ruby explains how he does his referrer tracking. Note on “referrer” vs. “referer” usage. Only use “referer” to refer specifically to the HTTP “referer” header. Unless you’re literally talking about the...
As you may have noticed, I’ve been shaking things up around here a little. Sorry for turbulence in the feeds! I’ve moved to a new server, and I’ve been trying out Movable Type again, but with mixed luck. Why? Because I’ve...