intertwingly

Sam, do you have the capability to send me a trackback for this post? I pick up my own trackbacks ok, but I don't have any from external sources yet...I'm wondering if no one has sent one, or if my code doesn't interop. :-

Greg, I already did.. I posted it as a comment on the entry that I was trying to trackback. My guess is that the problem is a mismatch between "URL" and "url"...

Ok...looks like you were using HTTP GET, and I only supported POST. I changed my implementation to support both - can you try it again?

Yup, my linkback parser is "million-dollar code". Has to be: the markup it takes as input is not under my control. Code is all I got.

Greg, it looks like that was it. To support old clients, I do a GET if the URL has a "?" in it. Otherwise I do a post.

Right on...very cool. Thanks Sam! Just out of curiousity, have you noticed if most of your incoming trackbacks come in as GET's or POST's?

Here's a philosophical question about trackback that's been on my mind since I first heard about it -- and it relates to the lazyweb idea as well.

If you want to get everything to coalesce, why not just use discussion group software?

I don't see what the pings get you.

Take it out a few years, do you think trackback will thrive and morph into something else? If so, what will it look like in a couple of years? And don't just say "very different" -- specifically what will it look like?

Dave, I don't know how to visualize it - most of my strengths are in backend text processing, so I won't really comment on the overall UI of the future, but here is how I see it.

We all chase our referrers. But if somebody out there writes something that catches our interest, wouldn't it be nice if we could chase THEIR referers?

Now realize that all of us post a mix of gems and garbage. Of course, no two of us will agree as to which is which, but that's what makes each of us interesting.

Radio's Referers pages tell us who found something interesting on this entire site, but to find out which thing they found interesting, you have to explore each one.

Mark's "Further Reading On Today's Posts" narrow this down to the granularity of a day, and makes an educated guess as to providing relevant contextual information.

Pingback narrows this down to the granularity of an individual post.

Trackback eliminates the need for an educated guess - the person issuing the trackback can provide the excerpt.

Of course, Mark's approach has the advantage that he doesn't require any cooperation others to make this work.

And all this is possible with blogs vs. instead of them.

Interesting. I'm not sure I want anyone else looking at my referers, although if you know where to look (lots of people do) they're out in the open with no password.

Also, I've rarely had trouble figuring out where on the page they're pointing to me from. When in doubt a view-source does the trick.

I like Mark's way of doing trackback. And I've tried writing code to do trackback, with mixed results. I think this is because there's two trackback protocols out there. Not sure.

I'm watching with interest to see what happens.

Dave, whether you want it or not, it is going to happen. For example, see:

http://www.technorati.com/cosmos/links.html?rank=&url=http%3A%2F%2Fscripting.com%2F

People have no trouble figuring out the context from a view source, programs have a bit tougher time guessing which permalink goes with which textual item.

Sam good point, except Technorati is missing an important bit of data, how much flow each link sent. An inbound link from Glenn Reynolds is worth a lot, and there's a huge diff betw a link from Glenn's blogroll (worth about 2 or 3 visits a day) vs a link from an item (several hundred).

Also, after a certain amount of time trackback is going to degrade, just like the DG on Scripting News did, with people flaming and the flames chasing away the thoughtful comments. It's one of the laws of open comment systems.

technorati has the ability to rank by blog authority...

And who is to say that the comment system remains completely open? For example, I know that Mark has some filters against spammers.

Finally, as they say, in the long run, we are all dead. So, while all comment systems may be destined for an ultimate heat death, in the meantime we can enjoy life. I, for one, have thoroughly enjoyed this comment system.

I enjoy your comments too, hey I post here. The signal to noise ratio is quite good.

Just for the record, my "further reading" system pinpoints incoming traffic to a specific URL. In my case that is a day's worth of posts, since URL==day. (Access logs don't receive the fragment identifier that would identify the specific post within the day.)

But if I had individual entry pages instead of multiple-entries-per-day, my "further reading" script would pinpoint flow per post, since URL==post.

I'm not sure if I'm making any sense. It's very late, and I'm very tired. The coarse granularity of my script is a function of my templates, not a function of the script. The limiting factor is the access logs, which only give page URLs, not fragment identifiers within pages.

Regardless, I like my way best, because my script also generates an RSS feed of the referrers to recent posts. I don't have to chase my referrers; they come to me in a nice little package.

For those who don't know, my "recent referrers" RSS feed is at http://diveintomark.org/xml/furtherreading.xml

Other available feeds are listed here: http://diveintomark.org/xml/

My trackback and pingback referrs are also available via RSS, as well as my comments:

http://www.intertwingly.net/blog/comments.rss

My assessment of the differences is that you don't get to see the links that are either never traversed or are only traversed by means that don't provide accurate referral information. And I don't get to see (via this mechanism anyway) entries that are produced by weblog authoring software that is neither trackback or pingback enabled.

Additionally, one can argue that trackback excerpts have a greater accuracy.

In the final analysis, however, I like yours better because trackback and pingback systems enabled systems aren't widely deployed yet. The reason I haven't developed and/or deployed such a system on my blog is that it seems to me that it would require continuous maintenance, not so much on the html parsing side, but to weed out the continuous growth of aggregators and republishers. For example, this one:

http://www.mdms.org/sam_ruby.html

Mark, you were making perfect sense. But then I already understood how your script worked, and how your weblog works.

What Sam just said about aggregators/republishers is what I'm thinking as well.

I get a lot of referrers from aggregation sites (like http://www.123aspx.com), which are useful, but the information on the referring page typically does not directly comment on my post. I would rather separate directly related information from other incoming links.

So I like to think of pingback and trackback as _intentional_ content-enhancing links, whereas other links from referrers may or may not contain additional content of interest to the readers of the post in question.

Although, Mark, I have to say you do a fabulous job of picking out the relevant content...I'm going to have to take a look at your code. :-)

Dave: "Take it out a few years, do you think trackback will thrive and morph into something else? If so, what will it look like in a couple of years? And don't just say "very different" -- specifically what will it look like?"

It'll look like a cheap way to have backlinks on the web. Backlinking was the critical thing TBL took out of his hypertext model to get it to scale up (and by all acounts, to howls of disapproval from hypertext cognoscenti at the time). In closed/semi-open hypertext systems, like Wikis, backlinking is an important way to find information.

Where it could get interesting is when you run the the web backwards, by walking the trackback graph. And start throwing some numbers onto trackbacks and you'll have Google sweating in no time - running numbers over backlinks is precisely the value add they provide.

The holidays (and associated downtime) are over and its back to business. This is a summary of the interesting and noteworthy news and posts that where made while I was out.... [more]

Trackback from tima thinking outloud.

Pingback from TIG's Corner

Please check the pages about...