Quibbling about TrackBack

A few nits to pick with Trackback in the UserLand environment (which is, after all, a first or second draft, and thus may still be a bit malleable).

First, a tiny bit of background from my perspective. It’s important to remember that there’s more than one sort of TrackBack.

The most visible sort is post-to-post, where one post expands on, comments on, or replies to another post, and so the reply post sends a TrackBack ping to the original post (or, depending on the conversation, to several posts). That sort of ping is intended to give a reader of the original post a path to follow to later posts. It’s also important to note that doing so doesn’t actually require that the later post link to the original. Most of the time, it will, since the usual pattern is Jane posts, John replies on his weblog, and needs to include a little summary and a link to Jane’s post, so his readers will know what he’s replying to, but it’s not a part of the TrackBack contract. Suppose instead that John has a FAQ post, explaining how to do something, and Jane publishes a post asking the world how to do that something. By sending a ping from his FAQ post, John provides a link to Jane and any readers of her post to his answer, but there’s no reason to have a link in his post to her question: people already reading the answer don’t need to know the question. In that case, autodiscovery based on links in the entry isn’t enough.

Though that’s the most visible sort of TrackBack, post-to-category, or category-to-post or -category, are also very useful ways to ping. Post-to-category pings, where an individual post sends a ping to a collection of pings, rather than a particular post, is commonly used for conferences, like the O’Reilly Emerging Technology Conference, where someone just making an entry or two might send pings directly from those entries, either by autodiscovery from a link, or manually, but someone live blogging the whole conference would be more likely to just add a category for the conference, and have any post in that category ping the ETCON TrackBack URL. Category-to-category (or post-to-category) pings are also what powers things like the Internet Topic Exchange, or my pinging of the MovableBLOG sidebar. Used that way, it’s like a finer-grained form of aggregation: Richard wouldn’t want to publish links to my every post by aggregating my whole RSS feed, but by pinging just the Movable Type related posts, I can let him aggregate just the on-topic stuff.

That said, when I talk about TrackBack, I usually mean post-to-post, and that’s certainly true when I say, yet again, A TrackBack Is A Comment. I say that a lot, but

Outbound Trackback is easy for both products, inbound will be hard for Radio, where the content system may be behind a firewall. Some kind of proxy will be needed, perhaps a Manila site.

I get the feeling that even if some of the philosophy is getting through, the practical aspects aren’t. A TrackBack is a post to a script at a particular URL, including the ID number for what’s being pinged, and a name, a link, a title, and some text. A comment is a post to a script at a particular URL, including the ID number for what’s being commented on, and a name, a link, maybe an email address, perhaps a title, and some text. If you have a comment system, and want to support TrackBack, it’s just a matter of writing a script that sticks together the title and the body text, tacks a [more] link on at the end, and stores it as a comment. A system that stores them separate from comments, but can intermingle them, will be more flexible, but with a little tweaking and another script, RadioComments.UserLand.Com could be accepting TrackBack pings without much trouble.

If a content management application supports outbound Trackback, as Manila does, it can notify another site that a post points to a post on the other site. You may see a list of inbound Trackbacks in a separate window linked to the post.

I’d much prefer “a post responds to” or “a post replies to” rather than “a post points to”. We already have a way to know that a post points to a post on another site: that’s what referrers do. We also have a way to know in a cleaner, quicker way, getting just a referrer from the permalink location without anyone clicking the link: that’s Pingback. And although MT’s basic implementation (and perhaps Radio’s as well) is stuck using a popup window, because MT’s single-threaded approach can’t always afford the time to rebuild every possible page where the ping might appear before it replies and dies, there’s nothing specifically TrackBacky about that. You may see a list of inbound TrackBacks in a separate popup window, or listed at the end of the entry itself, or intermingled with (local) comments.

When you post a News Item to the home page of the site, if Trackback is turned on (by default it’s not) we scan the message text for URLs, and then read each of the pages.

Danger Will Robinson! I hope you’ve got some safeguards in place that you just didn’t happen to mention, or that Manila’s a lot better about being able to suck down enormous quantities of data: probably the most commonly reported problem with MT’s autodiscovery is “I tried to save a post, but the script keeps timing out. Well, yes, I did link to a 5MB MP3 with autodiscovery turned on.” Better to pick some length (I like 100K) that’s the maximum you’ll ever read before giving up. Especially if

If Automatic Pings is not turned on, we just do the auto-discovery work, and later, the user, when editing the message, may choose to ping sites that are linked to, on a case by case basis.

If I understand that right, that there are two Prefs, one to turn on TrackBack, and another to automatically ping, so that it’s possible to choose what to ping and what not to ping, by saving the entry and then editing to choose what to ping, I like it, but… only autodiscovery? A person who only occasionally wants to ping would have to turn on TrackBack, and turn off automatic pinging, and then have his site download every single thing he links to, just for the odd ping every now and then. Isn’t there any room on the entry page to slip in a little textarea for TrackBack URLs? Add a third Pref: Trackback on or off, Autodiscovery on or off, and Automatic pinging on or off (with Automatic pinging as a sub-pref of Autodiscovery, so that selecting Automatic pinging selects Autodiscovery), so that shotgun pingers could send pings everywhere without thinking, frequent pingers could download everything they link to and then choose what to ping, and selective pingers could just do a quick copy-n-paste when they want to ping. If not, well, it’s a lucky thing Radio users tend not to do too many big link-dump posts with links to ten or twenty things in a post, or the traffic while upstreaming, downloading everything linked to, and then sending pings, will get a bit ugly on dialup.


Comment by Dave Winer #
2003-04-27 02:31:06

I didn’t understand the first few paragraphs, when I got to the Danger Will Robinson part of your post (hope it’s not really a comment, sorry) we don’t suck down MP3’s (it cares what the extension of the url is), and the checking is done in a separate thread (as explained in the doc) so the user doesn’t have to wait.

There is a new text entry box on the edit-post page for entering an address to ping. This is being discussed ad nauseum on the Manila-Dev list, or you can wait until it’s released, I’d be happy to set you up with a test site on one of our servers so you can review it. It’s definitely still fluid.

I just wanted to give Manila users the feature that MT users have. Beyond that I don’t know, at this time, where it will go or what it will be used for and why Referers aren’t enough. (I kind of think they are.)

Now a feature reuqest back atcha. Could you increase the font size on this site. My poor eyes can’t read the text without me having to copy it and paste it into an email message. In the old days I would say just leave off the font tag, but if I said that some wiener boy would flame me for not supporting standards.

Comment by Phil Ringnalda #
2003-04-27 11:24:20

Executive summary of the first part: if a link is the only way to ping, you’re limiting your users; if you can’t ping a TB URL with every post in a category, you’re limiting your users; if you can’t accept pings for more than just single entries, you’re limiting your users.

A textarea solves the first one nicely, thanks. (Wish I had the bandwidth to follow Manila-Dev, too, but there are limits!). Manila doesn’t do categories, does it? So the second only applies when you switch over to Radio: someone blogging a conference will want to create a category for it, and have every post in that category ping a particular URL, without having to include a link to a HTML page with the RDF to autodiscover or paste in the TB URL for it in every post. The third, accepting category pings? I can’t see how to graft in into either model (though Manila outside a News Item weblog could use it, I’d think). Skip it for now.

Why TrackBack over referrers? I think TrackBack is for readers, and referrers are for authors. I do display my referrers, but you can see down at the bottom of the page what an ugly messy thing they are. So far, six different URLs for your link to here, and while I really appreciate your link, which has brought in lots of people with different ideas, is there any reason why someone who has already read this entry should go read your link? That’s what I’m getting at by saying that not every link should be a TrackBack ping: someone who read this entry should go read Roger’s very different perspective, where for him a TrackBack is a less-identifiable comment rather than a more identifiable one like it is for most of us, but if you and two dozen other people linked to this entry (I should be so lucky) with ”just a link” links like yours, and all pinged, so that there were 25 pings, and only one was a reply expanding on the topic, then very quickly readers would learn that there’s no point in bothering to click through every TrackBack. ”Just a link” links are vital, we can’t live without them, I’m ashamed of myself for not figuring out a way to do more of them, but if they start sending TrackBack pings, it will ruin TrackBack for readers, and that’s who it really adds value for.

Comment by Roger Benningfield #
2003-04-27 03:09:46

Dave: Just a thought… I’ve written many an app that feeds various doc types to the client via a dynamic page and a little fiddling with the headers. So unless you’re filtering out anything with a .CFM, .PHP, .ASP, or .JSP extension, you may still find yourself parsing through 500k Word files and MP3s, looking for non-existant RDF. And even that won’t help with those clever folks who serve up PHP with an .HTML extension.

Comment by Danny Ayers #
2003-04-27 03:48:21

Thanks Phil – this is very helpful outside of the Manila context too.

Speaking as more of a salami boy, the font size isn’t a problem because Phil supports standards like CSS! Anyhow, you can set the font size through Internet Options -> Accessibility with Internet Explorer, or hold the ctrl key and move your mouse scroll wheel.

Comment by button #
2003-04-27 03:55:23

Phil: Sorry to intrude on supergeekdom as I am in the category of Push Here, Dummy!, but I have been meaning to express to you that I wish there were a more universal trackback thing. It just came up late last night when I dropped into Dave’s Harvard weblog. I wanted to leave a ref to one of his entries, so I just ended up leaving my blog url and a note that I had also posted on that case (a guy got fired). I have tried looking into the Internet Topics thing, but it seems like not enough people are using it yet. Please post a response to this on your weblog. Thank you.

Comment by Phil Ringnalda #
2003-04-27 10:56:11

Well, there, that’s all you can do right now, since Manila isn’t accepting pings yet. If it was, yeah, I could do an HTML form with a tiny bit of Javascript that would let you just paste in the TrackBack URL to ping, your permalink, blog name, title, and excerpt, and send a TrackBack ping. For that matter, since most implementations still support GET, you can ping almost anything by just typing a (long and complicated) URL in your browser. But I haven’t, because most people think TrackBack should be reciprocal, that if you are going to ping you should also accept pings. If you think of TrackBack as a conversation, where they say something, you reply, they reply back, and so on, then pinging without accepting pings becomes you saying something to them, with your fingers stuck in your ears so you can’t hear anything they say in reply.

Comment by Dave Winer #
2003-04-27 04:06:19

Danny, standards are great, but bugs are reality. That’s kind of what I was saying the other day when Zeldman and Pilgrim cut me a new anal orifice.

If I change the font size, stupid ole MSIE ”remembers” it, and from that point on every window I open is in the bigger font size, making the letters jumbo-size which is a bad joke given all I wanted to was read Phil’s brilliant words.

Life inside the locked trunk is great.

Comment by Phil Ringnalda #
2003-04-27 10:41:30

Bugs aren’t the only reality: there’s also bugs in bugfixes. I was a bit surprised that it was too small in IE, since I thought I had a bug in my CSS that was giving IE a bigger font size than Gecko. Turns out it was only the front page, not individual entries. Anyway, the smallish font was mostly to keep <pre>formatted code from breaking the layout with long lines, so now that I’m putting it in (I hope) scrollable but fixed-width boxes, I can muck around a little more freely. We’ll see: I’m sick of seeing it look the way it does, but I’m nobody’s idea of a designer, so I keep hoping someone will take pity on me and drop a pretty new design on me out of the blue.

Although I do understand that this isn’t a total solution, and people shouldn’t make unreadable pages in the first place, it’s still worth having around: the nostyle widget from htmlhelp.com gives you a right-click option to disable stylesheets in IE, for that page only and only until you refresh. Very handy for turning a page of 10px pink text on a red background into something a human can read.

Comment by Dave Winer #
2003-04-27 04:10:40

To Roger, you are right. There’s no doubt that with the current scheme I’ve implemented, the server will read some files that have no RDF bits, and never could, because extensions don’t always tell you what’s at the other end.

So what’s the answer? Make Trackback always manual? Then what’s the point of the RDF bits? So the user can enter the URL of the page instead of the Trackback ping URL? (Seems like a pretty small thing.)

I have a hard time answering this question because as I’ve said, I don’t really ”get” Trackback, but I know some users want it.

Comment by Roger Benningfield #
2003-04-27 05:23:22

Dave: Let’s face it… I’ve got Trackback implemented for a different audience, a different app server, and a different business model. So just because ”shotgun” autodiscovery spooks me is no reason for it to alarm you. If you’re confident it’ll be fine, seriously, you know your stuff a lot better than I do.

As you saw when you tried pinging my blog the other day, I don’t have the RDF bits in there by default. (I added them for your use, but not across the board.) And I don’t support the old, GET-based ping, either. It’s PUT or nothing, as far as I’m concerned. So having established that I’m an odd duck in the world of Trackback, I’d suggest turning TB on by default, while making autodiscovery a toggle.

FWIW, I’m very fond of TB. Beyond broad, site-wide traffic analysis, I’ve never been all that interested in referrers… they just don’t tell me much without a click-through. Trackback, OTOH, tells me a lot with every ping. I just find it handy.

Comment by Mark Paschal #
2003-04-27 15:43:04

You might do an HTTP HEAD to get the MIME type, then fetch with GET if it’s an appropriate HTML type. (There is a conditional HTTP for MIME types, but IIRC Apache doesn’t support it so you might as well not. Locked trunk etc.)

I have a similar app (in Perl, with LWP and HTML::HeadParser) that blogs URLs spoken in a chat environment, and it does it this way.

Comment by Phillip Pearson #
2003-04-27 04:28:18

The Python Community Server project (open source RCS implementation) has handled incoming trackback pings via the comment system for some time now; I’ve written a little about the technical details on my blog.

Anyone (Dave?) working on implementing the trackbacks-as-comments thing for Radio – please read this! It would be very cool if the bits PyCS has added on to the xmlStorageSystem API could be supported by RCS…

Comment by Dave Winer #
2003-04-27 04:33:24

Philip, have I ever told you what a joy it is to work with you!

I’ve bookmarked your message. I’m going to do a little more work on this and then pass it over to Jake. Then I’ll ask him what he thinks about an addendum to the xmlStorageSystem spec.

And anticipating a complaint from Danny, yes I am the sole maintainer of the xmlStorageSystem spec. No one else wanted the job. Believe me. ;->

Comment by Dave Bauer #
2003-04-27 05:24:21

I am implementing trackback right now, so this is definitely an issue for me.

I just have an idea that may or may not work, but is an interesting option (for me anyway.) It makes it a little for difficult for the user, but its just an option. Present a list of URLs to scan for trackback to the user with checkboxes. It gives a little more control, without making the user type or paste the URLs to trackback in again.

I think I probably will implement this as an option. I am not sure if there is a way to read HTTP headers to find out file size, I think there is, but I also think it is not always set. So you might still end up reading a 5mb file.

Comment by Phil Ringnalda #
2003-04-27 10:02:58

That’s pretty much the way I would like to see MT do it: with a single thread app, where you’re going to sit there waiting after you save an entry, what makes the most sense to me is to do autodiscovery, return the found TB URLs, and ask which to ping. For something like Radio and Manila, where returning immediately when an entry is saved is a design goal, I’d do it as another form button: my choice for workflow would be type an entry, click ”preview”, click ”autodiscover”, select which to ping, click ”save”. I suspect I’m sweeping back the tide wanting people to manually decide to ping, but that’s what I’d like to see.

What to do to avoid big untrackbackable files depends on how much you are willing to do to avoid it, which probably depends on how bad a thing it will be if you don’t (desktop app used by dialup users == bad, multiuser server app == not good, single-user server app == eh). You can start with a HEAD, look for a Content-Type: text/* header, and be pretty well set, though even so it might be worthwhile looking for Content-Length:, and if your HTTP library lets you, setting a maximum size for what you’ll GET, just to avoid files misserved as text/plain with no length even though they are huge and not text.

Comment by Kevin Marks #
2003-04-27 10:20:21

Here’s a suggestion:
Add a per post trackback URL to RSS. Generate it in Radio, MT, etc.
When a Radio (or Manila or NetNewsWire or whatever) user creates a new weblog post from such an RSS feed, the tool automatically does the trackback ping.
This would make it usable by ordinary humans who don’t want to copy and paste URLs around to comment on other people’s blogs.

Comment by Phil Ringnalda #
2003-04-27 11:41:20

I’m in (and have been for months, actually). Seventh element in every item is <trackback:ping> – if something wants to start using it, I’d be thrilled. Radio seems like the most likely candidate: while posting from the aggregator, mark the link from the <link> as already discovered, with the TB URL from <trackback:ping>.

As to copy-n-paste, truth is I almost never do: if I’m going to ping a single entry, I start my post with a Movable Type bookmarklet while I’m at that entry in my browser (since I’m likely to need a link and some quoted text anyway), and then it autodiscovers the link for that one entry. I only copy-n-paste when I’m pinging more than one entry with my one entry, or when I’m adding a ping to an existing entry. Probably part of the muddle about autodiscovery is that we only have one word for that sort, ”triggered by evidence of intention to ping with an option not to send the ping autodiscovery”, which is a very good thing, and ”shotgun pinging of every possible thing that can be pinged autodiscovery”, which I think will just devalue pings for readers.

Comment by Gary Santoro #
2003-04-27 12:10:14

This sounds attractive. I hope to use it in Radio. Trackback is a great benefit to both the author and plain old website visitor. It facilitates community and provides exposure to weblogs. It lets authors AND visitors know what is out there in the blogging system.

Comment by Dave Winer #
2003-04-27 13:22:03

Very interesting idea about the POST button in Radio and Trackback. I like it.

Comment by Viswanath Gondi #
2003-04-27 15:56:00

A Link is a vote cast by a site for another site. If linking is automated, we might have future problems, like authentication, which Roger Benningfield pointed out. Also, Porn sites are using scriptings referers site to post their url, without puting a link on their site, to increase their google rating. Now scripting appears in the searchs for porn. I think we should try to get Google allow bloggers to ping them and get their database updated when a blog updates. If google updates faster than it is currently doing, we can just use the ”related to” query ……… More

Comment by Dave Winer #
2003-04-27 16:25:46

Viswanath, that’s amazing. I couldn’t figure out why they wanted to be in my referers, and you got it. I wonder if Google has figured it out.

Comment by Phil Ringnalda #
2003-04-27 17:08:06

You don’t actually read me, do you? Just skim, maybe? I’ve got to get someone to do a ”Shorter Phil Ringnalda” like that ”Shorter Den Beste” blog ;->

I don’t know how many times now I’ve said that if you display referrers where a search engine can get to them, you need to GET the referring page and scan for an actual link before you display it. And when you get a comment on your MT weblog saying something like ”Great article, keep up the good work” from someone named ”Zip Codes” with the URL zipcodedownload.com or zipcodeworld.com? That’s comment spam, looking for Google to index it, too.

As to Google figuring it out, how can they? You link to porn. I link to zip code database sellers until I delete the comment. If they’ve already decided that the lowlife is a shady character, they can decide that since you link to him, you’re a shady character, too, but as to deciding that they shouldn’t register your vote for porn, I can’t imagine how they could tell one link from another.

Comment by Dave Winer #
2003-04-27 20:22:06

If we had a nice blacklist database of porn-spammers. But I imagine there’s a reason that won’t work too?

Comment by Phil Ringnalda #
2003-04-27 20:40:49

(What, exactly, was I thinking when I tried to set myself up as an expert in the ways of spamming porn merchants? Not bright, Phil.)

I’ve been told several times that I’m too harsh on the need for immediate action on spam links, so it may be that it would work, especially if it’s RCS and PyCS and maybe things like whoever’s Javascript that is that I can’t remember right now, all cooperating. But you would need to have a way for anyone to say ”this referrer is spam” in their referrer list, which then requires that someone other than them look at it to see if it’s just someone they don’t like, to avoid the kinds of crap that email blacklists do, blocking a whole ISP when someone reports real opt-in email as spam. My feeling is that just a fairly static list wouldn’t work well, long-term, because most search-engine spammers run big linkfarms with a ton of URLs feeding into just a few sites, so if you block 50, they just have to switch to the other 150.

But it sure wouldn’t hurt to try, and if you point to yours, I’ll add it to mine (which is currently just search engines that I don’t want to show circular referrers for, but I’d be happy to not even have to GET spammer pages to see if they have a link to me), and share any I find.

Comment by Phil Ringnalda #
2003-04-27 21:18:07

Cor, even when I’m not skimming and thinking about reading, I still can’t read. You mean ”Google could read our blacklist and not index links to them”, don’t you? That’s much more optimistic about their openness than I tend to be. I was thinking more in terms of a blacklist that we use to avoid showing a link to them in the first place, to show them directly that they can’t use us, rather than getting Google to make their use of us without value.

Comment by Dave Winer #
2003-04-27 20:24:24

And yes Phil I do just skim you. Admit it — you just skim me too. As does everyone skim everyone else. That’s why so many people flame me if the term CSS appears in the same paragraph as a negative word. I should try it some time, concoct a paragraph with the word CSS and some word like sucks, or stinks, and add lock and trunk, and throw in a Zeldman and standard for good luck, but actually be praising all of the above. How much you want to bet a holy jihad starts anyway?

Comment by Phil Ringnalda #
2003-04-27 20:48:43

Actually, I don’t. There are a few people I do skim, because most of their topics don’t interest me but a few do, but for the most part the people I read, I really read. Obsessively, every word, thinking ”what made them put it that particular way” read. I do misinterpret you, maybe less often than some people, but it’s not because I’m not reading. That’s what I’m trying to get across when I say that I’m a reader, not a writer. Not just that I’m not a writer, but that I’m a reader. On my best days, I do understand that that’s unusual, and that most blog readers don’t have time to read a dense four paragraph post, but those days are rare, and usually I just go on and on, throwing more words at what I really want to say.

Comment by Dave Winer #
2003-04-28 05:09:22

Phil, that’s wonderful! I usually write in such a way that it’s worthwhile thinking about why it says it the way it does. I was beginning to think that no one was reading.

Comment by Danny Ayers #
2003-04-30 16:43:36

My guess is a lot of tech people both skim *and* read deeply. The habit of RTFM in emergencies. Scan for keywords then zoom in really close.

People who have their work edited are more likely to spot nitty things like inaccurate translation, miscapitalization and pleonasms.

Comment by Gary Santoro #
2003-04-28 00:45:50

I’m speaking from more of an end-user perspective, and if blogging continues to increase in popularity, people will want autodiscovery. That way, (obviously,) blogs will get discovered when somebody posts.

The weblog authors on both sides, as well as the casual surfer benefit from the information. The network effects that result from this will boost blogging connections in general. It seems to me that the benefits that legitimate use of autodiscovery present outweigh the abuses that may result. After all, we haven’t quit using Google, email, and weblog comments because some people are spamming.

I personally don’t know the technical details, but I see the potential of Trackback in ”blogospheric” terms. – lol

Comment by Georg Bauer #
2003-04-28 02:47:27

There is already a means to put Trackback URIs into RSS feeds – the trackback module for RSS 2.0. PyDS already uses (and generates) it, so if people read PyDS generated feeds, there trackback URI field will be filled in automatically and so the trackback will happen without intervention. Would be nice if more tools would generate RSS with usage of trackback:ping, so that trackback usage (at least the post-to-post part) get’s even easier (and you don’t need autodiscovery with this!).

Comment by Phil Ringnalda #
2003-04-28 07:46:27

Yep, that’s what I’m using. Though it’s actually the TrackBack module for RSS 1.0 and 2.0, using native syntax in each.

I should have suspected that PyDS was using it: from now on, whenever I start to say ”but nobody actually does foo” I’ll check Py(C|D)S first ;)

Am I understanding the UI/workflow right, that you post from the aggregator, a textarea/input field/whatever gets filled in with the URL from trackback:ping, and then unless the user deletes it before saving the post, you send a ping? Not too bad, from my cranky perspective. I’d prefer having the URL hardcoded in with a checkbox that defaults to unchecked, since the bulk of aggregator posts aren’t the sort of thing that I think deserves a ping, but then, I tend to design completely incomprehensible UIs that do what I want, not what users want.

Comment by Georg Bauer #
2003-04-28 09:15:59

Yes, I just carry over the trackback:ping stuff to the URI entry field in the weblog posting form. So if the user want’s to suppress the ping, she had to delete it.

I actually never thought about checkboxes for this, as there are only maybe 5 feeds I read that fill in the trackback:ping stuff in their RSS feed and so it’s not that often filled at all ;-)

I do have lot’s of checkboxes though for preconfigured TopicExchange topics and categories and all those generate some kind of ping, too.

You are lost in a twisty maze of checkboxes, all alike …

Trackback by Big Damn Heroes (Tech) #
2003-04-27 02:59:54

Trackbacks Are Not Comments


”I say, yet again, A TrackBack Is A Comment.”

I’ve never spoken directly to Phil, but I think very highly of what I see in his blog. He’s one of those people who are enjoyable to read, even when I disagree. And this is one of those are…

Trackback by Raw Blog #
2003-04-27 03:55:32

Trackbacks are bits of string

Reading more at Phil Rignalda’s re. trackback, and his latest quibble. One of the commentors makes an interseting point, trackbacks

Trackback by Andrew Grumet's Weblog #
2003-04-27 08:19:16

Trackback again

There is a good discussion of trackback going on over at Phil Ringnalda’s. Interesting points: ”The most visible sort is post-to-post, where one post expands on, comments on, or replies to another post, and so the reply post sends a…

2003-08-11 17:03:19

Quibbling about TrackBack

Name (required)
E-mail (required - never shown publicly)
Your Comment (smaller size | larger size)
You may use <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <code> <del datetime="" cite=""> <dd> <dl> <dt> <em> <i> <ins datetime="" cite=""> <kbd> <li> <ol> <p> <pre> <q cite=""> <samp> <strong> <sub> <sup> <ul> in your comment.