Manila gets some lovin’

While I don’t regularly use Manila, I’ve always liked the idea of it, so it pleases me to see it getting lots of attention now that Dave’s using it at Harvard. I haven’t been keeping track of all the stuff that’s going on, but easy editing of navigation links is a really slick wiki-like editor, and up next is TrackBack and optional post summaries.

Dave’s take on post summaries and RSS is an interesting twist on my own: he says [t]he answer is to let the author make the decision, that while writing a post you should decide whether or not you will put the full text of the post in your RSS feed. My solution is to say that it’s up to the reader, to the extent that the author of their aggregator gives them the choice. I do a plain-text description of every post, even the ones where it’s damn near as long as the post, and put it in the RSS description element, and I put the full post in content:encoded (or xhtml:body, depending on the feed), and when I want to keep something to myself, I put it in Movable Type’s extended entry, which I don’t syndicate. That’s something I try to do as little as possible, mostly when it’s notes to myself I really shouldn’t be publishing, or something that will break my front page, but I do sometimes do it. There are certainly advantages to both ways of doing it, and I like Dave’s “you’ll syndicate full content unless you write a summary” far better than Movable Type’s default of “you’ll syndicate the first few words unless you make an effort to do something else.” But that wasn’t what I wanted to talk about.

I suspect that a lot of the movement in Manila is the result of putting a Manila developer in close proximity to some really smart users: most developers want to listen to their users, but when you have to interrogate them at length via email just to find out if their problem is your fault, their fault, or just that they don’t even understand what should be happening, listening can turn into a PITA and a huge time sink. If you have a bunch of users that you know are smart, whether or not they understand how you want your software to work, and if worst comes to worst you can just walk down to their office and see for yourself what they are doing, you should be set to do some cool stuff.

And interestingly enough, Evan now has his office in the midst of one of the biggest and best collections of really smart, really internet-savvy people around. And they are now Blogger users. The tiny little bit I know about that comes from referrers from domains that don’t resolve, and the odd spot of URL-hacking, but I do think it’s a bit odd that in all the talk about why Google would buy Blogger, I don’t remember anyone saying “maybe Google asked ‘how much for a site license?’, and when they heard the price, said ‘cor, that’s all? how much for the company?'” I don’t know, or really much care, what Blogger can do for Google, but I know that one thing Google can do for Blogger is provide a pool of smart users, capable of producing really good bug reports and feature requests, and that has to be good for Blogger and Blogger users.


Comment by Dave Winer #
2003-04-24 06:17:27

Right-o about smart users. Harvard is full of them. And the communication bandwidth is very high, nothing like face-to-face to foster understanding. Communicating over the Internet is fraught with difficulties. (Esp when people deliberately put noise on the net. No names mentioned.)

Sorry to hear that discovery for Trackback is the death of it. My implementation, which is working now, depends on it. As a result I can’t ping you because you don’t seem to support it. Why not? Remember I’m a TB newbie, so please be gentle. Thanks.

Comment by Phil Ringnalda #
2003-04-24 09:21:03

We need a few more terms for the various sorts of discovery: all I really object to is the shotgun form of pinging absolutely anything that can be pinged, without any user intervention to say ”yes, I said more about that” or ”no, I just linked to that, no reason for a reader of that to read my link”. Because the MT implementation of that calls it autodiscovery, I’m slamming that, but I don’t object to tools automatically discovering TrackBack URLs, only to automatically pinging everything a post links to.

I do support automatic discovery of the TrackBack URL through embedded RDF, though in my own curmudgeonly way. The default MT way is to include the RDF everywhere, archive pages and index pages, and to hide it inside HTML comments to keep the (X)HTML validator from objecting to it. If you scroll back to the start of my TrackBack category, there’s lots of thrashing around looking for ways to avoid that weirdness (doing it as RDF/XML, and then making it inaccessible to XML tools by commenting it out), and I’ve finally settled on my own odd solution: I embed mine directly in the XHTML, but my PHP pages sniff the user agent for a request, and if it’s an XHTML validator, the RDF doesn’t get included. So while view-source on most TrackBack enabled pages will show the RDF in a comment, on mine it’s just right there as more XML. It doesn’t really do anyone any good since everyone else comments it out, but it pleases me.

Then there’s the question of where it appears: I very strongly believe that a TrackBack is a comment that lives on someone else’s server (and that’s a useful way to decide whether or not to send a ping: is this something I would leave as a local comment if I didn’t want it in my blog as well as there?), and I don’t think that people should leave comments without first seeing what other comments have been left (same as not replying to email lists or forums without first reading other replies), so the only place I think people should be picking up TrackBack links is where my comments live, on the individual entry page. View source on this page (unless your browser sends a user agent that looks like the W3C validator), and you should see the TrackBack RDF. View source on my main page, where you can see an entry, but not its comments (and, if you are using an MT bookmarklet to start your entry, you’ll be picking up the URL for the main page, rather than the permalink URL for the entry you are going to ping), and you don’t get the TrackBack RDF.

So, if you pass your tool the URL for an individual entry of mine, and if it doesn’t depend on the RDF being commented out, then it ought to see it.

(Sorry that’s so long, I’m in a rush this morning and didn’t have time to make it shorter.)

Comment by Dave Winer #
2003-04-24 12:02:41

My crude little parser doesn’t depend on it being in a comment.

Comment by Phil Ringnalda #
2003-04-24 14:43:10

Hey, cool, a ping from Manila. Even autodiscovered, I’m delighted to see someone else in the game. Welcome!

And the digital magpie thing is lemonade from what I thought at first was a lemon from Sam Ruby. On second thought, I’m pretty sure all he meant was that when someone comes up with a new toy (TB, threaded comments in MT, BlogShares, IM notification of updates, you name it), I drag the pretty shiny thing back to my nest first, crow about what a pretty shiny thing I’ve got, and only then, only maybe, think about whether it’s something I can use. At least, that’s how I’ve chosen to take it ;)

Trackback by Dave Winer's Demo Site #
2003-04-24 13:09:12

Meet the digital magpie

Okay I have another Moveable Type site to test against.

Trackback by Dave Winer's Demo Site #
2003-04-25 04:52:55

Another test of trackback

I made a change in outbound trackback today, now it does the discovery-ping operation in a separate thread, so if the target is slow, the user doesn’t have to wait

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.