Flutterby™! : link management

Next unread comment / Catchup all unread comments User Account Info | Logout | XML/Pilot/etc versions | Long version (with comments) | Weblog archives | Site Map | | Browse Topics

link management

2002-02-04 19:24:30+00 by Dan Lyke 15 comments

It's bad enough when someone puts a web server at a domain name address rather than a machine address. It's worse when they redirect www.domainname.org to domainname.org with a "301" permanent redirection. But when they then stop doing that, and then start putting the previously "permanently redirected" URL in the weblogs.com list, well. Someone needs to give the admins of a popular nerd news site a severe bludgeoning with a clue-by-four. That updated time on the list to the right is becoming way[Wiki] more hassle than its worth.

[ related topics: Weblogs ]

comments in ascending chronological order (reverse):

#Comment made: 2002-02-21 05:35:02+00 by: Larry Burton

I'm not sure what the answer to this is, Dan. I fully appreciate how nice it is to have a list that lets you know where the new content is but I don't necessarily think weblogs.com is a good way to tell. Things just aren't standard enough for it to be reliable. The same problem of reliablilty exists with keeping a page of all the headlines on sites that provide RSS syndication, but in this case it takes longer to load a page of headlines than it does to just click through a list of links.

Oh, and what happened to Lake Effects?

#Comment made: 2002-02-21 05:35:02+00 by: Dave Winer

Which site is the problem? I don't understand what you're saying. Is it a critique of weblogs.com, or one of the sites that appears there?

#Comment made: 2002-02-04 21:57:55+00 by: jim winstead [edit history]

assuming that you're talking about http://slashdot.org/, it's not fair to pick on them for how they're listed on weblogs.com -- they're not pinging it. dave has a little radio tool he uses to help some of the sites that aren't pinging weblogs.com that he wanted to see updates for.

the /. listing at http://blo.gs/ sniffs the update time from their slashdot.xml file. and blo.gs tries to only list sites under either their 'www.'-prefixed or non-'www.' prefixed alternative, not both. (but there are old duplicates still in the database from before i implemented that filtering.)

#Comment made: 2002-02-21 05:35:02+00 by: Dan Lyke

Actually the real problem with /. is that they used to have a 301 redirect from http://www.slashdot.org/ to http://slashdot.org/. So my bookmark is to http://slashdot.org/, but they're listed on weblogs.com[Wiki] as http://www.slashdot.org/

If they hadn't dropped the 301, where the chain broke down would be more obvious. And really this is just me whining about the larger problem: A URL/URI is not a unique[Wiki] identifier, just a handle on one way to get there, so I need to either do manual bookmarks maintenance, or make smoe more likely-to-be-wrong heuristics for my data assimilator.

#Comment made: 2002-02-21 05:35:03+00 by: Larry Burton

Dave, the way Dan's doing his list, if the URL in weblogs.com doesn't match the URL in his bookmarks then it isn't going to show up as being updated. It sounds like a certain site isn't being consistant in the URL it pings weblogs.com with.

My problem is that some sites don't always ping weblogs.com when there is a change or ping it when there is a change in something other than blog entries. All of this seems to me to be outside of the control of weblogs.com.

#Comment made: 2002-02-21 05:35:03+00 by: Dan Lyke

So, Larry, should I be pinging weblogs.com[Wiki] when we get a new comment here? I go back and forth on this, I like the way Dave's change detection deals with it, but I'm not sure it's the right answer. Maybe a second entry, a "Flutterby - comments" one?

#Comment made: 2002-02-21 05:35:03+00 by: Larry Burton

On this site, the comments are as important to me as the entries so I would say yes for here. For some other sites I visit this would be a distraction. When you get right down to it, though, the less I think this site is updating the better off I might really be. ;-)

#Comment made: 2002-02-05 00:02:00+00 by: Dave Winer [edit history]

When they hand out canonical site IDs, a few months or years after all the weblogs use them, Weblogs.Com will use them.

#Comment made: 2002-02-21 05:35:04+00 by: dhartung

Larry -- I've been offline temporarily -- bad timing on the hosting renewal and maxed out credit cards. Soon! Glad to know someone misses me. :)

#Comment made: 2002-02-21 05:35:05+00 by: Mars Saxman

What exactly is "pinging weblogs.com", anyway? I keep hearing people talk about this but I don't know what it is. I went to weblogs.com once and it looked like a cross between blogger and livejournal...?


#Comment made: 2002-02-21 05:35:05+00 by: Dan Lyke

To get your web site in the http://www.weblosg.com/changes.xml file (and on the front page of weblogs.com), every time you update your site do an HTTP GET on the results of filling in http://newhome.weblogs.com/pingSiteForm

This tells weblogs.com[Wiki] to go check your site for changes.

#Comment made: 2002-02-21 05:35:05+00 by: Jerry Kindall

Mars, Weblogs.com is a list of Weblogs that have been recently updated. "Pinging Weblogs.com" just means telling Weblogs.com to look for changes at your site. If it finds any, your site appears on the front page of Weblogs.com for a while, until other updating sites push it off. Several other "blog tracker" sites also use the Weblogs.com list, so participating in the system can drive significant traffic to your site.

It used to be that Weblogs.com checked every site it knew about for changes every hour but that got to be impossibly slow, so Winer & Co. changed the site so you have to tell it whenever you update. There are a couple ways to do that; your content management system can do it via an HTTP or XML-RPC request (Blogger Pro and Movable Type do this), or you can manually do a ping at the site and just bookmark the result page for future use. The AppleScript I use for updating JerryKindall.com uses the URL Access scripting add-on to do it.

#Comment made: 2002-02-21 05:35:06+00 by: DaveP

If you're hosted on a Unix server and have static html files, you can use a script like http://davespicks.com/ping_weblogs.html via cron to periodically check the changing html file and automagically talk to weblogs.com, too.

#Comment made: 2002-02-21 05:35:09+00 by: starbreeze

Not that I'm defending Slashdot, because I get fed up with the users there so easily, but what's wrong with using a domain address instead of a machine address? The addie for my site is simply http://starjewel.org, altho you can use www as well. Slashdot actually had to create the www address after too many people complained about not being able to access it.

#Comment made: 2002-02-21 05:35:09+00 by: Dan Lyke

Actually, that bit's a piece of chauvinism for us old-school nerds who think a URL isn't there for human consumption, and therefore overloading the meaning of "domain name" to include "web address" is a bad thing.

It's just me being old and crochety.