Back to previous post: Ice pop

Go to Making Light's front page.

Forward to next post: Billy the Shake goes to the Demo Derby

Subscribe (via RSS) to this post's comment thread. (What does this mean? Here's a quick introduction.)

June 23, 2005

Baby, pull yourself together
Posted by Teresa at 07:00 AM *

Over the last week or two, my previous problems with Technorati have been supplanted by a new one: it gives me the same set of outdated incoming links over and over again, reshuffling their order, sometimes swapping one out only to swap it back in again an hour later.

I wrote my post about the giant popsicle meltdown yesterday morning. By early evening I learned that it was picking up considerable linkage. How? Not from Technorati; it hadn’t registered any of that activity. I found out because Patrick told me. So I went and checked the Making Light tracking statistics, and sure enough, more than half the incoming referrals were to the popsicle post. That’s heavy traffic.

I went and checked Technorati again. Once again, it gave me the same old reshuffled links it’d been showing me all week. The links it had missed, at least the ones I know about, included BoingBoing, Metafilter, Crooked Timber, Majikthise, and Sisyphus Shrugged. Julia’s tardily showed up at Technorati this morning. Going by what I’ve observed thus far, it may never register the existence of the others at all.

I’ve been in denial. No more. When Technorati can’t tell you that you’ve been simultaneously BoingBoinged and Metafiltered—an event that can shut down a low-bandwidth site—it is formally useless.

Bummer. I am definitely in the market for a replacement.

Comments on Baby, pull yourself together:
#1 ::: adamsj ::: (view all by) ::: June 23, 2005, 08:56 AM:

Tried the beta page yet? (Odd topic for the alpha post.)

#2 ::: Patrick Nielsen Hayden ::: (view all by) ::: June 23, 2005, 09:04 AM:

Yup, Teresa said grimly.

(My experience with the beta site is that large parts of it, like "sort incoming links by 'most authority,'" simply don't work at all; you wait forever for no result.) (We both realize this is arguably a variety of complaining about the free ice cream, and yet.)

#3 ::: adamsj ::: (view all by) ::: June 23, 2005, 09:05 AM:

"And how would I do that?" I hear you ask. Well, I have the answer.

I don't know whether this page finds something different behind it from what the proper page does, but there it is.

#4 ::: Lisa Spangenberg ::: (view all by) ::: June 23, 2005, 10:22 AM:

Teresa

Both on my own tiny site, and the much-uses sites on campus, I find using the web log (the real one, not a blog) produced by the web server itself in tandem with a very primitive javascript on each page gives me both the instant snap shot data (who's coming to the site right now, from where) and the more in depth stats I can sift and analyze at my leisure.

There are any number of "statistics" java scripts; I've used http://www.sitemeter.com/ in the past, and am currently using the free version of http://www.statcounter.com/. It's only good in the free version for the last thousand visitors, but that ought to let you know within an hours time if you've been linked somewhere new.

And yes, it's crude, and yes, of course I know that there are all sorts of ways to avoid being "counted"--that's fine; people have a right to do what they want with their own data; this is just a way of dipping one's toe in the river of stats to check current conditions.

#5 ::: protected static ::: (view all by) ::: June 23, 2005, 10:22 AM:

I've noticed that Technorati has a hard time distinguishing between 'newly-published' and 'recently-refreshed/recently-updated'... Using Blogger, for instance, whenever you update something it republishes the entire page and Technorati will see it as being 'new'. It seems like Technorati doesn't distinguish between 'created date' and 'last modified date', and my guess is that the various blogging apps out there don't make that distinction any easier.

#6 ::: PZ Myers ::: (view all by) ::: June 23, 2005, 10:52 AM:

The beta technorati doesn't help, because now the only technorati is the beta technorati.

I noticed when they switched over that it stopped my old plugin that used the technorati api from ever showing me anything new -- the same sites now seem to be permanently listed on my page. And my technorati rss is doing the same thing Teresa describes, with the same last few sites shuffling up and down.

Of course, maybe its because every weblogger suddenly discovered who I really am, and now I'm being shunned as a pariah. I don't know, because Technorati won't help me find out where I was outed.

#7 ::: sara ::: (view all by) ::: June 23, 2005, 11:00 AM:

I've been wondering for a long time why people haven't been noting the problems with technorati. In my case it's missing about half of what's coming my way. When I realized this I did a google for "technorati problems" -- and got next to nothing. This was maybe a month ago.

So I'm glad you're raising the visibility on this.

#8 ::: shana ::: (view all by) ::: June 23, 2005, 11:06 AM:

nice tom swifty, Patrick.

#9 ::: Teresa Nielsen Hayden ::: (view all by) ::: June 23, 2005, 12:06 PM:

Sara, I've been noticing Technorati's problems for a long time. It just embarrasses me to talk about them. I keep thinking I must sound like an egomaniac who's complaining about egoboo deprivation.

#10 ::: julia ::: (view all by) ::: June 23, 2005, 12:11 PM:

I ping them, too, so that's kinda weird.

I know that links to me frequently disappear and reappear later and the numbers don't update at all.

#11 ::: Patrick Connors ::: (view all by) ::: June 23, 2005, 12:23 PM:

I just dropped Wired from my RSS feed because all week they've been giving me the same links over and over again. I wonder if this is connected?

#12 ::: Christopher ::: (view all by) ::: June 23, 2005, 12:26 PM:

I've been having better luck with Blogpulse, http://www.blogpulse.com/. Though it's still not as useful as Technorati used to be, it's a damn sight better than Technorati is now.

#13 ::: Mac ::: (view all by) ::: June 23, 2005, 01:17 PM:

Gads, I'm never going to learn all this stuff. I was inordinately proud of myself just for learning enough html to code linked texts into comments.

#14 ::: Jules ::: (view all by) ::: June 23, 2005, 04:10 PM:

Have you considered a tool that just analyses your access logs and flags up any new referer URLs it spots? Something like this: http://awsd.com/scripts/logtools/index.shtml or perhaps one of the other scripts on the page I found it in: http://simplythebest.net/scripts/perl_scripts/log_scripts.html

#15 ::: David Sifry ::: (view all by) ::: June 23, 2005, 04:22 PM:

Hmmm, this sounds like a serious problem, thanks for alerting us to it. I've set the engineering team to work investigating just what went wrong and how we can fix it ASAP!

Sorry about the problems, and thanks for pointing it out. I hope you keep coming back to Technorati as we work to fix this problem. And please don't hesitate to let us know about these issues if and when they occur in the future, we really need and love to get your feedback, both positive (of course) but especially negative constructure feedback. We are working really hard to be of service to you, and data quality and performance is paramount in this.

#16 ::: Kathryn Cramer ::: (view all by) ::: June 23, 2005, 06:43 PM:

I keep the faith on Technorati. I have confidence in David Sifry.

The database meltdown I have been quietly chuckling over is the Truth Laid Bare, which for weeks listed my blog as one of the top 100, an obvious error, even more obvious if you had a look at the supporting info.

The highest I rose was to #84. I'm currently at #107. I think this is terribly funny.

#17 ::: julia ::: (view all by) ::: June 23, 2005, 08:33 PM:

Oh, TTLB is a riot. Technorati (which I know is missing links) shows me with a cumulative 400 or so from both addresses. TTLB shows me with something in the mid seventies.

Of course, if you only track blogs that link to Glenn Reynolds, you aren't going to see a lot of the blogs that link to me.

#18 ::: Stephanie ::: (view all by) ::: June 23, 2005, 08:55 PM:

I just went through, for the second time in a year, an evaluation of of the professional-grade log analyzers. Since this might be helpful to someone...

Free ones: Webalizer, AWStats, Analog (with Report Magic)

Not free: Urchin, NetTracker, WebTrends, and ClickTracks.

My favorite is Urchin 5, and I'm profoundly grateful that my host offers it as part of my plan. It isn't cheap -- around $900. Then again, the version of WebTrends that offers the same features costs somewhere around $10K.

Webalizer and Analog are the most common free ones, but I like AWStats better than either. The free ones tend to be inflexible, showing you only the most popular 10 pages on your site. If you want to see numbers on a new page that's getting a lot of traffic right now but hasn't yet surpassed your top 10... no go.

And then there's Refer, which doesn't produce pretty graphs of any kind, but is really great for tracking inbound links as they appear. ShortStat is also really cool. (This seems to be the most recent version.) I gather it's soon to be replaced by Mint. These two don't analyze logs, but rely on bits of PHP embedded in the page instead. Both are free.

Either Refer or ShortStat might be a good solution for you, Teresa, but in order to use them without breaking your permalinks, you'd probably need to do some server magic to parse .html pages for PHP code.

#19 ::: Stephanie ::: (view all by) ::: June 23, 2005, 09:05 PM:

Either Refer or ShortStat (example) might be a good solution for you, Teresa, but in order to use them without breaking your permalinks, you'd probably need to do some server magic to parse .html pages for PHP code.

(I'd written a longer reply, but Movable Type didn't like it, and I moved it to my site instead.)

#20 ::: Teresa Nielsen Hayden ::: (view all by) ::: June 23, 2005, 09:25 PM:

David Sifry gives off such penetrating Good Guy vibes that I feel guilty for even thinking of complaining.

#21 ::: Stephanie ::: (view all by) ::: June 23, 2005, 10:12 PM:

And my longer one posted anyway. Sigh.

*kicks Movable Type*

#22 ::: Jules ::: (view all by) ::: June 24, 2005, 04:33 AM:

I'm pretty sure there's got to be an easier way than embedding PHP in the code.

But, it seems you're running Apache and have PHP installed, so that is a potential option. The particular magic you want is to find the line in your httpd.conf file that reads "AddType application/x-httpd-php .php" (possibly followed by other extensions) and add " .html" to the end of the line. This should cause the server to process all your HTML files through PHP. Then you can include any tracking code for such things directly.

#23 ::: Stephanie ::: (view all by) ::: June 24, 2005, 09:12 AM:

Or, if you don't have access to httpd.conf, add "AddType application/x-httpd-php .html" to .htaccess.

#24 ::: Teresa Nielsen Hayden ::: (view all by) ::: June 27, 2005, 01:28 PM:

Note: On Beta Technorati's current list of top ten searches, clicking on item #10, "Odeo," yielded:

There are no posts that contain that text yet. Please try again later or add it to your watchlist to track future conversation.
Searching Google for weblog odeo got me 77,900 hits, including an entire weblog devoted to the subject.

Guess that'd be why they're calling this beta.

#25 ::: jami ::: (view all by) ::: June 27, 2005, 04:59 PM:

just when i was ready to start in on the "technorati's gonna kill google" conversation, technorati's killed technorati. my site never comes up for karl rove, no matter how much i talk about karl rove. do you think it's greed, like with google, where if you want your site indexed quickly, you better quick with the green stuff?

sure, we all gotta make a buck, but technorati's jumping the gun. if it's gonna be just as ineffective as google, some of us will just go back to google.

#26 ::: insider ::: (view all by) ::: July 22, 2005, 12:40 AM:

We've been stuck at the same # of links at Technorati for the past several weeks at a time that I know a great # of people have linked to our ebonics story. Don't even get me going on TTLB as they seem to use a random number generator to calculate our links (when the servers not busy that is).

Choose:
Smaller type (our default)
Larger type
Even larger type, with serifs

Dire legal notice
Making Light copyright 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010, 2011, 2012, 2013, 2014, 2015, 2016, 2017, 2018, 2019, 2020 by Patrick & Teresa Nielsen Hayden. All rights reserved.