Jon Udell: Now that we’ve shared our OPML, will SYO share it back so we can create and contribute our own data mashups?
Apparently, on the same day I reported that there were over three hundred people subscribed to TechMeme’s HTML page, every single last one of them, and a few more, unsubscribed from it, and subscribed to the XML feed.
While the intentions behind this particular change are innocent, the fact remains that SYO in its current state is essentially un-auditable.
While the intentions behind this particular change are innocent, the fact remains that SYO in its current state is essentially un-auditable
Doesn’t this apply to pretty much every online service? Or do you have the audit records for the Technorati Top 100, Feedster 500 or TechMeme’s current hot stories of the day?
Your comment seems pretty inflammatory given that this is the status quo in the online world today.
Your comment seems pretty inflammatory given that this is the status quo in the online world today.
Perhaps. But those other services have yet to surprise me. And after seeing the suggestion that someone at SYO has been manually changing some fields caused me to recall Jon’s prior unanswered question about the service.
Meanwhile, Michael Arrington writes (emphasis added):
SYO is completely objective and shows exactly what content people are actually reading right now.
Which receives a prominent link from Dave Winer’s ScriptingNews. Michael Arrington has some undisclosed interests in the matter, in that he has served as Dave Winer’s attorney.
Again, I can only presume that the only discrepancy I have found to date is completely innocent, but at a minimum, a disclaimer should have been added to Mike’s piece and Dave’s link.
Actually, I now think my suggestion from yesterday was off.
Here’s my guess as to how things work: They rank the most popular feeds by finding the most popular blog HOME url (not the feed url), and then assign the associated feed by picking one that an ARBITRARY user supplied. Ranking by url of blog is important because people subscribe to the same blog via multiple feeds. In the case of my site, a SINGLE user put in the bad feed URL, which led to the error. Of course the “right way” to address this is to add a step where they find the most POPULAR feed associated with the blog home url, something maybe on their todo list.
I know, this is all speculation (I haven’t checked with SYO), and somewhat orthogonal to your point, but I thought I’d offer a less sinister interpretation since it seems reasonably possible.
Ranking by url of blog is important because people subscribe to the same blog via multiple feeds.
If you scan down the right of this page, you will see that there are a number of popular blogs that are listed multiple times. I noted this previously.
When I was professionally engaged in this project, the biggest obstacle I thought Share Your OPML would face is this: After the early adopters get bored, what’s the value for more mainstream users?
Asking people to share their feed subscription data is like asking them to Share Your Books or Share Your Grocery Purchases. People who want to volunteer their personal consumption and can produce the data in an XML format seem like a rare breed, and I don’t see what users get back that justifies their time and effort.
To my thinking, that’s why the first incarnation of Share Your OPML failed to catch fire in 2004, though perhaps it was because Mike Arrington wasn’t around to tout it to a huge readership while neglecting to mention his professional relationship with Dave Winer.
Old injuries flare up around a new effort: winer-arrington connection, roger’s “sour grapes” review...
What web directories are "auditable"? The debate over ranking blogs are beyond the science of statistics. I judge directories by the benefits I get from them. Share Your OPML has disclosed a few blogs that I didn’t notice before. It works for me because I’m a geek and early adopter... The aggregator output page will help prove the concpet for another wave of adopters.
Critically, I do think that Udell hits the mark: The site doesn’t share the OPML. It’s got a one-way design like most commericially-oriented sites. It asks for donations and post the results in a static form... Maybe it’s just a lack of time getting the two-way feature out (serve OPML as an output). I’m hoping they will serve OPML files so I can import Scott Adam’s feeds and view the web from his preferences and then dump that list and import Arrington’s.
When I wanted an OPML file to play with an aggregator a few weeks back... I imported Sam Ruby’s blogroll from bloglines. Dozne’s of great geek blogs flowed before my eyes. I got aggregation in a way that I couldn’t by hand-crafting an OPML list... I didn’t see the point over clicking on my favorties until I had a quality list of a few hundred to scan... Then I added Scoble’s and overload (with a Microsoft slant) destroyed the benefits. It was then that I discovered that my aggregator tool couldn’t dump OPML and let be re-start.
So, cut Winer some slack guys... If the idea is good. Someone will fix the implementation and steal the audience. Patterns.
Rogers, it’s just very difficult to see you tell us that Winer’s shit smelt like flowers for several years and then now hear you say his flowers smell like shit. As much as I don’t like some things Winer does, Share Your OPML is awesome. I’m sure it’s not perfect and I’m sure it’s got bugs.
Darryl, you could be correct.
Sam, please implement retinal scan based captchas.
Randy, and analogy comes to mind:
“Hell hath no fury like a woman scorned”
I spent 4-6 months running an incarnation of the Share Your OPML web application and working with the OPML format, most of which preceded the legal debacle. I’m not going to avoid giving my take on the subject because people have a convenient reason to dismiss it.
The question I asked wasn’t even particularly critical: What’s the value for mainstream RSS users in submitting their usage data to Share Your OPML? Additionally: What’s the value for competitors like Bloglines in helping users hand this over?
Rogers:
“What’s the value for mainstream RSS users in submitting their usage data to Share Your OPML?”
Unkown. I can’t see any reason for average joe blow to go through the effort.
Darryl,
There is much value to the average joe. The biggest, I find is feed discovery. Upload your OPML, then go here...
You’ll find other people with similar interest. Then dive into their subscription lists and you’ll likely find some feeds that you didn’t know about, weren’t subscribed to and become a daily must read.
James,
You funny!
Then dive into their subscription lists
This would be made easier if SYO would simply “share it back” (as Jon Udell so aptly put it), as you could then simply import their subscription lists, or use them directly as reading lists if your aggregator supports this function. Mine does.
Wow... Randy and Rogers throwin' down in Ruby’s mud pit!
From a reader’s perspective... you guys are on the same team but there’s no management to re-focus you on how we keep score.
“Share Your OPML” will grow and establish a model for a service that
someone will re-create and re-target towards a mass larger audience.
Import/Exporting and aggregating blogrolls is a great idea.
Bloglines might be able to co-opt the effort in a few weeks but it
won’t have Dave Winer and Mike Arrington pimping for success.
Jeneane Sessum and other powerful blog voices will get it when the data they want is brought forward from the RSS geek rolls.
I imported Sam Ruby’s blogroll of one week and then re-booted today into Jon Udell’s opml list. What I want is for someone to export the OPML at Share Your OPML. Scrape the HTML page at:
and create an OPML file... Bela? Kosso? Sam? Anyone?
Then I can be Scott Adams for one week. Capice?
Sam, yes that would be nice.
McD, I like to get dirty :-)
What I want is for someone to export the OPML at Share Your OPML.
Technically, it is easy. Just add BeautifulSoup (and possibly mark as executable and/or rename to .cgi, depending on your server)
Legal? Welcomed by the community? Not so clear.
My feeling is that if SYO wanted you to be able to retrieve this data in OPML™ format, they would have provided it. Meanwhile, there is that legal statement:
Copyright © 2004-2006 Scripting News, Inc. All rights reserved.
Furthermore, the layout of that page could change at any time.
FYI: Bloglines has long had OPML import, export, and recommendations. Not that I would recommend Bloglines to anyone at this point.
“What I want is for someone to export the OPML at Share Your OPML. Scrape the HTML page ..”
The original Share Your OPML displayed user subscriptions as a web page and in OPML format. It also offered an OPML version of the “100 most-subscribed feeds” page.
Sam,
Thanks for the scaper! It worked like a charm... I may re-code it in Ruby as exercise and since I see that BeautifulSoup has been ported to Ruby as well.
I scraped the Scott Adams page... imported into Dave Winer’s newRiver aggregator and I’m happy.
“Share Your OPML” indeed.
Rogers: I feel your pain. It’s a sad story... All your work trashed. Like an ugly tech divorce with children.
McD