But Slava, isn't the whole point of a weblog to deliver a personal view on a subject? If I just wanted the news, I could read MacMinute and the New York Times. Instead, I want to know if people are actually excited to see journaling HFS and why people consider microsoft's testimonial shilling different than Apple's. The dry conventional media could never deliver this; Weblogs give us a chance to see the messy reactions to a story rather than just the facts.
Posted by Aaron Sittig at October 16, 2002 7:52 AMTrue, and I think I have not stated my point well enough. I am all for personal views, but many news sites/blogs just blatantly copy the story without adding anything from themselves...
Posted by slava at October 16, 2002 8:52 AMI'm with Slava here, but I think the reason it bugs me is that the aggregators don't aggregate by subject, but by site.
I've been thinking about creating an aggregator that would group posts by the URL they point to, or by keyword, rather than by who wrote the post. That way, you could see what every site you subscribed to thought about the Microsoft "switch" page, then about subject Y, then subject Z.
This is part of the attraction of Daypop, but a story has to be pretty widely 'blogged before it appears in their top 40, and it may get there because it's being discussed by sites I don't want to read (I can imagine a lot of people not caring about what the warblogging community, or the Mac geek community, or the insert-community-here community think is important). The reverse is also true: A story that's widely discussed in my circle of web feeds may never make Daypop.
Brent Simmons has already mentioned keyword folders as one possible improvement in NetNewsWire Pro, but I've never seen anyone discuss making the structure follow the content instead of being presented by site.
Posted by Frank at October 16, 2002 9:48 AMMaybe the answer is URL folders in your aggregator. If more than one weblog mentions a specific URL, then put both of the posts in a dynamically generated folder for that URL.
This sounds like the same problem that Trackback[1] is trying to solve, finding all weblog posts about a topic. But trackback assumes the story originated on a weblog with trackback functionality. This would work great for the unsanity.org story on Carbon vs Cocoa but wouldn't work for the eWeek rumors stories.
[1] http://movabletype.org/trackback/
Posted by Aaron Sittig at October 16, 2002 10:53 AMPerhaps a solution would be for google to use their algorithms that they're using to create http://news.google.com/ to identify blog articles on the same topics and rank them based on additional opinion comments made by the poster. I don't know.. I haven't really thought about this all the way, but it would seem possible for a 3rd party site to dynamically assign the id numbers as you mention.
Posted by Ben at October 16, 2002 3:02 PMIt occurs to me DRM might enable this. Who knew it could have a good side?
Posted by Pete Gontier at October 17, 2002 2:24 PM> It occurs to me DRM might enable this. Who knew it could have a good side?
That's because when it's evil, it's called DRM. When it's good, it's called cryptography.
Posted by Aaron Sittig at October 17, 2002 3:56 PMI must agree. I keep 3-4 Mac news sites in NNW and I am getting real tired of reading "Apple reports losses," "Apple refuses to go to Boston," "Apple to give OS X to Educators," etc. Repeated on every site.
It would be better if they added some sort of insight to the news instead of just reporting it, but they don't.
Posted by Etan at October 17, 2002 4:16 PMhmmm... mayyyybe if u concentrated less on fixing things that REALLLY mattered... and more on MIP for WS... u wouldn't go insane :)
Posted by ravedog at October 18, 2002 9:15 AMMASL (makeashorterlink.com) lets you type a three hundred character URL, any you like, into a form and get back a 25 character URL:
http://www.makeashorterlink.com/this/really/long/url/here
becomes
http://makeashorterlink.com/?R2C223732 (or http://masl.to/?R2C223732 :)
They assign a GUID to each URL pasted into their service. It'll work forever.
Movable Type (movabletype.org) developed a feature called "TrackBack" recently. It lets a blog reader say "I wrote about your article, at this URL" to a specific entry on a blog.
The technical mechanism? You send an HTTP POST to the specific comment's trackback URL, providing a URL to be added to the comment's collection of "pings".
Assign GUIDs to a collection of topics. Take a bunch of Movable Type blog articles, and decide which topics apply to which articles.
Then send a trackback ping to each article, providing a URL to the (guaranteed unique) topic.
It's kind of like choosing a Library of Congress call number for a book, then informing the book's author of the choice; once informed, they are certainly welcome to print that call number.
Everything2 (everything2.com) has a collection of topics, each with a GUID URL. Perfect for sending trackback pings, and a wonderful collection of 'newsbit' IDs.
The last step would be to tweak the Movable Type page's UI to allow you to select a topic for an article -- and since TrackBack is a public mechanism, anyone can do it.
I seem to remember that TrackBack was popular enough to where other blog software is working to interoperate. I'm out of date on that, though.
As an interesting note, I've been in the business of handing out GUID email addresses -- for instance, note the email address associated with this comment: unsanity-mt+43@crystalflame.net. I can guarantee that's entirely unique to this newsbit. The URL has a similar tag, for the same purpose.
(I've begun getting spam to the email address 'spam@crystalflame.net', as a side note. It got through someone's address filters. Hee.)
On a different note, I recommend you visit Daypop (www.daypop.com/top). Every time Unsanity releases software, URLs starting in "unsanity.org" appear scattered throughout the top 40. Daypop records a "chalk mark" for each blog post that links to a given URL. Problem is, this means the link "http://unsanity.org/" comes up every time -- not very unique, I guess.
Enter your GUID, which will show up in TrackBack pings. You could create a topic popularity index. When something flashes in the news -- Unsanity software release! W00t! -- the percentage of the day's articles reference the newsbit 'Unsanity' climbs.
It's a three-headed chimera of Daypop, TrackBack, and Everything2. Perhaps it'll meet your needs :)
Posted by Richard Soderberg at October 22, 2002 10:56 PMMaybe a utility could be built that would check Daypop and filter based on links to the same URL. Thus, you could say show me all of the blog entries that don't link to something on Daypop.
Just a thought as to how to accomplish this.
Posted by bbrown at October 24, 2002 1:25 PMSpeaking of repeats, this is at least the fifth time this same post has come up as new in my RSS reader - the comments change the RSS feed. Any chance you could provide a separate comment-free RSS feed for the kind of person that doesn't read the comments on slashdot?
Posted by Mike at October 24, 2002 5:57 PM"The Waypath Project's Related Weblog Navigation engine analyzes weblog entries to determine their core conceptual makeups, compares them with one another to find out how related they are, and presents you with its best guess as to what's related to your original input. This is done all automatically, using available technology."
Posted by Richard Soderberg at November 11, 2002 10:24 AM