If you think about it, what just took place in Afghanistan significantly informed what took place in Iraq. And what has taken place in Iraq is, in a sense, a -- it's a war, but it also is a -- an opportunity to learn about what we're doing well and what we're not doing well and what we can do better. It is like a giant laboratory where you can look at it as it's happening and as it evolves and finishes, and immediately take that knowledge and that information with respect to strategies and tactics and capabilities and weapons systems and munitions and logistics and all -- every aspect of this, and then try to back the lessons, the good lessons, into this system. I'm sure, I am positive that out of this we're going to end up finding ways that we can reduce friendly fire casualties. I am positive we can. There has to be a way to reduce that. And we know it's existed in every conflict, but without question, when this is concluded and we have the lessons learned, I'll bet anyone a dollar to a dime that we end up with ways that we can reduce those.See also related news.
When the Daypop Top 40 first came out, I noticed a fair ammount of redundancy. A single story (story meaning event, idea, or other meme) might occupy multiple spots on the list, such as Reuters articles on different news sites, and even the inclusion or exclusion of “www.” Not all stories were centered on a URL, but were rather spread out over a number of articles, blog posts, etc. In that case, some parts of the story would by themselves be too small to register on the list. I suggested to Dan Chan that he use topic clustering, to group the list into topics based on keywords. That way, all parts of one story could be seen at once, and conversely, a single story could easily be ignored if you weren’t interested in it. He thought the idea had merit, but nothing ever came of it. I later coded a prototype keyword-based topic clusterer, which I offered to the now defunct BlogMints (which could be described as an aggregator for blog comments), and nothing came of that either.
More recently, Daypop now has Top Word Bursts. This does group stories by keyword, however the stories aren’t garnered from the Top 40 links’ titles, but rather from the text written on blogs. Early on I suggested merging related keyword clusters, which Dan quickly implemented. It isn’t perfect, but few things are.
I also noticed that (at the time) there were quite a few listings on the Daypop Top 40 related to Macs, a topic I was uninterested in. I suggested to Dan that he create a separate top page for Apple (and ones on any other topics that were popular enough to warrant it), which could probably be easily done with keyword filtering. Again, he liked the idea, but it never happened. Much later, we now have Memeufacture, which includes Apple among its several topics. (Memeufacture and the Daypop Top 40 are among quite a few tools that do essentially the same thing; see my What’s Happening page for a comprehensive listing).
The Internet Topic Exchange is an idea I immediately latched onto. Anyone can create a channel (i.e. topic), and anyone can post thier blog entries into any channel. Many blogs include a system for categorizing entries, but this allows multiple blogs to work within the same topic framework. An individual can choose to read a Topic Exchange channel as if it were just another blog, with a focus that the individual wants, despite the fact that all the posts originate elsewhere. This blog has, for instance, included environmental stories, flash games, my ideas, and other topics. If someone is just interested in one of these, they could be reading a Topic Exchange channel that I ping, which would also include what other blogs say on the topic.
RSS aggregators (good listing here and here) are indispensable tools for reading the blogs you’re interested in. Add all the blogs you read, and then read them right from the aggregator, which also finds when the blogs update for you.
There are a fair number of public (online) topical aggregators around. For example, LISFeeds, on librarian topics and Weblogs At Harvard Law. The former (similar to a typical aggregator), uses one frame listing the blogs, and another frame to display the posts from the selected blog. The latter, on the other hand, lists individual posts on the same page, organized not by the source (the blog), but by the time of the post. This is better, in my opinion, as the blogs have already been selected for relevance, so no need to separate the posts in one from the posts in another.
Blog and RSS search engines are useful, but they search the whole blogosphere. Some newsreaders contain search engines, so that you can search just within the blogs that you read. Micah Alpern’s Trusted Blog Search Tool uses the Google API to do the same thing. It uses either an OPML file or Blogrolling.com blogroll (these list the blogs you read in your aggregator or the blogs you link to from your own blog, respectively), neither of which I happen to have. I’ll have to look into using OPML sometime.
Review| Tool | Includes | Examples |
|---|---|---|
| Blog/RSS Search Engine | whole blogosphere | Daypop, Feedster, others |
| Top Recent Links Indexer | whole blogosphere | Blogdex, Daypop Top 40, others |
| Recent Keywords Indexer | whole blogosphere | Daypop Top Word Bursts |
| Top Topical Recent Links Indexer | blogs chosen by creator | Memeufacture |
| Closed Metablog | blogs or individuals in a group | Metafilter, The Start of Fee |
| Open Metablog | posts by opt-in blogs | Internet Topic Exchange, Blogpopuli, blaxm! |
| RSS Aggregator | blogs you choose | see here and here |
| Public Topical Aggregator | blogs chosen by creator | LISFeeds, Weblogs At Harvard Law |
| Personal Blog/RSS Search Engine | blogs you choose | within aggregators, Trusted Blog Search Tool |
For quite some time, I have wanted my own Blogdex (top recent links indexer). Virtually all RSS aggregators now “cluster” posts by source; that is, you can view headlines from one blog at a time. As I mentioned earlier, I would rather they be sorted by date, regardless of the source. Even better, I want clustering by source, link, and keyword, all from within my aggregator. By clustering by link, I mean using the blogs in my aggregator to create a top recent links indexer, instead of the whole blogosphere as Daypop, Blogdex, and others do. Similar to Technorati, each link should be followed by the posts which include it. T