I need to buy a tape recorder so I can catch all my weird ideas in the moment instead of relying on my poor memory to recall what Past Roy was thinking.

I went to the Chapel Hill meetup of bloggers at Cafe Driad tonight. I'm not much of a conversationalist, so I mostly shut up and just listened to what other people had to say. Because the whole concept of easy self-web publishing is so new, I think a lot of the bigger problems haven't been dealt with effectively on a lot of self-publishing platforms. What will help platforms in the future is how they deal with inevitable information overload. As each weblog/website [I use these terms interchangeably because I honestly think they boil down to the same thing] grows older, the information on the site grows and grows. I have over 1000 posts on this site alone; can anyone effectively locate a post about a specific topic? [Answer is no]

The next big step for all developers is how to manage the information overload in the blogosphere (I wrote about this earlier briefly). How do we quickly locate users based on similar interests, geographic location, and hot topics [a tool for tracking specific current events]? Technorati allows you to search all the blogs they've indexed for specific keywords. The problem with Technorati [and it's really not their fault] is that they're taking a "one-all" approach to this aggregation. Every blog in the world is being searched for a specific keyword. If becomes impossible to find information about local news or local events. Furthermore, there's no way to identify quickly if a user is from your geographical area. Blogging also needs to serve as a tool for connecting people in real life [one of the many lessons I'm learning].

An idea

What would be a cool tool to build [which probably means I'll end up making a rough copy then giving up] would be to create a DMOZ-type directory of bloggers [a blogging yellow-pages, perhaps] that would let users quickly identify bloggers in their area. Weblog tools would 'ping' the centralized site whenever a user updated their site.

The way I envision this would work would be to let bloggers use the "ping" feature found in many weblogging tools [weblogsUpdate over XML-RPC] whenever they update their site. The centralized site [I'll call it BYP for now for Blogger Yellow Pages] would accept the ping and throw it into a queue. If the site's URL was a new site, then BYP would examine the page for four critical pieces of information:

• The RSS feed
• The FOAF file
• Geo header data in the HTML file
• general metadata [keywords, title]

Unlike most aggregation tools [and I'm not aware of too many, so correct me if I'm wrong], this wouldn't serve just as a "here is a list of the most recent posts by this human-edited list," this whole system would be automated.

Furthermore, one of the main reasons I don't use aggregator tools is because there is absolutely no context for the user. Who is this user? Do I really need to visit their site to read about it? What compounds this problem is that a lot of users don't have visible "About Me" pages [including this one!] so the context of the weblog is completely lost.

Real quickly, for those who don't know what FOAF is, FOAF is a standardized XML format for telling people who you are. There was some hoopla about it a while ago, but the project seems to have simmered down. I find FOAF useful because it's a standardized format for exactly this purpose - identity context.

The FOAF file would help eliminate this problem. BYP would capture the FOAF data and create a Friendster-like profile of the user within the system itself. So not only BYP would you see a listing of "the most recent posts," but you would have a quick view of the community of characters themselves without having to visit each blog individually.

The geodata would help identify users with location; if they didn't input a location, they could manually input their data into their BYP profile. Since this is a geographically-driven tool instead of a interest-driven tool, the primary concern would be with grouping users based on location. I'd have to find an API or tool that would convert longitude/lattitude values to specific metro locations, but that's a problem for another day.

The meta-data would help separate users within each location to interest groups. Do you have a blog about politics in a local area? They define one of your meta keyword values as 'politics,' and BYP would aggregate that into a category called 'Politics' within the 'Chapel Hill' area.

BYP would *not* be a post aggregator [Bloglines kicks too much butt to do that]. At the very most, I imagine it could take many different feeds and simply identify what *types* of information is offered by each user [and maybe show the last 10 posts or so per user]. However, it would offer an OPML file for any specific category within BYP that you could import into your aggregator of choice.

A separate tool that would be cool

There was an interesting discussion on the value of 'Carnivals' tonight. Carnivals are in essence a post that simply sums up the "best of" a specific post that is human edited. Someone sends out a request for the best posts regarding a specific topic; people then e-mail this person with all the posts, and this one poor soul sifts through the posts and creates a long running post with short excerpts.

Why not help build tools to simplify this tool? It gets to the essence of information overload: How do you identify "valuable" posts within the blogosphere that are worth reading? Eventually someone is going to have to read it and make a determination himself. I don't think there is an automated way of doing this - for comments and such, the /. system of moderation works fine, but with the very nature of the blogosphere being spread out on so many systems, no one could agree to a single classification for "rating posts." Someone is going to have to read them.

The tool would allow someone to create a "Carnival of ______" thread on a centralized site. People would then trackback to this post with their individual posts. The site [which I shall now refer to as CA for Carnival Aggregator] would grab the RSS feeds to the site and store the specific entry within one single location [like a blog aggregator!]. The person who created the Carnival thread would set a time limit on when people can submit trackbacks to the Carnival thread.

The thread creator would then go to the site and read all the posts on one page. A checkbox and a text form would let them quickly aggregate annotate good posts, and then the site would generate all the HTML for the whole Carnival thread.

The creator could then post the generated HTML of all the links and annotations on their site, and we'd have a Carnival!

But the cooler thing here is that site now has *archived* the carnival. With the temporal nature of many blogs, a lot of useful data gets lost as server and sites go down. CA would have all this data with all necessary metadata stored on its servers so someone doing research in the future could reference all the threads.

This system could be expanded to any "hot topics." For example, imagine if someone created a Carnival for the Tsunami crisis a while ago? People who actually lived in the area coudl trackback to the main site, and bloggers could follow the progress on a day-to-day basis [assuming someone created a Carnival a day]. This would also be really cool for stuff like elections where you could actually track the opinions of people as an election goes on [e.g. where was the tipping point for Dean's popularity?]

Imagine integrating this with the BYP above. I'll leave the implications for you to consider.

Another tool that needs to be built

One of the first big features I built into Tabulas was the creation of a crossposting tool to Xanga and Livejournal. One of the issues that I think a lot of bloggers are dealing with is the issue of dealing with multiple blogs and how to crosspost to multiple ones.

Honestly I think this is a *huge* flaw with many blogging platforms. The very nature of many personal blogs is in the wide variety of topics that the author writes about - I write about poker, personal stuff, tabulas stuff, and current events all on one blog.

The proper blogging platform shoudl allow you to use categories to create separate pages that only show these items. Furthermore, there should be separate RSS feeds for each category. This helps prevent people from having to register multiple blogs everytime they want to write about a separate interest.

But there is a strong legitimate need for some people to run multiple blogs. I'm not disputing this fact, I just think that a lot of the times the multiple blog issue is the result of limitations in the blogging platforms. In any case, someone was asking about if there was any way to cross-post between different platforms.

The short answer is, "Yes, Virginia. there is." Almost all blogging platforms support XML-RPC to some level. The only problem is that you have to download a client, then configure it to post to all these blogs at the same time.

But blogging is about posting from anywhere! Having to download a client to a local computer really is going to limit the capabilites of bloggers.

Someone needs to build a website that lets you create a listing of blogs that you own. You would post your username/password to all the accounts on this site [assuming you trust them!]. Since there really are only a few "big" blogging tools, the site would recognize when you input the site URL what type of blogging platform you used and would automatically configure to the proper XML-RPC location.

You would then post an entry to a WYSIWYG browser form, then check all the blogs you want to post to. Voila! It posts to all the sites that you list ... and it's easy. No need to copy+paste to 3 or 4 separate blogs.

Good thing Tabulas does this already with Blogger, Livejournal, and Xanga. I've been personally abusing this feature for a long time - I know a lot of my friends troll Xanga a lot more than Tabulas ... and I'm not locking them into any particular platform. View where you want to see my posts - I'll do my best to post everywhere!