What’s going on in the Tabletsphere this morning? Here’s a snapshot courtesy of the converstation grouping service we’ve been working on:
(Click on the image above for a better view.)
I’m enjoying dabbling with on this “conversation tracker” app. Although there’s much more to do to make it robust and compelling across a wide range of domains, it’s interesting enough to see what it generates up to this point.
I’ll be curious to see after using it for awhile how “sticky” the application is. Will I naturally go to it each day? Or more than once a day? Or will it slowly become more and more idle, gathering dust in the bit bin?
No matter. At least I’m learning something along the way. Here are just a few of the things I’ve noticed so far:
1. RSS feeds are all over the place in terms of how closely they follow the standard. And it’s not just small developers that don’t adhere to the format correctly. Some of the big name news sites fall victim to format errors. It takes a lot of trial and error to get things working well. It really shouldn’t be this way.
2. Feeds that redirect links are getting more and more popular. It gives an RSS tracking service a chance to see if a link has been clicked on. The tracking service then redirects the link as needed. It means that you have to be careful to get the correct link when trying to build a system that analyzes links though. The original link is what you really want–not the tracking link. Currently there appear to be two sites that are the most common in providing this kind of “service:” feedburner (most popular) and pheedo. What if there become more and more of these services? It means more and more special cases.
3. Some feeds are overly complicated–that’s just my opinion. MSDN blog feeds are a case in point here. What’s going on? Is formatting that much of a problem?
4. After working on a system that reads RSS feeds I appreciate even more all the work that the browsers do to properly handly mal-formed HTML or XHTML, even when the content has significant errors. You have to be able to intelligently handle all forms of bad syntax. What do you do when an anchor tag contains jibberish for an href (url) attribute? Well, that’s simple enough–ignore the link. But, what do you do when CDATA sections are missing? For now, the code tries to fix things up, but if the challenge is too great, that article in the feed is skipped.
5. I would have thought by this time that everyone understands how useful full feeds are. Many of the major publications are culprits here–providing partial feeds. Some go further. Apple’s PR feed, for instance, is just a list of headlines. Why not have a full feed? I don’t get it.
6. A system like we’re building here is based in part on the number of links–so links are important. Too few links and the generated list gets less interesting. The workaround is to leverage other metrics, but in some ways
7. Links are a sign of conversations. However, for some topics there don’t appear to be all that many conversations going on–even when there are lots of bloggers. Take digital photography, for example. There are many great sites and blogs, however, the bloggers don’t seem to link all that much. Instead, other metrics, such as number of comments, become more and more important.
