Wednesday, November 06, 2002

I updated my storyRoll macro, posted yesterday. The updated version, available for download, now accepts a category file name as a parameter. If omitted, it will build a list of stories (as before). If provided, it will build a list of category-specific stories. The parameter must specify the file name for the category, that is, the category's directory name. For example, assume you have a category named "All About Cats" and the category's directory name is "allAboutCats" - that is, "www\\categories\\allAboutCats". To get a list of stories within the AllAboutCategories directory, invoke:
Drop that line into a #homeTemplate.txt file that is specific to your category to maintain a list of category-specific stories separate from any other category or those on your "home" site.
5:34:59 PM      

I have an idea on how to break the feedback loop from Zoe to Radio. I could inject an aggregator callback that removes all of the Zoe news items from those elgible for Radio's new email aggregator.

Not sure how kosher this idea is. The email aggregator accumulates news items using the storyArrived callback. I could tack on my own callback to remove these items one by one. I would just have to ensure my callback runs after the mail callback. Not a problem, naming conventions control the order of execution. or I could use the postScan callback to remove the whole table of news items from the Zoe service. This is probably simpler, but must run prior to the mail aggregator's postScan callback (since that is where the email is finally sent from).

I'm going in!
4:55:02 PM      

I have an interesting feedback loop now. When Radio scans my newsfeeds, it sends an email which is then indexed by Zoe resulting in an updated newsfeed from Zoe. Radio catches this at the next scan, sending another email containing the "new" items from Zoe (items already indexed by Zoe). One downside is that Zoe is just indexing stuff it has already indexed. Another is that when I search Zoe I am going to get some duplicate hits that I really should not get. I could break this loop by not scanning the Zoe newsfeed. OTOH, I kind of like having that newsfeed as an easy way to post email items to my weblog.
3:03:15 PM      

Update on Zoe indexing Radio news feeds. It's working! Radio sent me an email with headlines from the hourly scan, Zoe picked it up and indexed it. While I could have found most of these via Google at some later point in time, that would not necessarily apply to my own posts (since Google may rank them fairly low). I like having Zoe index these posts because a) it provides a nice way to index my own posts; b) it keeps much of my writing (email and Radio) indexed in a single location.
11:34:56 AM    Google It!  

A new feature for Radio this morning, the ability to send posts to email. While this is cool for keeping up with newsfeeds while on the road (and on the wrong side of the firewall to work with Radio), it also provides a vehicle for indexing those feeds through Zoe. I am going to enable this feature and send the newscan to my own email address. Zoe will index the incoming mail (posts from newsfeeds I regularly read). Further, I think I will subscribe to my own feed (so my posts get indexed also). I'll set up an Outlook rule to categorize these Radio-driven emails so I can easily delete them after they have been indexed.
10:01:50 AM    Google It!